Author: admin_casinoutansvenkl

This is The Stepback, a weekly newsletter breaking down one essential story from the tech world. For more on the bleak state of the video game industry, follow Andrew Webster. The Stepback arrives in our subscribers’ inboxes on Sunday at 8AM ET. Opt in for The Stepback here.

Microsoft closed out Summer Game Fest with a bang. The company’s annual June showcase was packed with crowd-pleasers: Halo, Gears of War, Fable, a translucent Xbox, and even some pleasant surprises like new Persona and Crazy Taxi games. It was the kind of event that harkened back to the boisterous days of E3, when the industry was in a healthier place and game reveals were cultural events.

Just three days after the showcase, new Xbox CEO Asha Sharma warned of a “reset” at Microsoft’s gaming division, which would require “making hard choices.” The weeks that followed were filled with reports of impending layoffs, studio closures, and game cancellations. Ninja Theory is reportedly one of the studios on the chopping block, despite having just revealed a new game at SGF. If all this comes to pass, Xbox will be a shell of its former self.

After muscling its way into the console space nearly 25 years ago, Microsoft’s gaming division is at its lowest point ever. And the fallout from some disastrous decisions is going to get very ugly in the coming weeks and months.

It wasn’t always this way. With the arrival of the original Xbox in 2001, Microsoft seemed poised to be a viable contender in the space, with all of its resources helping it play catch-up with the likes of Sony and Nintendo. Major exclusives like Halo and a prescient foray into online play through Xbox Live helped to solidify this position for a time. But Microsoft flubbed the launch of the Xbox One in 2013 with an ill-fated push into non-gaming features like TV, and the brand has never really recovered. With the oft-confusing Xbox Series X / S generation, the company only fell further behind.

There are many reasons for this, but arguably the most damning was Microsoft’s extremely expensive push into subscription services. On paper it made some sense: Streaming services like Netflix were upending the film and TV landscape, so maybe the same could happen for gaming. Microsoft made some absolutely gigantic bets on this unproven future, spending billions of dollars to acquire studios and publishers in an attempt to build out a large library of content for Game Pass that would lure subscribers.

And while Game Pass proved popular initially, it ultimately plateaued, which meant that Microsoft spent all of that money on a business that didn’t grow anywhere near as large as it expected. (The service currently has around 30 million subscribers, while Microsoft had hoped to hit 100 million by 2030.) This misguided play also coincided with the “This is an Xbox” marketing campaign, which suggested that Xbox wasn’t a single console but rather a suite of Game Pass-capable devices, leading to even more confusion around the brand.

Just how bad are things? As Sharma and Xbox’s chief content officer Matt Booty wrote in the “reset” memo, “Excluding Activision Blizzard King, over the past five years, we have spent over $20 billion on ongoing investments in our content, platform, and hardware subsidy, but our annual revenue has declined nearly half a billion during that time. Going forward, this cannot continue.” The Activision deal, meanwhile, cost $68.7 billion. The company spent all of that money just to make it even less clear what an Xbox is.

This past February, there was a major shake-up at the Xbox division. Former boss Phil Spencer, who oversaw the brand through the Game Pass push and its many costly acquisitions, retired, while former president and COO Sarah Bond left the company. Despite some uncertainty around her lack of experience in the world of gaming — her prior role at Microsoft was head of the CoreAI division — Sharma’s early days provided some cause for optimism. She appeared willing to listen to fans on things like backward compatibility and exclusives, scrapped the unpopular Microsoft Gaming branding in favor of just Xbox, and moved the brand away from controversial AI features. She also made some strange and superficial changes, like restyling Xbox as XBOX.

But it’s clear the issues at Xbox run much deeper than a simple name change can fix. Sharma inherited a business that spent colossal amounts of money and had little to show for it, and now the bill is coming due. What makes this especially tragic is the sheer pedigree of the game studios that are being impacted. My colleague Tom Warren reported that Microsoft was mulling over closing at least five studios, which includes the likes of Arkane — best known for the wildly influential Dishonored series — and Double Fine Productions, a beloved team behind cult hits like Psychonauts, and more recently Keeper and Kiln. That’s multiple teams filled with talented individuals responsible for some of the most notable games ever made. Now they’re being discarded because of poor decisions they had no part in.

But even amid this apocalyptic landscape, Xbox’s issues feel particularly existential. Its hardware and subscription businesses are both faltering, and now it’s decimating its game development teams as well. Tom reported that the impending layoffs are expected to start next week, and it’s not clear yet just how widespread they will be. Part of the uncertainty is that we don’t know exactly what will happen to these studios; some may be hit with layoffs, some may be closed entirely, and some may be spun off as independent entities.

Whatever happens, though, Xbox will look drastically different once it’s all over. And given the dire state of console gaming, these might not even be the last changes for Microsoft’s gaming division.

• Sharma has done a lot of work to clean up the messaging around Xbox, but plenty of confusion remains, particularly when it comes to the company’s console exclusivity strategy.
• At the same time the Xbox is struggling, a new player is entering the space, as Valve launches the console-like Steam Machine.
• As always, Nintendo largely operates in its own parallel universe that has allowed it to largely weather the current storm.

• Bloomberg’s Jason Schreier has done some excellent reporting on the turmoil at Xbox, and has also helpfully condensed everything into a video on his YouTube channel.
• Matthew Ball is Xbox’s new strategy officer, and in an interview with The Game Business he explained how the brand is thinking about the next console, currently codenamed “Project Helix.”
• Speaking of execs, Booty talked to Game Informer following the SGF showcase to try and explain Xbox’s ever-changing strategy around exclusives, saying that “We want there to be a reason to believe and a reason to buy Xbox.”

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Andrew WebsterClose

  Andrew Webster

  Senior entertainment editor

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Andrew Webster

• ColumnClose

  Column

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Column

• EntertainmentClose

  Entertainment

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Entertainment

• GamingClose

  Gaming

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Gaming

• MicrosoftClose

  Microsoft

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Microsoft

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

• The StepbackClose

  The Stepback

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All The Stepback

• XboxClose

  Xbox

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Xbox

The Swift Observatory was launched in 2004, but recent solar storms have pushed its orbit lower, and it’s in danger of burning up in Earth’s atmosphere as soon as this year. To try and stave off its demise, NASA has enlisted Katalyst Space Technologies. The company’s Link spacecraft launched Friday with the goal of intercepting Swift, which has no propulsion system, and boosting its orbit back to its original position. Right now, Swift is circling at an altitude of 224 miles, and Link is aiming to raise that by about 150 miles.

Using a three-armed spacecraft to lift a satellite 150 miles higher into orbit is challenging enough, but the speed with which Katalyst pulled the mission together makes it even more impressive. NASA required the company to rush the job because Swift would be too low to save by October. $30 million and nine months later, help is on the way for the $500 million Swift.

The Matic is our favorite robot vacuum by a pretty comfortable margin. If you’ve been thinking about buying one, you may want to plan on doing it sooner than later. The company will raise its price by $250 on September 9th, going from $1,245 to $1,495. Matic told The Verge that the new price reflects its rising costs for memory and other components for the vacuum, which it says are now tenfold.

Those who buy one directly from the company will get a year’s worth of replacement bags worth $96 at no extra cost. Each refill contains 12 bags, and they ship for free. Additionally, Matic has increased their return policy from 60 days to six months.

Our reviewer Jennifer Pattison Tuohy praised the Matic’s “human-like navigation,” noting in her review that it got stuck only twice during six months of use. In her cluttered (Jen’s words, not mine), three-story home with pets, thick rugs, and high transitions, it handled difficult layouts with ease while delivering powerful suction and a great self-cleaning roller mop. It’ll continue vacuuming even if its water tank runs dry, and it’s quiet enough to let it run without becoming a distraction if you work from home.

It also requires very little maintenance. Rather than relying on a big multifunction dock, it carries its own water tank and stores dirty water in a disposable bag, so you’ll only need to empty and clean its dirty water reservoir. It even drives itself to the sink when it needs more water. Unlike many competing robot vacuums, it can also operate entirely offline, with maps and other data stored locally instead of in the cloud.

The US Department of Energy reportedly deleted about 6,000 pages related to energy conservation as a historic heatwave tears across the country.

The deletion was suspiciously timed, following Republican outrage over Mayor Zohran Mamdani asking New Yorkers to help reduce strain on the grid by setting their AC to 78 degrees. Republicans like Ted Cruz (who has famously fled severe weather in his home state), Nikki Haley, and Representative Nancy Mace (South Carolina) quickly pounced, framing the request as socialism and an act of war on women in menopause (the Republican Party is notoriously friendly to women’s health).

Of course, this is pretty standard advice during a heatwave. It was the official stance of the Department of Energy that Americans should set their thermostats between 75 and 78 degrees, and Republican governors in deep red states like Texas have issued the same advice in the past — including current governor Greg Abbott.

The deletions by the Trump administration are broad and indiscriminate. While pages that would support Mamdani’s request to lower thermostats were deleted, so too were pages about water conservation, types of insulation, and its solar decathlon challenge. The Internet Archive preserved the pages that have been lost.

Temperatures topped 95 degrees for four straight days in New York City, including two days over 100 degrees. Those sorts of temperatures put significant strain on the electrical grid, especially when more people are home during a holiday weekend. Setting thermostats to 78 degrees can help prevent blackouts that would leave people without air conditioning and vulnerable to the heat.

On average, extreme heat is responsible for more deaths in the US than floods, tornadoes, and hurricanes combined, according to data from the CDC and NOAA.

Erick the Architect is a founding member of, and the primary producer for, the legendary Flatbush Zombies. He’s toured the world, performed on Kimmel and Fallon, played Coachella, and collaborated with everyone from Joey Bada$$ and the Rza to James Blake and hardcore punk band Trash Talk. But perhaps the most unexpected collab was with Apple, when Erick popped up following Tim Cook’s final WWDC presentation to rap about apps. That was just a precursor to him dropping his new disco and reggae-tinged single, “No Doubt (I’m In Love).”

The new track, produced by Yeti Beats and Federico Vindver, is definitely a shift in tone from the darker, grittier, more boom-bap-grounded sounds Erick is known for. But that’s part of what makes it so compelling. Erick is still looking to experiment and expand his palette this deep into his career. That sense of adventure doesn’t stop him from getting a bit nostalgic for physical phone keyboards and the GameCube, though. It also turns out that Erick was one of just a handful of unfortunate souls who spent their hard-earned money on the Nokia N-Gage.

What is your most indispensable tool?

A moleskine book and a pen.

What is the first app you install on a new phone or computer?

I usually go for Dropbox first so I can pull up the thousands of files I have stored floating around on the internet somewhere.

What is one thing you wish you could change about your phone?

I miss typing on my phone with a physical keyboard like I used to do using a BlackBerry.

What sites do you have pinned to your tab bar?

ESPN, Behance, MyFonts, Fanatics, Topps, eBay, Discord, UPS, FedEx, Whatnot, Plex

How many tabs do you have open right now?

What is your happy place online?

Uh… the screen that says “Your order has been placed” when you buy something online

What is your favorite gadget you’ve ever owned?

Which was the most disappointing?

What game do you have the fondest memories of?

Resident Evil 4 for GameCube. I played this game so many times until my eyes were bloodshot red, and my fondest memory was the huge TV that I played it on. The TVs back then were so big they were actually pieces of furniture. When it came time to move it, you needed like four people to lift up those big tube TVs. Anyhow, I love that game, and I think I’m probably one of the best people at it!

Which tech trend do you wish would go away?

I’m not really a fan of virtual reality and putting things over my eyes to augment my reality… sorry to anyone who is obsessed with those things!

What is one thing you wish you had created?

I wish I had created the song “Bohemian Rhapsody.” It’s such a ridiculous song in the most beautiful way. I can’t compare it to any other piece of music. If I did, I’d have to reference a ton of different songs to make this one. I think if I were the creator of it, people would have no idea what song to expect from me, and that’s pretty cool.

What creation are you most proud of?

If I had a child, my answer would be that — but since I don’t, I would have to say the project I released in 2011 called “Almost Remembered.” It was the catalyst for all of the music I eventually learned how to create, and it gave me the confidence to continue to pursue being a producer and artist. I considered myself an amateur back then, but it was the most creative I’ve felt to date in terms of experimenting with new sounds.

Which are you least proud of?

I don’t want anyone to google the name of the thing I am least proud of hahaha

What’s the best piece of advice you’ve ever received?

My mom told me that everybody has a season, and although it may not be your season now… You have to consider that when it is, that season may last forever. Be patient and wait on it.

What is your current obsession?

What do you do when you need to focus?

Turn my phone off or pretend I don’t have one anymore.

What do you do when you’re feeling stuck?

Again, turn my phone off or pretend I don’t have one anymore.

When was the last time you went somewhere without your phone?

I take walks without my phone all the time, and I use my digital audio player that isn’t connected to the internet whatsoever. I can focus on the music entirely and not be distracted by incoming texts, emails, or social media.

What’s the last piece of physical media you bought?

I bought a bunch of records at VinylCon! a couple of months ago, and I’ve been collecting Absolute Batman and Invincible comics.

What do you think is worth splurging on?

Food and anything you like to collect.

What would the tagline for your biopic be?

“The man with too much on his mind.”

What’s the last GIF or meme you used?

No arguments.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Terrence O’BrienClose

  Terrence O’Brien

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Terrence O’Brien

• EntertainmentClose

  Entertainment

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Entertainment

• InterviewClose

  Interview

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Interview

• MusicClose

  Music

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Music

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

I know a vocal group of people who swear by the number pad on their keyboard. And yet, for years I haven’t cared about using one — until I put my hands on the Epomaker RT98. It’s a mechanical keyboard with a charming retro aesthetic, a fun CRT-like screen, VIA compatibility, a nice typing feel, and most importantly, a modular number pad that can be moved to either side. It’s a clever compromise in a budget friendly-ish board, allowing you to further customize your setup without losing the functionality of the coveted “tenkey.”

But like many mechanical keyboards out there, it’s got some unique quirks and tradeoffs.

$119

The Good

• Ambidextrous number pad
• Nice-feeling and -sounding silent switches
• Even nicer creamy switches
• Magnetic CRT-style screen is silly but fun

The Bad

• Plasticky build
• Moving the number pad is a slightly fiddly process
• Printed legends on the keycaps look low-quality and not sharp
• Around the starting price of nicer aluminum keyboards

The RT98 is Epomaker’s most recently crowdfunded mechanical keyboard, costing $119 for a prebuilt model configured with either the brand’s own Creamy Jade linear switches or Sea Salt Silent V2 linears. You can also choose to have your number pad on the right or the left when it ships to you, but being modular, this is something you can change on your own.

Southpaw keyboards with left-handed numpads aren’t new (Keychron has made them easily accessible for years), but this keyboard’s switch-hitter status is rare. There were a couple of old Asus ROG Claymores and the Mountain Everest Max with swappable / removable number pads, but they were costly and garish gaming keyboards. The Epomaker RT98 is much more understated and charming.

A little before and after. Left: The RT98 in right-hand numpad orientation. Right: Southpaw mode.

To fully test the RT98, I started off using it in right-handed mode with silent switches before moving the numpad to the left and swapping to the Creamy Jades. The typing experiences from both switches are solid. The RT98 is designed with the same hallmarks of most modern enthusiast keyboards: an internal gasket mounted design with layers of foam under the PCB to give the typing acoustics a fuller sound — even in a plastic chassis. It uses a solid PCB and polycarbonate switch plate without any flex cuts (cutouts allowing the whole board to flex under pressure). This makes the RT98 feel a touch sharper compared to other boards that have a pillowy bounce to their typing.

Whether you pick the RT98 with silent or creamy switches, you’re in for a fine typing experience that sounds nice. The Sea Salt Silent V2 linears have a smooth quietness to them, though that’s slightly betrayed by the louder plate-mounted stabilizers in the space bar, enter, shift, and backspace keys. It’s a fine setup for not bothering a desk neighbor or disrupting a video call with your on-the-down-low typing, but the Dry Studio ATM 98 I recently reviewed (while much pricier) is a better-sounding silent board with much nicer keycaps. Epomaker’s included PBT keycap set has nice beige nostalgia vibes, but their dye-sub printed legends look cheap and not very crisp.

The RT98 is an absolute treat when you use it with Epomaker’s Creamy Jade linears, delivering a crisp, marbley tone that’s always pleasant. When I hear “creamy” I think of the super popular F75 and similar boards from Aula, a brand that collaborates with Epomaker on some keyboards. The F-series has a bit of that sought-after raindrops-on-a-window tone. The RT98 and its Creamy Jades play in that ballpark but have a richer, darker tone than the F75 Max I’ve briefly used. I prefer the Epomaker.

The typing and acoustics are very good here, but the main draw of the RT98 is that modular number pad. I don’t work in data entry, so typing the occasional digits on the number row is fine by me. But I don’t mind the idea of having a numpad off to my left and out of the way of my mouse hand, which I can always do with a dedicated wireless pad. So I was excited to use the RT98 in a southpaw layout.

1/25The RT98 comes with a simple nylon-wrapped USB cable, a small hex wrench, a couple spare switches, and alternate Mac keys.

Changing the position involves removing the bottom screws, popping off the top cases from both the main section and numpad, swapping the internals, and reattaching everything. It’s not the most labor-intensive process, but you do have to temporarily unplug the cable for the detachable screen’s pogo pin connectors and reposition the numpad’s ribbon cable. I couldn’t find instructions in the box on doing this, but I found Epomaker’s step-by-step video tutorial and had it all done in about 30 minutes. I’m always slightly worried about messing up ribbon cables, but the process — while fiddly — was easy enough. But it did reveal an uncomfortable weakness: Prying this plastic thing open and pressing its clips back into place was nerve-wracking. I’d love to see this board made with metal.

I imagine most people will likely buy the RT98 in the numpad orientation they prefer and just leave it that way. Or, like me, change it only once. But there is something cool about not being locked into one form of 1800-layout board. Part of the fun of mechanical keyboards is how customizable they are, from the keycaps and switches to the stabilizers and down-the-rabbit-hole modifications you can make — especially if you’re an enthusiast who just mixes things up once in a while to feel alive.

1/20The RT98 does have a nice retro aesthetic, as its name implies.

The RT98’s silly little detachable screen is another way to have some customizable fun. It can show the date, time, and connection statuses or loop some built-in cat GIFs. Connecting the keyboard to Epomaker’s web-based driver allows you to sync its clock and customize the three GIFs. It was easy enough to load up some silly GIFs of my own and watch them endlessly loop. There’s not much other use for this CRT-like keyboard screen, but it’s fun. The magnetic nature of it would be more useful if Epomaker creates an ecosystem of attachable accessories, but company reps never replied to our questions asking if that might be in the cards.

PreviousNext

1/2Give me a screen for GIFs on a keyboard and I’m absolutely going to upload silly anime clips to it.

The RT98 is a very good keyboard with a cool and unique hook. But it costs anywhere from $40 to $60 more than various Aula or other entry-level boards, and its build quality is only a tad nicer. For $119, you’re around the starting price of fancier keyboards with aluminum cases. It’s just $10 shy of something like the Wobkey Rainy75, which also includes much nicer keycaps. You have to really want the modular numpad and retro vibes of the RT98.

There are so many mechanical keyboards out there, and between various sizes, layouts, and optional number pads and macro pads, you can arrange your workspace in limitless ways. The Epomaker RT98 is just another cool option with yet one more way to customize how you use it. It’s a fun thing to tinker with and enjoy, even if you’re likely to just set it and forget it.

Photography by Antonio G. Di Benedetto / The Verge

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Antonio G. Di Benedetto
• KeyboardsClose

  Keyboards

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Keyboards

• ReviewsClose

  Reviews

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Reviews

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

The Ikko MindOne Pro is delightfully small.

I keep calling it a square phone, which isn’t quite right; the screen is square, but the phone itself is slightly rectangular. The camera flips up so you can use it for selfies — you can even open it partway to use as a stand or a kind of PopSocket. There’s a Clicks-style keyboard accessory that also adds a magnetic ring and a headphone jack. I tried so hard to like it, but this phone is a miss no matter how you look at it.

I used it like a normal phone. I downloaded a minimalist launcher and tried using it as kind of a dumb phone. I put the keyboard case on. I took the keyboard case off. Nothing feels quite right.

Hip to be square.

I first met the MindOne Pro at CES back in January. It’s made by Ikko, a company based in Shenzhen that has mainly produced earbuds and audio accessories until now. The MindOne Pro is the company’s take on a small smartphone, or an AI gadget, or… both I guess. It ships globally and is priced at $499, though it’s currently marked down to $429 on the company’s website.

As charming as the hardware is, I noticed some yellow flags right off the bat when it got significantly warm as I was setting it up. Initial setup is always a lot for a phone’s processor to chew through, but the MindOne got noticeably hotter on that first day out of the box than most phones I use. A week later it has settled down a bit, though it doesn’t take much to warm it up again.

The flip-up main camera does double duty for selfies.

It also acts as a kickstand, which honestly rules.

Battery life is also pretty dismal. I watched it drop from the mid 90 percents to the 60s over the course of an hour and a half, and I wasn’t doing anything super taxing. Scrolling through reels, surfing around on Google Maps, and streaming some music — all over Wi-Fi — was enough to eat through a third of the phone’s battery in less than two hours. That doesn’t feel great.

The bad news just keeps coming. The novel camera design is a good idea, but the camera itself just stinks. Color processing is all over the place. It usually handles daylight okay, but photos taken under dim indoor lighting look too green.

The phone’s square screen is nice in theory — at least until you remember that the mobile web is built for vertical rectangles. The phone’s default behavior is to fill the entire screen area, which means cropping into the middle of vertical videos and websites. It’s not ideal and the onscreen keyboard fills up more than half the display area.

1/3The default screen aspect ratio and resolution make this interface nigh impossible to use.

Thankfully, Ikko includes some controls in the quick settings shade to help mitigate this. There’s a toggle for resolution, which fits more content onto the screen, and a toggle to change the display to a vertical aspect ratio with black bars on the sides of the screen. It’s helpful for the occasions when you run into text boxes and date pickers that just don’t work on the square screen, though it means you’re dealing with a smaller display area inside of an already small display.

Maybe that’s the point, though. I can appreciate the MindOne more if I look at it as a minimalist phone or an alternative phone to use on weekends. I like the idea of something like a Light phone that simply doesn’t run the mind-numbing apps like Instagram that I can’t seem to quit. But sometimes you need an Uber app, so I have trouble with all-in minimalist phones that feel too limited. The MindOne could be an alternative — sure, you can open the Instagram app, but it’s such a bad experience maybe you’ll do it less. Maybe I lack willpower, but I found myself scrolling Instagram just as often, except having a worse time. The best minimalist phone is a cellular smartwatch, etc. etc.

The keyboard case just didn’t “click” for me.

I even have a hard time recommending the keyboard case. Again, it seems great in theory. There’s a little battery inside and you can flip a switch to have it charge the phone up while you use it. But somehow the keys are more fiddly than just typing with the onscreen keyboard. Maybe I’m just out of practice because I swear I was pretty fast on my Blackberry Curve, but I actually found it to be slower and fussier than virtual keys. The headphone jack is a thoughtful touch, though.

There’s a whole other launcher on the MindOne Pro dedicated to AI apps that I’m not sure I get the point of. It offers a chatbot that allows you to switch between LLMs, as well as a notes app. Ikko being a Chinese company, I wondered how it would handle questions that the Chinese government might not like. I asked it if Hong Kong is part of China; it responded in Spanish for some reason and said that it couldn’t help me with my question. It even comes with a global eSIM that you can toggle on — it’s free to use with the AI launcher, but you need to pay to use it for messaging and the like. The connection is slow, at least in my neck of the woods in Seattle.

You can’t deny that this is an appealing-looking phone.

As charming as the concept is, I think a phone this shape was always going to have a tough time. Fighting to look at webpages and apps built for rectangles through a square-shaped window is just a losing battle. Maybe instead of a square-ish phone what the world really needs is a small, rectangular phone with modern bells and whistles. You hear that, phone makers? I am begging literally any company to make the iPhone Mini again with a USB-C charging port and a battery that doesn’t die halfway through the day. The MindOne isn’t that phone, at least for me. Maybe it is for someone who’s easy on the battery and isn’t picky about camera quality. The rest of us will just have to keep looking.

Photography by Allison Johnson / The Verge

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Allison JohnsonClose

  Allison Johnson

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Allison Johnson

• Hands-onClose

  Hands-on

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Hands-on

• MobileClose

  Mobile

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Mobile

• ReviewsClose

  Reviews

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Reviews

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

Over the past week, a new fanworks movement has kicked off, with the aim to root out authors using generative AI. But the detection methods being implemented are questionable, and any fanfic writer could be caught in the crossfire.

Broad distaste around the use of Claude, ChatGPT, and other AI tools has long been a thing in creative communities, including the world of fanfiction. Readers and writers have passed around tips for spotting supposedly AI-generated works, citing anything from em dashes to the broad concept of purple prose. But on June 29th, an anonymous X account called @heatedrivalryai promised a seemingly more reliable solution. It posted a skin — similar to an extension — for the popular fanfic repository Archive of Our Own (AO3) that would purportedly identify coding artifacts left behind by Anthropic’s Claude bot.

“When a Claude-generated response is pasted directly into AO3 from Claude, the text is wrapped by a Claude-injected code ‘font-claude-response-body,’” said the @heatedrivalryai account. “Its presence indicates the use of Claude definitively.” When a user visits a page (like a work of fanfic) with this code, the skin turns the entire background red.

Several test posts have been published to AO3 that allow users to check if it works. The screen immediately turned red when I tested the skin against these examples myself, and I published a Claude-generated short story to run my own experiment just in case. The red screen appeared when I directly pasted from the chatbot into the editor and vanished if I pasted text (including the exact same generated story) that didn’t come straight from Claude.

As far as “red flags” go, this is pretty hard to miss. Image: AO3 via The Verge

The Claude detector post was accompanied by examples of fanfic where the artifacts were spotted, which the anonymous creator said was meant to demonstrate the system works, not “create an environment of mistrust or accuse particular users.” But fanfic communities have quickly mobilized to publicly name and shame writers whose published works were flagged by the tool, and its creator certainly doesn’t consider AI a positive thing. “Fandom is a uniquely connective, collaborative space. It thrives on the human element and the creative spark which drives it and feeds off it,” they said. “If we unknowingly allow AI to corrupt these spaces, what will be left of them?”

Anthropic did not respond to my request to verify if the fan-made Claude detector works as described. The methodology here does look sound, however, and our own testing backs it up. There’s no apparent reason for the Claude code to be present in a story if the bot wasn’t used somehow. But there’s a clear risk of both false negatives and overgeneralizations.

The code wrapping is only preserved if text is copied directly from Claude into AO3’s editor, so it won’t catch anything edited in Google Docs or Microsoft Word and then moved to AO3 — and as someone who writes for a living, I can testify to how risky writing straight into a CMS is. Some writers who have been flagged have already updated their works to remove the artifacts, and future works can easily evade the tool.

Conversely, the tag doesn’t reveal how heavily Claude was used in a given work. That flashbanged scarlet screen could mean the entire story was fully AI-generated, or that an author pasted a few human-written sentences into Claude for spell-checking or translation, then moved them back into AO3.

That hasn’t mattered to some fandom members, who view any use of generative AI as an inexcusable betrayal to the wider creative community. Many people cite concerns over the environmental impact of the technology and how it’s trained by scraping the open web, which likely includes fanworks uploaded to platforms like AO3.

This particular tool’s applicability is limited — AO3 isn’t the only platform for publishing fanworks, and Claude is just one of many AI models. At least one person claims they’ve written separate code that can detect “Claude, Deepseek, and some ChatGPT” usage, but they haven’t released that solution to the public or explained how it works. I asked Google and OpenAI if their models leave any traceable artifacts in text generation that could be detected by similar means, but they haven’t responded.

In fact, it’d be highly surprising if a universally reliable system existed. I’ve been reporting on the issues surrounding AI detection for a few years now, and to my knowledge, there isn’t currently a reliable technological solution for distinguishing generated text from that typed out by human hands. Systems like C2PA Content Credentials and Google’s SynthID are making some progress toward identifying generative AI in images, videos, and even audio, but these rely on invisible watermarks and metadata that don’t carry over for copy-pasted text.

AI companies have every incentive to at least solve the problem internally

That could change in the future, and AI companies have every incentive to at least solve the problem internally. Early models were trained on text indiscriminately scraped from the internet, and as human writing is crowded out by its synthetic counterpart, they could risk a “model collapse” scenario that would degrade the accuracy of outputs.

For now, though, fandom communities are still mostly relying on vibes. Most fanfics aren’t judged by a tool like the AO3 skin, but by “tells” that could include anything from specific sentence structures — like the notorious “it’s not X, it’s Y” — to overuse of flowery metaphors. (At least nobody in fandom, so far, has benches becoming men.) But we have to remember that AI often Writes Like That because it was trained on stuff that real people have written. It’s trying to replicate us. I’m not bold enough to share my own AO3 bookmarks, but I’ve definitely read some overly grandiloquent fanfics in the pre-ChatGPT internet days that wouldn’t pass this dubious sniff test.

The best solution for distinguishing AI works on AO3 is already available: the site’s robust tagging system. A “Created Using Generative AI” tag exists, and many authors do include it to disclose the use of tools like Claude. That requires honest transparency, though, and there’s little incentive for honesty given the backlash. It’s also worth remembering that fanfiction is supposed to be a hobby, not a regulated industry.

With these efforts to prevent AI from taking eyeballs away from genuine human-driven creativity, authors who don’t conform to what’s deemed to be an acceptable quality of writing may become innocent victims of the ongoing witch hunt. At least one writer has already been caught up in this because another person they trusted to edit their fic did so using Claude. So if the next fanfic you read feels a little robotic, just bear in mind that it might not actually be the product of a robot.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Jess WeatherbedClose

  Jess Weatherbed

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Jess Weatherbed

• AIClose

  AI

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All AI

• AnalysisClose

  Analysis

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Analysis

• CreatorsClose

  Creators

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Creators

• CultureClose

  Culture

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Culture

• EntertainmentClose

  Entertainment

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Entertainment

• Internet CultureClose

  Internet Culture

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Internet Culture

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

Despite my initial skepticism, I’m now sold on wireless Qi chargers that add integrated fans to keep your phone cool while charging. I figured they’d be too loud, or too weak, or too gimmicky, but I’m a convert after spending a week with the new $59.99 Kuxiu D5 Qi2.2 charging dock. Its active cooling system has kept my phone from heating up, unlike every other Qi charger I’ve tried. Good thing, given my own history with an overheating phone.

Last year I fried the logic board on my titanium iPhone 15 Pro. I was on a sweltering train and trying to edit a 4K video on my phone, which was charging from a magnetically attached Qi power bank. Oof, my iPhone is getting really hot, I thought, just as the screen went forever blank. The repair bill was €660.33 (over $750), for which I paid nothing because it was just inside the two-year warranty. Apple’s “genius” wouldn’t say if the phone overheated, but I’m not keen to tempt fate twice.

Heat is the enemy of batteries

Heat is the enemy of batteries. High temperatures accelerate the chemical degradation of lithium-ion cells, permanently reducing how much charge they can hold over time. The electromagnetic induction employed by Qi chargers is inherently inefficient, producing far more heat than transferring electricity through a cable plugged into your phone. That heat can cause phones to throttle the wireless charging speed in order to protect the battery.

Despite knowing this, I keep coming back to Qi charging docks and power banks because they’re so damn convenient. I’m also addicted to Apple’s StandBy feature that turns my phone into a desk or bedside clock and widget machine when docked and charging in landscape orientation.

Fortunately, companies like Anker, Aukey, ESR, and Kuxiu have started building active cooling systems directly into their Qi chargers. They feature tiny heat sinks and ultra-quiet internal fans that actively blow the heat away from the back of your phone while it charges. This keeps the battery significantly cooler and allows for faster charging speeds. Powerful magnets ensure perfect alignment between the phone’s receiver coil and the charger’s transmitter coil to maximize efficiency and minimize heat waste.

The display on the D5 shows the charging status of up to three devices, and whether the active cooling is enabled.

A look from behind at the cooling fan beneath the Qi2.2 phone charger.

This button lets you switch off the display, switch off active cooling, and retract the watch charger.

The 45W GAN charger is a great in-the-box addition for such an inexpensive Qi2.2 charging dock. Here the watch charger is shown recessed.

The fan inside the Kuxiu D5 operates silently, but it can still be turned off if you want. The display is useful, but the little winky-faced screensaver is not. I do like knowing the time and how much charge each device is pulling, in addition to the total power draw, but the animated movement is distracting when caught out of the corner of my eye. It’s also a bit too bright for my taste when perched on my nightstand. Fortunately, I can turn the display off with a tap of a button.

I like that the Kuxiu D5 ships with a 45W USB-C PD charger in the box. Otherwise, you’d have to guess if whatever charger you have lying around will support the dock’s max total output, which can deliver up to 25W to Qi2.2-compatible phones, 2.5W to Qi-compatible watches, and 5W to Qi-compatible earbuds. It’s compatible with several generations of phones, watches, and earbuds sold by Apple, Samsung, and others.

Kuxiu calls the D5 a “5-in-1” charger, which I find misleading since it’ll only charge up to three devices simultaneously. It seems to include the display and fan in its total. By that count, I’d be happier if this was a 4-in-1 device, since the added cost for the display won’t add much value for most people, and it can only cause issues down the road.

The watch charger is also a little silly in its over-engineering. A long-press on the dock’s sole button kicks off a tiny motor to slowly push the watch dock out the side. I guess it adds to the blocky dock’s durability when traveling, but not by much, and it can only be retracted when the dock is plugged into the wall.

1/7The display when charging a single device.

Overall the Kuxiu D5 is a fine Qi2.2 charging dock made better by active cooling. It’s equally at home on a desktop or nightstand, and priced below similar docks from Anker, ESR, and others.

Active cooling on Qi chargers is a good idea whose time has come. Phones sold by Huawei, Xiaomi, and Oppo have supported wireless charging speeds of 50W and above for years. But that’s only possible when placed on proprietary charging docks capable of whisking away all the extra waste heat. Now the body responsible for Qi is working on a global 50W standard for release by 2028. Manufacturers will likely need to adopt active cooling to reliably sustain those higher charging rates.

Fine by me. Whatever it takes to keep that vital Qi life force flowing.

All photography by Thomas Ricker / The Verge

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Thomas RickerClose

  Thomas Ricker

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Thomas Ricker

• Accessory ReviewsClose

  Accessory Reviews

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Accessory Reviews

• ReviewsClose

  Reviews

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Reviews

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

The AI startup, best known for generating images, released a behind-the-scenes video of its dunk-tank ultrasound scanner, which it plans to deploy in spas and hopes will transform medicine with cheap, detailed, radiation-free imaging. The nearly 20-minute tour comes from tech YouTuber Marcin Plaza, who also happens to be an engineer at the company.

Plaza frankly describes the scanner as scores of ultrasound probes “hacked apart and slapped on a glorified hot tub with an elevator in it,” connected to off-the-shelf computers and Raspberry Pis. The video shows more of the hardware and the team building it, but largely glosses over the physics and imaging questions experts raised when Midjourney first announced the project.

Those experts told The Verge that Midjourney had shown little evidence it could overcome the well-known limits of ultrasound, a technology that has been around for decades, or generate the kind of detailed images it has suggested at the scale and speed it is promising. The company has emphasized that the scanner will launch as a wellness product focused on body composition, rather than as a diagnostic medical device, which would require FDA clearance and clinical trials. Midjourney doubled down on that approach in the video: head of medical Tom Calloway said the focus on body composition would let the company “speedrun” and open right away once testing is complete. But the video still leans heavily on medical language, asking what physicians could do with access to frequent scans taken over time.

Calloway did not seem especially concerned about clearing up any confusion in the video. “I don’t think there’s anything to really clarify,” he said, promising frequent blogs with progress updates. CEO David Holz, meanwhile, said Midjourney’s lack of investors gives the company freedom to pursue the scanner. “No one can tell me not to do it,” he said.

Since its inception, Apple TV, née Apple TV Plus, has built a reputation on quality over quantity. It has far fewer shows and movies than the likes of Netflix or Disney Plus, but generally speaking, the projects it does put out are quite good. It’s a strategy that has brought comparisons to the HBO of old, and 2026 has featured a particularly strong mix of new hits and returning favorites. It took a few years, but Apple TV seems to be finally hitting its stride.

That’s a pretty solid lineup on its own. But what has really helped flesh out Apple TV’s offerings is that those shows are coming out alongside the continuation of existing properties, many of which have been chugging along for several years. The last few weeks have included new seasons of the sci-fi detective series Sugar, the postapocalyptic thriller Silo, and Star City, a spinoff of the long-running alternate history series For All Mankind.

The next few months look like a similar mix. Dark Matter, Slow Horses, and Ted Lasso will all be back with new seasons. Meanwhile, there are brand-new thrillers like Last Seen and the Anya Taylor-Joy-led Lucky, along with a pair of buddy comedies: Ryan Reynolds and Kenneth Branagh in Mayday, and Matthew McConaughey and Woody Harrelson in Brothers. Maybe this year Apple will even finally put out its long-delayed series The Savant, which explores domestic terrorism.

It took a few years to get to this point. That required landing on not only a handful of hits that could go on for multiple seasons, like Slow Horses and For All Mankind, but also exploring different genres. Apple has been particularly successful in the realm of science fiction, but many of its most talked-about series — like Ted Lasso and Widow’s Bay — are comedies. At this point, the service offers a strong mix of both lighthearted fare and more traditional prestige television. It’s a trend that looks to continue, at least for the foreseeable future. The likes of Silo and For All Mankind are coming to an end, but new seasons of heavy hitters Pluribus, Severance, and awards darling The Studio are waiting in the wings, while a much-anticipated adaptation of William Gibson’s cyberpunk classic Neuromancer is also incoming.

What’s particularly notable about Apple’s focus on premium-style television is how it differs from the growth mindset of its competitors. Netflix is the prime example. With the end of series like Stranger Things and Squid Game, the streamer has become less reliant on tentpole TV shows. Instead, reality television, live programming, podcasts, and even interactive games have become a big focus as Netflix attempts to reach a large, broad audience. Disney, similarly, has been experimenting with turning Disney Plus into “the immersive, interactive digital centerpiece of the company,” which will involve focusing on things other than traditional programming to lure subscribers.

Apple TV obviously isn’t perfect; its film offerings leave a lot to be desired, and upcoming features like the toy car adaptation Matchbox don’t instill a lot of confidence. It has also begun expanding into areas like live sports, including Formula 1. But on the television side, the streamer’s approach is refreshing. While nearly every subscription service seems steadfastly focused on appealing to as many potential customers as possible, Apple TV is simply a place for good TV. Signs point to that potentially changing in the future, but for now the service is the closest thing streaming has to a prestige home.

At the event “The Briefing: AI for Science” earlier this week, Anthropic announced Claude Science, a new “AI workbench for scientists” that pulls fragmented tools and datasets into one environment, and generates figures and visuals. Anthropic, already dominating the industry with its popular coding tools and powerful AI models, framed the launch around what it says is AI’s potential to “dramatically accelerate the pace of scientific discovery and the development of healthcare interventions,” and touted a long list of biotech and pharma customers already using Claude.

Anthropic also went a step further, saying it would develop drugs of its own. Head of life sciences Eric Kauderer-Abrams said the company will focus on discovering treatments for “neglected” diseases.

AI companies have been eager to court science and pharma customers — OpenAI, Amazon, Google, and others have their own life sciences tools and platforms. But Anthropic’s planned move is one of the most direct public attempts by a major frontier AI company to actually develop drugs itself. It puts it in the unusual position of selling software to other, potentially competing drugmakers. Anthropic joins a broader race that includes AI-first drug companies like Insilico, Google DeepMind spinout Isomorphic Labs, biotech startups, and Big Pharma companies building or buying AI tools of their own.

Anthropic has provided very few specific details about what it hopes to accomplish in the drug development space. At the event, Kauderer-Abrams didn’t say what the company would do if it finds any promising drug candidates. Anthropic did not respond to The Verge’s requests for comment seeking more details, including what diseases it plans to target first and whether it would partner up with other companies for lab work, animal testing, clinical trials, or manufacturing.

AI is applied at “every single stage of drug discovery.”

Experts told The Verge that the uncertainty surrounding Anthropic’s plans reflects a broader uncertainty around the AI drug boom itself. “AI drug discovery” can mean many things. It “is a really broad term,” explained Namshik Han, a professor at the University of Cambridge and cofounder of AI biotech startup CardiaTec. AI is applied at “every single stage of drug discovery,” he said, from finding new compounds and improving them to supporting research, data analysis, clinical trials, and even manufacturing. Every major drug company will be using AI in some way, he said. Matthew Todd, a professor of drug discovery at University College London, echoed the sentiment that AI already pervades drug discovery and research, calling it a “catchall phrase” given its broad array of uses.

AI is undoubtedly changing drug development. Han pointed to the numerous initiatives by pharma giants like AstraZeneca, Novo Nordisk, and GSK, and said AI can already help generate possible drug ideas, such as by suggesting new molecules that could interact with parts of the body like cell receptors that are already known to be involved with a particular disease or are targets of existing drugs. Todd said it’s immensely useful for speeding up research and helping “road test” new drug ideas. Given Anthropic’s work on frontier models, the company would presumably use generative AI to search across vast chemical and biological possibilities and help researchers make connections that would be difficult or slow to find otherwise, potentially suggesting new drug ideas, identifying new disease targets, or finding new uses for existing drugs.

But that is still a long way from an AI-designed drug reaching patients. Todd said the field is “a long way off” from an AI-designed drug being approved by regulators for human use. He added that the drug discovery process would not run autonomously, with human input and supervision required throughout. Todd and Han both noted the lack of publicly available, high-quality experimental data, such as how various chemicals behave in the body, could slow drug development efforts as well, stressing that even for well-studied areas of biology there are still large gaps in our understanding of how things work.

AI models “haven’t yet come close to making experiments unnecessary.”

AI is not positioned to fix many of the slowest parts of drug discovery. Frank von Delft, a professor of structural chemical biology at the University of Oxford and head of protein crystallography at the Oxford Centre for Medicines Discovery, said people are right to get excited about advancing AI models, but they “haven’t yet come close to making experiments unnecessary.” Drug candidates still have to be tested in the real world for efficacy, toxicity, and whether they have practical properties allowing them to be prepared, stored, and delivered safely as medicines. All of that requires skilled workers, a lot of money, and time, especially clinical work in humans — a point when many promising drug candidates fail. If Anthropic wants to develop a drug, von Delft said, it is “going to have to spend a lot on experiments.”

It’s possible Anthropic is willing to try. In the last year, the company has been actively hiring biologists and building its own wet labs, and as of writing it has several live applications hiring for life sciences roles. Han said Anthropic has been “actively recruiting” in the area too, adding that several of his academic colleagues had been approached by the company. Without naming names, Han said he thinks Anthropic has successfully hired a few candidates away from Big Pharma and prestigious academic institutions.

With all of this complexity, whatever disease Anthropic picks, any payoff is likely a long way away — at the very least, the better part of a decade, given how long it typically takes a new drug to go through clinical trials. There’s “always a big lag time” with testing medicine, Todd said. “It takes time to show experimentally that something’s safe.” No AI-designed drug has yet made it through clinical trials and FDA approval to reach market. Some AI-developed candidates have entered clinical trials, but it’s hard to know how much AI contributed, where in the process it was used, or whether those candidates outperform conventional drugs. AI can speed up part of the search, but drugs still need to prove themselves the old-fashioned way: in slow, methodical experiments that take place in the real world.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Robert HartClose

  Robert Hart

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Robert Hart

• AIClose

  AI

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All AI

• AnthropicClose

  Anthropic

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Anthropic

• HealthClose

  Health

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Health

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

• ScienceClose

  Science

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Science

This is Optimizer, a weekly newsletter sent from Verge senior reviewer Victoria Song that dissects and discusses the latest gizmos and potions that swear they’re going to change your life. Opt in for Optimizer here.

Europe is melting, the eastern US is currently trapped in a “heat dome,” the Midwest has the corn sweats to look forward to, and if you’ve never felt the oppressive, sticky misery of monsoon season in Asia — consider yourself blessed. But as folks duke it out on social media about who has it hotter (and whether air conditioning is ethical), it’s time to address a nonnegotiable truth of surviving summer: hydration.

Surely, staying hydrated isn’t a controversial topic. Something as simple as drinking water can’t possibly be twisted by the wellness wild west. But alas, a recent TikTok search proved me wrong. “Sometimes water alone just isn’t enough,” an influencer named Grace, a “holistic nutritionist,” says in a partnership video with Liquid IV. She mixes a packet in, stirring with an aesthetic glass straw while spitting general truths about why electrolytes are important. “It’s about what your body does with the water that matters.”

Next, I was blasted with a clip of actress Mayim Bialik’s podcast emblazoned with the words “You’re hydrating wrong!” In it, Bialik’s guest, who is billed as an “exercise physiologist and nutrition scientist,” claims that drinking a lot of water won’t hydrate you properly. You’ll just pee it out, she says, because plain water doesn’t have the “ideal pressure” due to a lack of sodium and glucose. Rounding out my cursed TikTok search, another so-called nutritionist stated that ice water is actually dehydrating based on traditional Chinese medicine principles because the “body cannot hold onto the water.”

These are all peak examples of the wellness grifter playbook.

It’s true that the body needs electrolytes, minerals that hold a positive or negative electric charge when dissolved in water: things like potassium, magnesium, and sodium. (Sugar isn’t technically an electrolyte, but it too is necessary for hydration.) They’re important because they help maintain appropriate fluid levels within your cells, contract and relax muscles, and play into several other chemical reactions within your body. You lose electrolytes when you sweat and whenever you pee as waste products. These are all facts that influencers and wellness brands use to establish credibility by letting you in on the “real science” of how hydration works.

The next turn in the playbook is then convincing you that “drinking more water” — a long-established, common-sense way to hydrate that’s often recommended by doctors — is not good enough. There is a more optimal way, a secret that you’re about to be let in on. To finish it off, you’re either fed a dubious but harmless hack (e.g., room-temperature water is better for hydrating than cold water) or told the electrolyte powder they’re shilling will hydrate you so much better or faster than plain water could.

Liquid IV can be useful to have around. This presentation, however, is a bit of science-washed marketing. Screenshot: Liquid IV

Take Liquid IV’s science page. “Hydration is essential,” the site says, as a video of stylish scientists clad in chunky glasses and crisp lab coats plays in the background. “We elevate it with science.” Specifically, “breakthrough technology, clinical programs, and scientific leadership.” Scroll a bit further down the page, and you’ll see the words “Water is the least studied nutrient” and “Just add science … We love water (it’s a phenomenal hydrator), but sometimes water alone isn’t enough.”

The site then goes on to describe Liquid IV’s four-step hydration philosophy. It boils down to: (1) making drinks tasty so you drink more; (2) adding “hydration multipliers,” aka a proprietary blend of electrolytes; (3) taking credit for sodium in the mix helping you retain water; and (4) taking credit for the mix’s other electrolytes doing what electrolytes do. A separate page for its clinical study results says that Liquid IV helps people rehydrate quicker and retain fluids longer than water alone. Which, again, is how electrolytes work. It also says that the sugar-free version doesn’t spike blood glucose levels, which, to state the obvious, is an expected result from a sugar-free beverage.

This is effective at selling the message that, in a vacuum, Liquid IV — and other electrolyte mixes — are better than water at hydration. But it’s also missing a lot of key context.

The FAQ on Liquid IV’s site notes that the product is safe for daily consumption “during or after tough workouts, while traveling, after a night out, or for heat-related rehydration.” That’s not “replace all your water with Liquid IV.” It goes on to suggest one serving a day and to consult a doctor for further questions. That’s all fine, but electrolyte imbalances are a two-way street. There’s such a thing as too many electrolytes.

JUST ADD SCIENCE! These are facts about electrolytes, but it’s missing some key context to sell you on buying in. Screenshot: Liquid IV

Take sodium. The CDC says that the average American gets about 3,300mg of sodium per day, which is above the federal recommendation of less than 2,300mg. Liquid IV has over 20 flavors, so the nutrition label varies, but generally a single serving has roughly 500 to 520mg of sodium — or more than 20 percent of your daily recommended intake. Barring a condition or medication that affects your sodium levels, like POTS, you probably get enough of this particular electrolyte from your diet. And salt isn’t the only ingredient to be mindful of. Ask any marathoner or cyclist. Chugging too much of a sugary sports drink can trigger a nasty case of the runs.

But say you are someone who does need to be extra mindful of hydration. Overdoing electrolyte mixes is easy to do and can have unintended health consequences. My spouse suffers from severe, chronic migraines. In the summer, they’re often triggered by dehydration and heat intolerance. Migraines are horrible, but my spouse is doubly unlucky in that theirs often come with intense nausea and vomiting. It’s a vicious cycle where dehydration worsens the migraine, but the nausea makes it so they can’t actually rehydrate themselves. Several times, I’ve had to rush them to the emergency room for a saline drip. So in an effort to be proactive, last summer they started drinking Pedialyte and Liquid IV daily to prevent heat-related migraines.

It was a genuine attempt to take better care of themselves. But they ended up giving themselves high blood pressure from all the salt — which can happen with too much electrolyte supplementation, along with other issues like increased cardiovascular strain.

Hear me out: plain water is good enough for Theo Hernandez of France at the 2026 World Cup. Photo by Marc Atkins/Getty Images

There is a time and place for electrolyte mixes and sports drinks. Mainly, when you need to quickly rehydrate yourself after losing a significant amount of bodily fluids in a short time and your daily diet can’t compensate for it. So, if you’re sweating buckets because you’re active outdoors for several hours in the hot sun? Yeah, an electrolyte drink is appropriate. If you, a healthy human, decided for some cockamamie reason to go for a 30-minute run in extreme heat, but will immediately go back into air conditioning to chow down on Grandpa Rick’s barbecue? You probably won’t die drinking a Powerade but could also stick with plain water. If you caught norovirus and are gastrointestinally suffering out both ends, go with God and get that Gatorade. But you don’t need to go crazy about how you hydrate. You don’t even have to spend a ton of money on electrolyte packets. Half a lemon and a pinch of salt in your water is an effective, budget-conscious alternative. (Or pickle juice, if that’s more your vibe.)

I deeply understand the desire to “hydrate better.” I’ve passed out in hot weather, sometimes while long-distance running. It’s not fun! The few times it happened were so terrible, it motivated me to be better about hydrating in hot weather. It’s why I’ve tested multiple wearable sweat patches that claimed to help optimize my electrolyte intake when I was training for half-marathons in sweltering heat. It’s why I’ve reviewed smart water bottles of many flavors and variations. (Adding flavoring is also a valid “hack” to drinking more water.) Earlier this year, I installed an at-home urinalysis lab in my toilet because of course I did. I’ve invested in a ton of sports hydration packets and energy goos. Ironically, what’s helped most is a regular Owala bottle. Apparently, I like drinking water better out of a straw.

Which is a fancy way of saying I’ve firsthand tested several “optimized” ways to hydrate. My verdict is common sense, scrolling past influencers, and in most cases, drinking some nice cold water are good enough.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Victoria SongClose

  Victoria Song

  Senior Reviewer, Wearable Tech

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Victoria Song

• ColumnClose

  Column

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Column

• HealthClose

  Health

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Health

• OptimizerClose

  Optimizer

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Optimizer

• ScienceClose

  Science

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Science

12 months, 16 days, 21 hours, and 54 minutes after I first heard about Trump Mobile’s T1 Phone 8002 (gold version), I’m finally holding one in my own hands.

That’s right, The Verge’s Trump phones have arrived. And yes, I do mean phones – we have three of them: We ordered two, paid for two, and received three, all of which were sent to the wrong address. You can’t beat service like that.

My phone came with a free Trump Mobile SIM card, even though I’ve not signed up to the company’s mobile plan, plus an A4 quick start guide that was quite clearly printed on a standard office printer. Inside the box, there’s a SIM tool, clear plastic case, 33W charger, and a gold-and-black braided USB-C cable that feels remarkably cheap. Then you get a second set of instructions, this time a tiny booklet that covers everything from initial setup to how to make a phone call or take a photo with the camera. I get the feeling this phone isn’t intended for a technically minded buyer.

As for the phone itself, my first impression was how oddly tacky it is. And I mean that quite literally — the gold, plastic back panel has a slightly sticky sheen to it, and has already picked up plenty of dust and fingerprints in just an hour or two. Even more encouragingly, my phone has a tiny scratch in one corner that was there before I even took it out of the box.

Setup went pretty smoothly at first. It installed all my usual apps from my Google backup to go alongside the preinstalled Truth Social, claims its February Android security update is the latest it needs, and unlike YouTuber Quinn Nelson my fingerprint sensor works just fine. I’m based in the UK though, and so far the Trump phone isn’t playing nicely with my British SIM card — I can make phone calls just fine, but it won’t connect to LTE or 5G. I’ll have to see if I can fix that, or if the Trump phone is strictly for Americans only.

Either way, I’ll have a full review of the T1 Phone next week. Let me know in the comments what you want me to test.

Photography by Dominic Preston / The Verge

Got inside information on Trump Mobile or the Trump phone? Reach out securely from a personal device to tips@theverge.com, or see our How to Tip Us page.

Despite the fact that smartphones have become impressively capable shooters, standalone point-and-shoot cameras are enjoying a renaissance. The tiny Kodak Charmera is still wildly popular, while influencers are scrambling to find aging Canon cameras on eBay. Godox, a company best known for its photography lighting products, is the latest to join the simple camera craze, but its new C100 stands out from the crowd by skipping a color preview screen for a transparent LCD that doubles as an optical viewfinder.

Although the C100 product page on Godox’s website is thin on technical details, including its sensor resolution and video capabilities, it shows off the camera’s viewfinder, which can display information like battery life and shooting settings. The device can even function as a light meter, measuring the brightness of what you’re framing and suggesting exposure settings that can be used on another camera.

The 65-gram camera can capture images and videos in four different aspect ratios, including 16:9 and 1:1, which are all stored on a microSD card up to 128GB in size. There’s no wireless connectivity, but you can transfer images to a PC or mobile devices over USB-C. The Godox C100 has launched in China but there’s no news on if it will ever see an international release. There’s no pricing info on its product page, either, but according to Digital Camera World the C100 is just ¥199, or around $29, making it even cheaper than the Charmera.

After success with the pig eyes, the team members then tested their device on human eyes. They first collected 12 eyes from six people who had died. In each case, one of each pair of eyes was put in the device, while the other was not. Again, the perfused eyes did better—and their retinas were preserved.

Cosma and her colleagues hope that their device could offer scientists a new way to study eye treatments—one that doesn’t involve experimenting on living animals. They also hope that, with some improvements, the ECaBox might provide a way to maintain and revive donated human eyes for whole-eye transplantation.

Whole-eye transplants have been attempted in the past, mostly in research animals, with limited success. In May 2023, a team at NYU Langone transplanted an eye along with part of a face to a man who had survived a high-voltage electrical accident that resulted in the loss of much of the left side of his face, including his left eye, two years earlier. Although the man recovered well, he wasn’t able to see out of the transplanted eye.

We won’t know whether eyes treated in the ECaBox could do any better until they have been transplanted, says Tessier. 

In the meantime, Cosma and her colleagues plan to use a newer version of their device to collect more human eyes for research. “We are planning to develop a portable, surgery-room ECaBox to minimize [degradation] in heart-beating donor eyes, when they become available,” they write.

The Fire HD 8 that launched in 2024 was the last new addition to Amazon’s budget-minded tablet lineup, but the company has quietly updated the Fire HD 10 that debuted the year before. In 2023 it was offered with multiple storage configurations that each came with 3GB of RAM, but the 32GB version now ships with 4GB of RAM, and a small price bump from $139.99 to $154.99.

The Fire HD 10 with 64GB of storage still only comes with 3GB of RAM and the other specs for both tablets remain the same, including a 10.1-inch, 1,920 x 1,200 display, a 2GHz eight-core processor, a 13-hour battery, and expandable storage through a microSD card. The refreshed version is also only available for purchase with lock screen ads, but those can be removed after the fact by paying a one-time fee. If you’ve been eyeing this tablet but don’t want to pay a $15 premium for it, Amazon still sells the original 3GB/32GB model for $139.99, but stock may sell out.

As an early indicator of the need for the faster rhythm, it issued two security advisories dealing with a number of critical vulnerabilities on June 30 — the fifth Tuesday of that month: APSB 26-28 and APSB26-29. It is not alone in issuing out-of-sequence patches for urgent fixes: In April, Microsoft also released one to react to a particular threat.

Adobe said in a blog post that it is responding to the increased level of threats: “Twice-monthly bulletins will enable us to keep pace with the era of frontier AI. More vulnerabilities found means more fixes to deploy and a once-a-month publication window is no longer fast enough to stay ahead of our adversaries. This new cadence is the direct result of investing in improved vulnerability discovery.”

The new schedule will be effective from July 14 and will apply to every advisory that includes a formally published CVE requiring customer action.

The attacks all came from a single source, an IPv6 address range controlled by internet provider LSHIY LLC, Huntress said in a blog post. LSHIY has since terminated access for the customer using the IP addresses involved in the attack.

Huntress had been monitoring spray attacks for some time and had noticed a slight increase from June 12, and then a sudden spike on June 22 when 30 of its customers were affected.

The attackers replayed validated credentials via the OAuth ROPC (Resource Owner Password Credentials) flow. This takes a username/password at the /token endpoint for a tenant and mints a new user-delegated token, once provided with the correct credentials. This was possible because multi-factor authentication (MFA) had not been configured to handle the techniques deployed by the attackers.

It’s a big year for America. It’s the semiquincentennial, otherwise known as America250, and the United States is cohosting the World Cup. But spectators at these events — and the millions of people who live in the cities hosting them — may not realize that they, too, are being watched.

From Kansas City to New York, the US cities hosting the World Cup have been ramping up their surveillance capabilities in the months leading up to the tournament. Security measures are at an all-time high in Washington, DC, which isn’t hosting the World Cup, but is home to a series of spectacles this summer. The Fourth of July festivities in the nation’s capital will have an unprecedented level of surveillance. Law enforcement agencies say they can’t take any risks during these once-in-a-lifetime events — but privacy advocates warn that some of this surveillance won’t be limited to this summer’s celebrations.

Both the Fourth of July fireworks on the National Mall and the July 19th World Cup final in New Jersey have been designated National Special Security Events (NSSE) by the Department of Homeland Security, the most stringent security designation the agency gives. This isn’t unusual for major sporting events — the Super Bowl is always given an NSSE designation — but it’s a first for the Fourth of July. The UFC fight at the White House in June was also an NSSE, as was the official UFC watch party on the Ellipse.

Attendees at the Fourth of July fireworks show on the National Mall will have to pass through airport-style security checkpoints and won’t be allowed to bring folding chairs or coolers. Counter-drone measures will be in place, The Washington Post reports, as will bomb technicians, countersnipers, and medical personnel from several federal agencies. While attendees will notice these security measures, others could be close to invisible — including camera networks that track their biometrics.

The measures at the National Mall appear to be a response to criticisms of lax security at the White House Correspondents’ Dinner, which was infiltrated by a gunman who allegedly shot at a Secret Service agent.

There will be similar measures in place at the World Cup final, which Donald Trump is expected to attend — and where he will reportedly present the trophy to the winning team.

“This is going to be security-o-rama regardless of whether the president goes,” Jules Boykoff, author of Red Card: The 2026 World Cup, Sportswashing, and the FIFA Greed Machine, told The Verge. “If the president goes, that’s just an extra lacquering of security.”

Boykoff, a professor of political science at Pacific University, said there may be an increased ICE presence at the World Cup final as well, and pointed out that ICE arrested rapper 21 Savage at the 2019 Super Bowl — another NSSE — claiming he overstayed his visa.

Anne Toomey McKenna, an attorney who specializes in privacy and biometric surveillance, said the NSSE declaration may also make it easier to justify collecting communications data under the looser standard of Section 702 of the Foreign Intelligence Surveillance Act, instead of the more stringer requirements of the Wiretap Act.

Andrew Giuliani, executive director of the White House task force for the World Cup — and son of Rudy Giuliani — has said there will be heightened security at all the World Cup matches, even the ones Trump doesn’t attend. “You’ll have multiple perimeter checks from security. You’ll have checks while you get onto public transportation to make sure you’re a valid ticket holder,” Giuliani told the Atlantic Council’s Frederick Kempe. “Soccer fans — or futból fans — they generally like to come to stadiums late, in the 15, 20 minutes or so before the game.” But Giuliani said ticket holders should know that gates open three hours before kickoff and plan to arrive early so they don’t miss kickoff.

The surveillance isn’t limited to one-off events and in fact involves building up a massive apparatus across the country. Through FEMA, the Department of Homeland Security gave $250 million in grants to states that are hosting World Cup matches, much of which was used to buy counter-drone equipment, according to The New York Times. The FBI has also been training local law enforcement agencies on drone mitigation. According to Giuliani, the Fan Fests in all 11 host cities will be covered by counter-drone technology. It’s unclear whether these cities are using the same tech that led to an airspace closure in El Paso earlier this year.

This is going to be security-o-rama regardless of whether the president goes.

New York City — technically one of the host cities, even though the matches are taking place across the river in New Jersey — spent $6.5 million on counter-drone technology. In Kansas City, Missouri, authorities have confiscated at least 16 drones since the World Cup began.

“The general rule with the World Cup and Olympics is that local and national police forces use the sports mega event like their own private cash machine,” Boykoff said. “The World Cup creates a state of exception that allows for all manner of securitization processes.” And in many cases, once these tools are in place, they remain. Paris, for example, enabled AI video surveillance ahead of the 2024 Olympics — and is keeping it in place through the end of 2027 despite privacy concerns.

Similar camera systems have been installed throughout the US ahead of the World Cup, even in areas far beyond stadiums. Kansas City also planned on putting cameras equipped with facial recognition on some city buses, even though the state government refused to fund the project over privacy concerns. The city initially went through with the program anyway, saying it would help identify missing persons and could thwart human trafficking attempts during a major international sporting event. City officials said that the images captured are checked against active missing persons alerts and only retained if there’s a match.

“Privacy is always a tricky thing,” Tyler Means, chief mobility and strategy officer at Kansas City Area Transportation Authority, told The Washington Post. “We’ve always had cameras on our buses. It’s just new technology. I think in time it’ll smooth over and people will realize, ‘Well, it didn’t really feel any different.’”

The cameras aren’t operational yet because of backlash and technical delays, but Kansas City plans on implementing the program later this year — even though the World Cup will be over by then.

The America250 celebrations will be monitored by thousands of law enforcement officers, including National Guard troops and FBI agents, many of whom will be wearing body cameras. Several cities have expanded or reactivated CCTV systems ahead of the World Cup. Seattle reportedly reactivated dormant cameras after FBI and Seattle Police Department officials briefed the mayor on “credible threats” during the games.

McKenna said the increase in surveillance at these events isn’t unwarranted given the increased level of risk, but said there’s an issue with how biometric data is gathered and retained. McKenna noted that British Columbia, which is also hosting the World Cup, has regulations around how long surveillance footage from matches and other events can be retained — rules the US lacks.

Though CCTV has been around for decades, advances in camera technology — and AI integrations — have made these systems incredibly sophisticated. Early footage “told us a lot about what was happening, but it really wasn’t that different from what a police officer standing on the street could see themselves,” McKenna said. “That’s how the law in the US reached the conclusion that CCTV systems are okay — because it happens in a public space, so there’s no real reasonable expectation of privacy risk under the Fourth Amendment, which protects against unreasonable searches and seizures.”

But cameras reach much farther than they used to — they can tilt, pan, or zoom, and can often see several miles away. They can be equipped with thermal imaging devices and facial recognition technology, all of which may be accessible to law enforcement. Some AI software can even analyze people’s facial expressions and claim to predict a person’s behavior, McKenna said.

“We have increasingly advancing AI systems with analytical capabilities that can merge so much data and detect things from the footage that before we wouldn’t know,” McKenna said. “An officer on the street wouldn’t be able to identify every person walking by, but facial recognition technology software is very common, and it can be utilized together with the footage that is being taken and collected by CCTV systems.”

All of this information can be sent to federal fusion centers, where information is shared between local law enforcement and federal agencies like ICE and the FBI. McKenna explained that when there’s more information-sharing between local law enforcement and federal security agencies, “we lose control over how that information is used.”

“That’s part of the protection we’re supposed to have under our laws — that information that’s been collected for national security purposes not be used for domestic law enforcement purposes,” McKenna said. “We have increasingly seen a blurring of national security measures becoming part of domestic law enforcement.”

There are still a couple weeks left in the World Cup. But there’s no telling how long all the surveillance data gathered around the matches will be stored, or how it will be used.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Gaby Del ValleClose

  Gaby Del Valle

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Gaby Del Valle

• PolicyClose

  Policy

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Policy

• PrivacyClose

  Privacy

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Privacy

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

YouTuber Jon Prosser has finally filed a formal response to Apple’s lawsuit made against him and another defendant over allegedly stealing iOS secrets. In his response, Prosser denied that he “planned or participated in any conspiracy or coordinated scheme” for the “purpose of injuring Apple.” However, Prosser admitted to recording a FaceTime call showing unreleased iOS software and sharing revenue from his YouTube videos about the leaks with the person who showed him the information. Prosser also argued that the other defendant “is completely responsible” for the alleged disclosure of trade secrets.

Last July, Apple claimed in its lawsuit that Prosser and Michael Ramacciotti had a “coordinated scheme to break into an Apple development iPhone, steal Apple’s trade secrets, and profit from the theft.” Ramacciotti, Apple alleged, showed Prosser information about iOS from a development iPhone belonging to an Apple employee, Ethan Lipnik, in a FaceTime call. Over the course of a few months in 2025, Prosser hosted three videos on his Front Page Tech YouTube channel detailing an unannounced iOS 19 with elements resembling the Liquid Glass interface that Apple would go on to demonstrate at WWDC as part of its renamed iOS 26 operating system.

Apple didn’t immediately reply to a request for comment. When asked about the case earlier this year, Apple spokesperson Jacqueline Roy told The Verge that “We don’t comment on active litigation.”

Months after Apple filed the lawsuit, a clerk entered a default against Prosser after he hadn’t formally responded to the complaint, meaning that the case could go on without him. But Prosser said in an April filing that he would finally be retaining counsel. In June, Prosser agreed to sit for a deposition, and soon after, Prosser and Apple asked the judge in the case to set aside the entry of default. He did, and gave Prosser 10 days to respond to Apple’s complaint.

Below is Prosser’s full response, if you’d like to read it for yourself.

As the parent of two little girls, I often think about how their childhood is different from mine. The seven-year-old is learning about AI at school. The five-year-old is given internet-based homework every week. And they are both absolutely repulsed by the idea of smoking.

That was not the prevailing sentiment when I was young. Smoking was a central part of our culture. Which is why the UK’s recent passing of a generational sales ban on tobacco products feels like such a big deal.

This is what’s described as an “endgame” approach. While many tobacco control strategies—such as taxation or gory imagery—aim to reduce consumption, policies like the UK’s are designed to eliminate it entirely. It’s a new approach, and no one knows whether it will work. But it’s an enticing prospect—and it’s starting to look a lot less radical.

Find out why generational tobacco bans are gaining support.

This story is from The Checkup, our weekly biotech newsletter. Sign up to receive it in your inbox every Thursday.

You do your own time

—You do your own time is a short story by Elizabeth Bear, an award-winning speculative fiction author.

There we were, a regular murderers’ row of librarians. Turning around in the nave of our library to greet the sound of footsteps, pistols leveled in case whoever was coming in didn’t respect sanctuary.

I pulled down a solid-state drive full of biographies and case studies of people who had spent time—and sometimes their whole lives—in labor camps or chattelhood. It was illegal to possess, and the feds used smart agents to track down and obliterate any copies. Which was why we were sending one to the stars.

This is what’s described as an “endgame” approach. While many tobacco control strategies—such as taxation or gory imagery—aim to reduce consumption, policies like the UK’s are designed to eliminate it entirely. It’s a new approach, and no one knows whether it will work.

The Maldives was the first country to implement a generational smoking ban, in November last year. It’s too soon to say how that has panned out.

Nor do we know if these laws will even last. In 2022, New Zealand passed a similar generational sales ban as part of a broader anti-smoking law. But it was never enacted—the law was repealed by a new government in February 2024.

In the UK, both major parties support the ban. But Nigel Farage, whose right-wing party has seen a recent surge in support, has promised that “the generational smoking ban will not last long if Reform gets the chance to start rebuilding our mismanaged country.”

Chris Bostic, an attorney and former policy director for the advocacy group Action on Smoking and Health, says he and his colleagues began promoting the idea of a generational ban in the United States 11 years ago. Back then, they struggled to win support, even from major health charities. “People said we were crazy … [and] that this was impossible,” he says. Opponents argued that bans would infringe on personal freedoms.

“The public health argument is: Well, what about freedom from addiction?” says Britta Matthes, a tobacco control researcher at the University of Bath in the UK. Most people who smoke began when they were teenagers, want to quit, and wish they’d never started. Tobacco is arguably the most harmful consumer product of all time. It will kill half its users who don’t quit, according to the World Health Organization.

It also kills people who don’t smoke. Of the 7 million who die from tobacco every year, 1.6 million are nonsmokers who were exposed to secondhand smoke, according to the WHO.

To be eligible, companies must have a minimum of 5 IT employees and a minimum of 100 total employees. We consider IT employees to be those IT workers who provide technology support and services to their own company — or to multiple companies through their work at an IT service provider. Workers who would *not* be included are administrative support staff for the IT department, staff who work in communications or PR for the technology department, IT contractors, or those staff whose primary role is in product development for outside sales.

Best Places to Work in IT is a global program. We ask that companies submit no more than one survey within any one country. If your company operates in multiple countries and you would like to submit a survey for your location only, please note this in the company name field (e.g., “Foundry North America” or “Foundry Germany”). If no location is specified in the company name, we will assume that the entry represents all locations worldwide.

In most cases, we prefer to have the parent company, rather than subsidiaries or affiliates, apply for the Best Places to Work in IT list. However, a subsidiary or affiliate may be eligible, providing that it stands out as a separate entity from the parent company, with separate business functions, IT leadership and so on. A subsidiary may also be eligible to apply separately if its parent company is a holding company. In those cases, the parent company and subsidiary may be able to apply separately. We encourage companies to complete the nomination form or contact us at bestplaces@computerworld.com, and our Best Places research team will evaluate the submissions on a case-by-case basis.

Frontier transformation

Microsoft Frontier Company will place 6,000 experts with customers to “co-design, co-innovate, deploy and continuously improve” AI systems based on their specific business goals, Judson Althoff, CEO of Microsoft Commercial Business, wrote in a blog post.

The new offering focuses on what Microsoft calls “Frontier Transformation,” helping customers build an intelligence platform based on their proprietary data and internal expertise, workflows, and decision-making processes. Based on FinOps principles, the offering helps users “observe, govern, manage, and secure” AI tools across their stacks, and their intelligence compounds over time, Althoff said.

Microsoft Frontier Company is a “model-diverse, open, heterogeneous” platform, Althoff noted; customers can choose their own models: ChatGPT, Claude, Microsoft Copilot, or other open source or industry-specific models.

The video game disc is dead, and Sony’s been planning to kill it for some time, according to a report out of Austria. The man who leads Sony’s discmaking operations, Sony DADC president Dietmar Tanzer, told ORF Salzberg that the company’s Thalgau plant produces 600,000 discs every day, half of which are for PlayStation. But since it’ll only be making 10 percent of that volume in 2028, it’s planning to retrain all 300 employees to work on optical microlenses instead.

Thalgau isn’t just one of Sony’s disc plants. It’s where the disc-making division is headquartered, and appears to be its only remaining wholly owned disc manufacturing facility. Sony made discs in the United States for decades, originally in Terre Haute, Indiana and later in New Jersey, but it closed the latter plant in 2011 and moved all manufacturing from Indiana to Thalgau in 2022. Today, the Indiana facility markets itself to automakers who need help packaging and assembling headlights and the like instead.

This transition didn’t happen overnight. A behind-the-scenes video from December 2024 shows that the Thalgau plant was already working on microlenses as of then:

Those lenses, too, are created using discs:

ORF Salzberg writes that Sony has now invested €30 million to manufacture these microlenses, and that mass production may begin “as early as next year.”

Microlenses are theoretically used in all kinds of emerging applications where you might want to bend light, including headsets, but it appears that Sony may cater to automakers here, too. The head of Sony’s micro optics division gave ORF Salzberg the example of “a car turn signal that is projected onto asphalt.”

All of this is to say: Sony didn’t make this decision in a hurry, and it isn’t likely to change its mind despite the predictable backlash. It’s been winding down disc manufacturing for decades, and it’s ripping off one last band-aid with PlayStation.

According to Sony DADC’s website, it has produced over 26.4 billion discs to date — the vast majority, 23 billion of them, were made between 1983 and 2022 in Terre Haute, Indiana.

We’ve got a special Decoder today — I had the chance to talk with Amy Lanzi, the CEO of Digitas North America, in front of a live audience at the Uber Villa at the Cannes Lions advertising festival in the South of France. I know, it’s a hard gig, but I do it for you.

Amy has been on Decoder three times now, and she’s one of my favorite people to chat with — she is clear-eyed about what the advertising industry really is and does for brands and what all the money sloshing around the ad-supported internet really accomplishes. You’ll hear her say that she thinks the traditional chief marketing officer role is done for and that her job is driving business results using data and analytics.

Verge subscribers, don’t forget you get exclusive access to ad-free Decoder wherever you get your podcasts. Head here. Not a subscriber? You can sign up here.

That might sound straightforward, but it was a shocking statement at Cannes, which is where the entire advertising industry gathers every year, drinks rosé, and convinces itself of outrageous nonsense. This year, the big trends were creators, and, of course, AI. And Amy, alongside her parent company Publicis, aren’t holding back when it comes to calling out the AI nonsense for what it is.

Publicis actually put out an ad before Cannes listing all the false promises being made about AI when it comes to advertising, so I asked Amy about that, and what AI might actually be good for, beyond just generating slop and slop headlines. After all, Meta and the rest of the big platforms were all at Cannes talking about generating more and more ads with AI — something that threatens almost every other company in the industry.

Of course, we also talked about the creator economy and how all the creators at Cannes were openly calling themselves marketers — essentially turning themselves into small ad agencies of their own. On top of that, the biggest creators in the world almost all launch their own products — something Amy and Digitas see as an opportunity, as those companies will need operational scale and excellence if they’re going to be successful over time.

There’s a lot in this one — like I said, Amy is as sharp as they come, and I really enjoy talking to her about how the money really works.

Okay: Amy Lanzi, CEO of Digitas North America, live at Cannes. Here we go.

This interview has been lightly edited for length and clarity.

Hello everybody. I’m Nilay Patel. I’m the editor in chief of The Verge and the host of Decoder.

Hi. Amy Lanzi, CEO of Digitas, part of Publicis Groupe.

I’m very excited to be here with Amy. Amy, we have to stop meeting like this. We only talk during Decoder podcasts.

I know. We do. We have to do this more. We both live in New York. We should get together more IRL.

Yeah. We only travel long ways to talk to each other in front of live audiences, which is wonderful. Thank you all for being here. Thanks to Uber for having us. I have a lot to talk to you about. It seems like every year we hang out and the entire advertising industry is in a new form of chaos. And that chaos doesn’t really get resolved. We just move on to the next form of chaos. This year, it’s AI.

I don’t think that’s a surprise to anyone in this room. Everywhere here at Cannes, the conversation is about AI. Let’s just get started.

Publicis put out what you would call a fake ad, a documentary ad. It’s called “The Wrong Promises” and it is basically just a series of vignettes of promises people are making in pitches, including “you don’t have to pay us until you win a Lion,” which is incredible. And it says “this is real” at the bottom.

Tell me about what you are hearing in these rooms that are leading to these crazy AI promises getting made.

Thank you for bringing up our fun video. It was really designed to stimulate conversations like this. In the business right now, it’s crazy. I’ve been in this business for a long time and we are seeing all kinds of partners offer wild things in the pitch process.

It’s different than it’s been before, whether it’s all the things that were, of course, hyped in that video, but also just insane commercial deals that are just generally bad for people and the business. And they’re coming in all different types — whether it’s about free AI, free platform, free whatever — that are creating a dynamic that is not good for all of us in this industry. Because we all need to work together. It is a people business and all of those things really, long-term, create a people problem.

Not to get all Toy Story 5 here, but the conversation that you and I are constantly having is about the pressure of the tech platforms on the media ecosystem, whether that’s publishers, whether that’s agencies, whether that’s creators at some point. We’ll come to that.

But the idea that the platforms have enough scale to promise you business results and then deliver them — whether or not that happens, but they certainly can make those promises — is leading to some of these outcomes. And some of these promises about AI and what it might be able to do, is there any reality to that? Or is that just a reaction to the pressure the platforms are putting on the ecosystem?

Whenever this conversation comes about, the promise of AI, I always go back to the promise of programmatic. How many of you remember when programmatic was a thing? No more people, it was all going to magically—

No one in this room admitted that they remembered that programmatic advertising was a thing.

Yes, right, exactly. And that was a time when it was all just going to magically happen and it magically still needs people, still has the nuance of brands and the marketplace and all of the things that we do to define our partnerships and what are going to be game changers. I go back to that because I feel like the AI story is the new programmatic story, with the promise of everything just being absolutely automated, and that absolutely did not happen.

When programmatic was rising, there were all those promises at that time as well and now we are living that. But what is different here is it’s coming from either agency partners or tech and platform partners. It creates a different chaos to where you started.

Publicis and Digitas also have huge investments here. You were early, right? You bought Influential in 2024 to do analysis of how the creator ecosystem was doing. You have Digitas AI, and it’s been two years that you’ve been into that. How do you think about those products and those platforms in an ecosystem that is full of these promises?

For me — and I’ll talk about Digitas AI in particular — we started to make our people better unicorns, as we talked about the last time we met. It was really to say, “Okay, everyone look at what’s on your desk, what’s in your day, and think about what you absolutely could build an agent to do and that way we free up our time to do this.” That’s where we started. And then what was fascinating is the magical things that were able to be built by the young talent in the industry that is solving a working problem but eventually a business problem.

That still holds because every day, as a Gen Xer, I always say, “Hackers wanted.” If you have a hacking mentality and you’re curious, you can actually do better things than I did when I was a hacker in that age. It enables us to use our agents, use our data to get to better ideas, better workflows that are more of a surprise and delight to clients than what you might have brought in the first round. Because you can do many rounds before you actually get to the final product and that’s how you get to the unique outcome.

You can tell that Amy is a Decoder pro because she has led directly to the Decoder questions. I’m very curious if AI, at least in the enterprise context, is a top-down or bottom-up change agent. You just restructured Digitas, right? You put in a bunch of new roles. You have a new chief intelligence officer, a new chief systems officer, and my favorite, a chief transformation officer. He turns into a robot. What are these roles?

[Laughs] I hope he hears this because that’s going to make him laugh. Anyway.

It’s very exciting. The toy line is going to be great. What are these roles for? How are you changing the structure of Digitas?

You’ve always inspired me by asking “how do you structure yourself” and “how do you make decisions?” The first time we talked about it, I had been in this role for just a little bit of time and we had set our culture, which is, “We’re fearless, inventive and generous.” That still holds on how we make decisions.

Org structure-wise, we needed to figure out exactly what you just said: how do we do things in the underpinning of the agency that will enable all of our different practices to be able to scale faster? To leverage intelligence — and, of course, intelligence is code for AI in this story. We are moving — so that’s why we have our chief intelligence officer, which was a pickup from data and analytics — into liberating that into an intelligent platform that all of our agents are built on and then also all of our employees use.

The second is systems thinking. We took the traditional COO role, turned that into a chief systems officer role and really redefined both of those roles because all of our client conversations are really about marketing systems. You need to have someone who can come in and understand that and meet them where they are. Then we think about the marketing flows that we’ll marry to the system that works within some of their tech choices, et cetera, as two examples. The third, the chief transformation officer is — and this is a big conversation here — around clients wanting to transform, particularly CMOs transforming into CGOs.

I’m a journalist. My job is not to innovate. It’s unfortunate.

I just ask the questions.

That’s why we’re good friends. I make up a lot of things.

I was just coming from the Forbes CMO Council thing and there was a whole conversation on the rise of the CGO. What that means is CMOs, formerly, were really responsible for making marketing magic. Campaigns and media investment. Now they’re responsible for building capabilities and they need a partner to help transform their solutions, so it’s a people problem as well as a tech problem. That’s why we have that discreet chief transformation officer role.

You are a data person, you’ve always been a data person. I’m really curious about that. The idea that the marketing function is now so directly tied to business results, I see it on the tech platforms all the time. You and I have talked about this before.

Mark Zuckerberg will just look every single one of you in the eyes and say, “I’m going to kill you.” He does it without remorse. His vision is that you will just pay Meta and it will make creative for you, it will find the audience for you and they will deliver you business results and almost everything in the middle, from money in to money out, gets automated away by Meta’s intelligence stack.

The idea of the chief growth officer feels of a piece with that. That the marketing function is now a business function and it needs a data layer, it needs all these things. Where do you see the creative fitting into it? Where do you see the classic form of brand building?

It’s a fundamental part of that capability that needs to live, like the marketing capability now, as a system. Brand building and the fundamentals matter more than anything. What you stand for and how you show up, what you’re the answer for, especially with LLMs, is really important and it has to be authentic for the brands to grow.

That’s still the case. You just can’t stop there. It needs to be attached to a system and ideally, you have some data intelligence layer so I’m learning more and more about you and I can then redefine my marketing stories to reshape how I’m showing up if it’s not working, if it’s not selling more candy, if it’s not selling more cars.

How we change faster versus the CMO being responsible for building these campaigns and then moving on — that role is a dying role and it should be. It should be that marketing drives commercial value, marketing drives shareholder value.

You heard it here first, CMO is a dying role.

The previous version of the CMO. Sure. Just like the whole, we are in a chaotic world. We’re in an area of convergence. If you are a CMO and you’re not thinking about your data layer, you’re not thinking about your full stacks, [you’re not thinking about how] you can get closer to the consumer and get to better creative, better use of creators, you are not going to be doing well in the future. Whether it’s CMO 2.0 or you’re moving into that CGO role, that’s really the future of that type of function in an organization.

The audience here at Cannes knows this, but for the Decoder audience in their cars listening to this, the conventional wisdom was that CMOs only ever had two years on the job, maybe less. You were dead the moment you showed up, you got to execute one big campaign, you found the bathroom and then you were out.

I hope you heard everyone in this room laugh.

That was the most knowing laugh in the world. This seems like a much longer-term role. If you show up as the new CMO and your job is to transform the entire business from the bottom up over and over again, it’s a lot of instability. Do you think these are longer term roles?

If you do this right and you create a marketing system, you will be able to live beyond the two years path, because you’re creating a growth engine. If you only want to do two tent pole marketing campaigns, that’s probably not going to go well after two years.

I basically have two themes that I want to talk about. One is data in scale and one is creators. This sits right in the middle of them. And I just keep thinking about this phrase that I’ve heard Meta use the most, but the other platforms use it a lot too, and it’s “creative is the new targeting.”

I hear that and I don’t make ads. We make two podcasts a week. We don’t do scale at that level, but I hear “creative is the new targeting” and I think that is a demand for output like no one has ever heard before — “We’re going to find your customer for you; you just feed the machine with as much as you can.”

It seems obvious to me that you have two choices in that world of “creative is new targeting.” Meta’s Andromeda system is the biggest example of this, and it’s that you’re going to feed it with AI creative and you’re going to let them do it, or you’re going to have your own AI creative system or you’re going to let a million creators and influencers do it for you. How do you plug a growth engine into a world where that’s what the platforms are all saying?

The closer you are to your consumers and all the places they’re spending time from a data identity, however you’re systematically doing that, is the most important thing so you can see how they are reacting to… what did you say?

Creative is new targeting.

I’ve never heard this phrase. I don’t like it.

Yeah. I figured you wouldn’t.

It’s terrible, right? It’s terrible for our business — “creative is targeting.” Only because it sounds horrible. Creative is emotional and lovely and magical. How can it be emotional and magical if it’s targeting you? Those are two words that don’t go together.

The brands that are doing really well are deciding not to do that or to do it in a way that feels so contextually relevant because they know so much about Gen Z on [Meta’s products], that they’ve decided that they need to do this to be able to drive velocity for a certain type of brand that feels fit for that brand.

But we are not in a world where we need more content. Does anyone want more content? We are not in a world where we need more content. That idea, creative targeting, assumes that we have endless amounts of content and impressions as a concept. Those are two worlds that still need to be decoupled, but when they’re put together, it needs to be in a way that you are constantly learning. “How did that do? Did I gain more attention from a certain type of growth audience I’m trying to get to?”

Can I push you on this? Not to support Meta’s ideas here.

Let me just make the argument because you and I have talked about it before. The way, in particular, Mark Zuckerberg talks about it is at some point they know more about the viewer of Instagram than anybody, or they want to know more about the viewer of Instagram than anybody, and they will just deliver the right creative to that person at the right moment.

This is the same argument as programmatic. Only now they have acres of GPUs with which to do it. The argument there is yes, the army of unpaid teenagers will make the fun stuff that you watch and the advertising will just show up magically in the platform and find the right person at the right time.

Again, I read that as, “You should just let Meta make the creative for you.” You just give them a picture of the shampoo bottle and they will put it in the right place, on the right background. For me, it will be pictures of trucks jumping over things and exploding and then shampoo will appear at the end. They’re like, “That’s shampoo for me.” It would work. If any of you need an idea to target me specifically, that would work.

No agency is ever going to make something that niche, but for example, Publicis just bought LiveRamp to take all that data out—

It’s still in the situation.

Yeah. I mean, it’s out there, but the deal is not done.

Presumably the goal of buying LiveRamp, which is a giant data platform, is to lever up against Meta and say, “Actually, we know as much about these customers as you do. We can distill information across all the platforms and we will actually do a better job for our clients of placing the creative in the right place at the right time.” This implies that Meta will participate. Why would they do that?

Because people are complicated. And today, look at how often on this beach there are different players on the beach. Everyone doesn’t stay in this position forever and people are complicated. I would offer that if I started seeing your wonderful ad 50 times, I would say, “I don’t want to see this anymore and I actually don’t like that shampoo.”

That’s what happens when it’s just left to go on its own without someone thinking about suppressing messaging, without understanding consumers better because you actually win hearts and minds when you’re doing something that’s not as expected. Because people are complicated.

It’s like we say to our clients, “You need to work with all of them because all of us spend time in all these different places for different reasons. Why does someone choose to spend time on Pinterest versus [a Meta product] versus Amazon versus Walmart Connect? All of them are different roles, in terms of how the platforms work for you and your life. So why would you put all of your dollars in one place?”

For me, those players are shifting in and out depending on what we think that growth audience is doing and how they’re thinking about intersecting with that category.

Are they good at participating in an ecosystem in that way? This is something that strikes me every time.

Some of them are and some of them are not. The more they start to see headwinds, they lean more into participating. Ideally we want to do data collaboration with these partners because it’s better for our clients that we can say, “We knew more about this growth audience on the platform — I’m not going to say X because I’m not saying X —- on platform Y. We know more about them and this is—”

No one knows anything about X.

Right. “And this is how we’re going to grow your brand.” That’s a good joint story for a platform and an agency partner to come in to talk about because we’ve been able to actually share data and look at that in a way that’s beneficial for the consumer in the end.

There are only two business models, right? Bundling and unbundling. In the agency world, there’s a lot of bundling going on right now, right?

There’s a lot of bundling going on.

Obviously Publicis is huge. Omnicom just bought IPG. CAA is out here saying they’ve invented creators this year. I don’t know if anyone heard. They’re very proud of themselves, but they have a $250 million fund to take stakes indirectly in creator businesses.

There’s a lot of scale. There’s a lot of bundling going on. What do you think is driving that?

It’s the desired state of a system. Hence why we have a chief systems officer and there are capabilities that we built. Last time or two we met, I talked about building the commerce capability and the acquisition of Profitero that I led. That was really to get to a unique set of data so we could build a better commerce capability and advise our clients on how they should be investing in some of these retail media partner spaces at the time. How to win the digital shelf was really the business problem to solve when we made that acquisition.

That holds true for the acquisition of Influential. How do we create “creator” as a channel? How do we harness the creator in a way that’s also easier for clients? If you don’t bring those things into a bundled offer, it’s actually really hard for clients because there are seven different versions of Profitero calling on them. There are seven different Influential partners, or those types of partners that are calling on them, and then you’re looking at a decentralized set of technology, and it’s, again, ideally not tied into a single data identity spine.

The more you have them connected, the more you can actually say, “Did this really work? Did it really drive someone to buy more things?” That’s the reason why the capabilities are now being bundled. Think about them as capabilities versus assets in this conversation.

Talk more about Influential. That was a big acquisition two years ago. I’m assuming it’s been integrated. That was the big creator play for you. I know you were instrumental in that acquisition.

It feels like the creator economy has taken on a new shape this year. We are hearing more and more about the biggest creators in the world and the biggest platforms. Influential was much more of a scale play, right? It had an idea that you could go and do a lot of analysis of what was happening on these platforms at massive scale. How’s that working out?

It’s working out very well because we have taken their technology and connected it to our identity so that I can now see you, I can see who you follow, and then I can understand how to get closer to you or have either the type of right creator choice or right creative decisions that can get you to feel an emotional connection to my brand.

It’s also really that the creator as a channel is also super efficient. You can also have creators talk in a way that is a little more authentic to the audiences, where a brand maybe can’t go out with that message. But it’s very complicated. Before we had Influential, it was very hard. Every different brand partner had a different creator solution. We still are interoperable. We work with other ones that are out there like CreatorIQ, but it is easier for us if it’s connected into our system because then we can get to see if it actually worked faster.

What’s the step after the creator interaction? You open one of these platforms, you scroll some videos, you see some creator telling you to buy some product. Is it then to go search? Is it then to go open a retail app directly?

You mean for the consumer?

I mean, it depends. One of the hottest topics we’re getting right now is TikTok shops, which is another interesting topic we can have another day and that’s because creators are saying, “Don’t you want to buy this thing?” It’s a one-click opportunity to go get it.

For us, when some brands that are more used to selling through traditional retail channels, it’s a pretty big lift to build that out. It still is. But the desired state is, “I want to buy your kit of parts, whatever you’re talking to me about in your creator video, and I want to be able to buy it and do it then.” That’s a move we’re starting to see. We also just see “follow me and learn more about me as a creator.” It’s actually a cheat for CRM because if I’m a brand and you’re one of my power influencers and I want to really have an owned relationship with you, this is a way for us to get closer to you as a brand.

The reason I ask that question is I’ve run around the past two or three years asking every web host and web company why anyone should make a website now. If I was a new creative and I was starting The Verge over again in 2026, I would not make a giant fancy desktop website. I would probably be a TikToker. My dancing is pretty good. It’s not great, but I would probably start there.

The answer from all the web companies I’ve gotten is, “Oh, because you want to do retail and you want to escape fees on platforms. You want a different fee structure.” The best answer I ever got was from the CEO of Squarespace, who said, “A lot of people start websites with us just for the appearance of legitimacy.” Having the dot com — or I guess the dot ai now — gives you an appearance of legitimacy that a customer will go and transact against versus just dropping a TikTok shop and who knows what sort of Temu product you’re going to get.

I look at that and I say, “Okay, you’re going to go to the web.” That’s a move everybody wants you to do and the web is fundamentally under threat by agentic search, by AI overviews — everything everyone knows about. Do you see that changing in your funnels?

Interestingly, we are seeing more and more briefs for revising websites.

To do generative engine optimization (GEO)?

Yes and no. It’s actually pretty interesting. This is something that the market talks to you, clients come in, they look for new things. Who would have thought, if you asked me at the end of last year, “Are you going to do a lot of website work?” I would have said no. We’ve seen a lot of it for two reasons. One, the future of websites is knowledge sites. Websites were built to contain a bunch of random information that didn’t really enable you to show up well in a single response to an LLM’s prompt.

So part of it is that, but also if you’re going to be a real brand and it’s literally your legitimacy point, you really need to have your own assets. You want to own that and eventually you don’t always want to have to funnel everything like packaged goods through a retail partner. You want to be able to have your own relationship. That doesn’t mean everyone’s going to buy from that site, but the idea is that it is a site that, one, does its job of feeding the LLMs but also builds your brand and gives information, too. Old websites pretty much weren’t really built in a way that makes sense for modern consumers.

Who are you giving information to? Is it consumers? Is it AI agents?

No. I mean, yes. If you have to build, we build for machines and humans.

What do you think happens? Every time I go to Google I/O, I sit down with CEO Sundar Pichai and I ask him. Every single demo of search for years now has ended in a transaction. I don’t know if you have noticed this.

Every Apple demo ends with trip planning. Every Google demo ends with trip planning that ends in buying shoes. You realize the audience is partially Verge readers who just love tech and partially it’s other developers. And then a huge part of it is industry analysts and their investors and they want to see the revenue return on every single feature that Google announces and so they all end in transactions.

Over time, those transactions have taken place closer and closer to Google’s core platform. This year they announced intelligent shopping carts that are just going to track prices and execute a purchase for you. You’re talking about rebuilding websites. There’s a world in which the website is just the backend database for a very complicated Google system that is doing all of the things that you’re talking about. Is there a hedge against that or is that train just on the tracks?

It’s a little bit of a hedge against that. I would say that for many of our packaged goods friends, one of their biggest threats is Amazon. That is the Google in this story because if you are so dependent on a partner like Amazon that also can — and we like our partnership with Amazon — but they also are a place where, because of the size of their marketplace, if you’ve built a brand that has a story, is a known brand, some of these marketplace brands can immediately win because of how you’ve invested and built the category.

So you need to have your own assets to be able to think about what is your next story, what is your next move so that you don’t get commoditized. That applies to both Amazon and Google, to a degree.

Do you think they’re competitors in this space in that way?

I don’t know. I don’t do any shopping on Google, so I don’t know if I would compare them to that.

Have you ever planned a hiking trip on Google? They’re great at it.

[Laughs] Actually, no. I use this thing called Mindtrip. It’s my favorite. I learned about it in The New York Times. You go to Mindtrip and it’ll magically plan your trip. It’s pretty good. It can do most things and then it breaks in certain areas.

But this idea that you’re going to have a bunch of agents just planning the trips for you, buying all of the gear that you might want, or tracking prices for you, which is another big promise that everyone has. And then the brands might pay for preferred results in all of those spaces. That does seem like a more direct competitor to Amazon than not, right?

Absolutely, it does. I just think there’s the planning part and the transaction part. Those are two different things. I think on the planning, you’re competing. Amazon is trying to get up. They’re going to get added transaction into that planning component — when you think about all the things that they’re doing with their own creators and even just their move on full funnel and content, that’s really to make Amazon your total destination. But I think in the agent-to-agent buying space, the concept of trust is still very important. What you’re willing to do and not to do, I think as a consumer is still TBD.

Here’s my wonkiest tech question for you. Google is very, very proud —

[Laughs] This is hard. Oh, Lord.

I got it. It’s fine. Because again, is AI top-down or bottom-up? I understand how it’s bottom-up. We’re going to empower all of our smartest people to do agentic things in our workplaces. And then there’s top-down, which is that Google is very proud of its universal cart standard, where agents are going to be able to access carts across domains and compare prices and do all that stuff.

That’s a corporate decision. That’s a strategy decision and they’re pretty proud that everyone’s on board. Their argument to me is, “Well, we just needed something. And this is the something.” I hear the tech industry say, “We just needed something.” And I think, well, of course, but it’s this something that you control and everyone might need something, but no one is thinking it through.

At the end of this, one player will control it. When your clients and partners think about, “We’re going to adopt these standards for agentic workflows, for how commerce will be done in the future,” is there any pause to say, “Google’s going to control the standards again?”

Yes and no. The clients I speak with today are not as worried about Google as a commerce partner as they are about the transformation of their more traditional partners, like Amazon or other retailers, that have transformed from being a retail partner to being media partners.

That transformation is occurring over there, their media partners —

It’s changed this dynamic of what a retailer does and what a supplier or manufactured brand does. That dynamic has changed and that makes the negotiation between them such that they need each other to grow. When they’ve transformed into media partners, this changes that equation significantly. This is the whole retail media conversation. “Are they a media partner or are they our retail partner?”

That’s still much more the conversation.

Has AI changed the dynamic of that at all?

Yes, because the more AI-oriented products they have, the more it puts focus on media investments with them because of these magical products. It’s not necessarily solving a core shopping problem.

So let me flip to the other side of this, because so far we’ve talked about tech and data and scale, and I think that is the promise of AI. “We’re going to automate everything. It is going to be magic. We’re going to do programmatic again.” All you need is data and they need your money.

The other side of this — and I’m just walking around here at Cannes — is creators everywhere. Something vastly more human, vastly more bespoke. My theory — and I was just wrong on this, so I’ll cop to it — I thought creator rates would fall drastically this year, as there was more AI content, as the platforms got more and more punishing about what would work on their algorithms, as they started doing their own content.

Instead, I look around and everyone here at Cannes is just so proud about how the cost per minute of video is skyrocketing again. What’s going on?

Again, it’s all about supply and demand. Once creators start moving into a more famous space, then suddenly they’re worth more. Last time we talked about this, we talked about the difference between a creator, an influencer, and a celebrity. I still think that holds true.

Unilever went out, probably this time last year, saying that they were going to spend 50 percent of their dollars with all creators versus creative. To me, an interesting conversation is, “What is creative now?” If you’re putting it all in the hands of a creator, what does it mean to be a brand that delivers creative and how do you make those things work together?

Because we can’t have another thing that also prices us out of the market and you get to a point where you’re ceding so much control to the voice of the creator that that’s probably not going to build your brand over time.

The creators see themselves as marketers, at least the very best ones here. I mean they’re here at Cannes.

They’re all small media businesses of their own. Some of them are very large media businesses of their own and they’re very open that what they do is marketing for brands. That comes right up against agencies. It comes right up against your creatives. All the creators know they want more control. They want the brief from the brand. They want business results. And they say, “Leave me alone to do it. I’m just going to do it however I want to do it.” What’s that dynamic like as you start working with the bigger and bigger creators?

We still hold true to brand fundamentals. Strategy matters. What the consumer insight is for that brand has to be something that’s an exchange between a brand and a creator in order for the creator’s work to really build what’s unique about that brand. It’s still the idea of freedom and a framework that that creator is briefed on. They control how they’re going to do it, but there are some tone and voice and things that, if a creator’s actually going to buy into this, they need to buy into that. That still is really the exchange. That’s always been the exchange. Do you remember when UGC started in Tongal? Do you remember that whole thing?

It was the same thing. I was working on Johnson’s Baby and we were trying to figure out how to really create authentic first-time expectant mom stories. We had this great debate about how we’re going to get the new creators — like mom bloggers, was the language — to get them to tell these stories. So we had this: this is what the brand stands for, here are some things, but you need to buy into this and believe in it. That exchange still exists. We just call them creators now. And then you can do it in a more scaled way, through a system like Influential.

I guess maybe my new prediction is that we’ll see a more barbell shaped creator economy, where the biggest creators are making more and more money than ever and the middle gets squeezed. Either you’ve got to get big, you’ve got to be able to command rates because you’re a celebrity, or you end up towards those UGC rates, where you’re just doing infomercials on TikTok and that’s the business. What’s left for the middle?

It’s a really good question. I don’t know the answer to that. I do think there’s so much demand for this space and there’s a lot of demand for the middle. So you could use it as a way to eventually launch yourself into this upper echelon of creators. Someone said to me that creators are daytime stars from soap operas, which I thought, “This is pretty interesting.”

Because they’re in your house?

Because they’re just so on. It’s not noise but they keep you company, they keep you busy and then all of a sudden, one of those people becomes a movie star. Which, to me, is a little bit of that middle of what you just articulated, where we know the creators. I’ve watched my daughter over the years, and I’ve watched all of her creators [grow]. Emma Chamberlain was a creator on YouTube who’s now a star, doing Estee Lauder ads or whatever she’s doing. She’s really a celebrity now, but she started as a tiny creator making adorable shows on YouTube.

One of the things that you and I were briefly talking about before we came onstage was at some point creators hit a scale where their businesses need to be businesses. That’s where you might need an agency to come in and actually operate your business. That scale is still pretty huge. It might just be MrBeast at that scale, and he needs to operate a full business.

But I see lots of creators. I think it’s fascinating that the creative economy is the only one where people routinely transition from selling bits to selling atoms. It’s crazy to me that the Paul brothers are like, “Bottled water is a good business to be in.” Historically not a great business to be in. Shipping water around the world is vastly more expensive than just making one more video, but there’s no business in making one more video that scales.

You have to make physical products and sell them to people and that scales. Is that becoming a bigger and bigger line for you? “We’re going to go find the biggest creators and make them products and scale that up”?

The last time or two we met, I had predicted that, at some point, there are going to be these creators that become enterprises, let’s say a MrBeast that actually needs an agency, and it has happened. We are now starting to see briefs from these scaled, started-as-a-creator businesses that have now turned into enterprise businesses, where they need an agency partner and they also need someone to help them to figure out how to invest their media so they can grow. They’re like real brands.

MrBeast himself cannot flip over the bars at Walmart.

I just think people need consultants. You need someone to help you.

No one has ever looked me in the eye and said, “People need consultants,” before.

[Laughs] I call myself a consultant. Look me in the eye. Yes.

I have here Accenture just bought Whalar, which is like an all time nonsense Decoder phrase, but okay.

Yes. You, as an enterprise owner or as an owner of a company, you need help to figure out how you’re going to get to your next growth path. This is what’s happening with some of these businesses.

These creators that have turned into an enterprise who are wondering, “What is next? Gee, it’s expensive to ship water all over the world. What do I do? How do I do this? What types of partners do I go to? How do I figure out ways to grow my brand in unowned ecosystems?” Because that’s the quickest way to grow your brand, not just in your own channels.

There’s this term that I’ve been hearing called the influencer cliff. I want to give credit. The TikToker who came up with it is named Carmen Vicente. I’m only giving her credit because no one gives me credit for Google Zero.

I’m trying to reinitiate karma into this world. It’s her term. It’s great. If you’re listening to this, we’ll put the video in the show notes. But her thesis is that creators can do the branded content because the audiences are there for the real content and they will tolerate the ads in support of the work that they’re there to see. And then when the creator transitions to selling their own product, this misaligns the audience from the creator, and when the brand integrations get too overpowering, it misaligns everybody, and then the creators have to issue apologies and they’ve fallen off the cliff. I see this all over the place.

I see this with tech creators. I see this with fashion creators. Once you start shilling the product directly, your relationship changes, something goes horribly wrong, maybe you backtrack all the way. There’s some class of creators that’s able to do it, that makes the transition to selling products directly. You’ve seen a lot of these now, you’ve seen the briefs. What qualities let you take that next step?

I think a unique product is really important, versus a me-too product that a creator has decided they wanted to sell. That’s important. You’ve seen some creators where it’s that they’re into sports or they’re an athlete, and they’re moving into a space that they feel like this is something that made them a better athlete and is a unique part of the market.

If someone’s following for that, then that makes sense that you would buy into that. When it feels like someone woke up and wanted to just put their persona on a certain brand and it doesn’t feel like it’s been their talk track to you as your friend, then it doesn’t work.

Do you see any particular products that are going to change categories?

One of the things that I think that has been really interesting but is not talked about, is actually a unique example and more of a celebrity example, are brands like Fabletics, Kate Hudson’s brand. It’s super interesting. It’s right in her lifestyle play and her desire to be healthy and all of that. And it was really her launching Fabletics, not her endorsing Fabletics. That’s not really a good creator example. I have to think more about that one.

I’m very curious because I think that’s coming.

As particular rates at the top end of the curve go up, the idea that you can just launch your own brand and collect those margins directly, I get it. It makes rational economic sense. But I’m wondering if the audiences are going to take that at scale.

Even if you just look at traditional brands that are trying to scale DTC-like brands, it is very hard to do because it’s very hard to be found and the distribution part is very expensive. There is a reason why Walmart is incredibly successful, because they’ve solved the route to the consumer. And that is very challenging and expensive to try to build that out, even if the product is incredible.

You mean distribution of product out in the channel?

Because I was going to ask about that. You talked a lot to the web. I feel like I have to ask you because I brought up Google Zero. Is that true in your world as well? It’s definitely true on the publisher side that traffic from Google is going down. Is discovery traffic to all those websites going down as well?

To publisher websites, yes. We are definitely seeing a change in partnerships from our publishers that are trying to understand how they can create custom content that is not as dependent on search, but it’s definitely something that is a big conversation that’s happening here.

I would say in this case, if you are a publisher and you really have an authentic, true voice, people will follow you. They’ll come find you. Versus if you’re trying to fight it out in search, that’s going to be very difficult.

All right. Well, I’ll be in the bar at the Carlton later, if you want to find me. Okay. We’ve got a few minutes left and then we’re going to do some audience questions. But I have a lightning round for you.

[Laughs] Okay. This hasn’t been a lightning round? Oh my gosh.

This hasn’t. No, this is the depth. All right. I have talked to a lot of tech CEOs and AI CEOs over the past year, and there’s a furious debate over whether AI has a marketing problem. Sam Altman will say, “We got to do better marketing. We got to buy a podcast for $200 million.” Great haircuts from those guys. I think about them every single day.

I don’t mean this as an insult. Those are some of the best haircuts in the world. I’m working on it. I don’t know if $200 million of podcast haircuts is going to fix it for OpenAI. Sundar Pichai will tell you it’s not a marketing problem. Satya Nadella will say we have to earn social permission. Do you think AI has a marketing problem?

I think AI is misunderstood. It’s a marketing problem and the problem it’s solving. I’ll say that one of my favorite people to follow who talks about AI — my friends here are going to know what I’m going to say — is Shelly Palmer. If you follow Shelly Palmer, he published something recently about AI being electricity and how electricity is really what fundamentally changed the industrial revolution.

The companies that were fast — and he referenced Ford in this — to really reshape how they work and were able to get to more efficient production as well as speed to market, will win. I think that’s the lost conversation on AI, which is, for me, all of us have too much to do on our computers and our desktops, if you think about it. You can use AI to do those things, and you don’t call me because you’re like, “Hey, did you clean out your email box?” Things like that are not necessarily game changers that are advancing all of us.

Also we have a lot of busywork because all of the work is distributed everywhere. The more we can use AI to connect those things and get those things done enables us to do more things like this. Once we use AI to transform the way we work and to be more efficient in how we’re getting to the outcomes, it’ll then make time in all of our lives to do the things that are going to be game changers. It’s a marketing problem and a misunderstood problem.

Is that going to stop the kids from booing at college graduation ceremonies?

I think so. Right now — and my daughter’s a third-year — they’re worried about their jobs because all they hear, to your point about a press problem or PR problem versus marketing problem, is that AI is going to be the thing that replaces all of them. I would say we want new young talent because if I watch my 12-year-old, he’s been doing Google Slides since he was eight, because of COVID.

All of them got right into Google Classroom and they all learned how to co-work and use many different apps at once and produce presentations easily. Anyone that’s coming out of school we want — we want bright, curious hackers. We’re not not hiring them because we need them. We need people as the lifeblood of our business.

Do you think AI creative is going to get good enough?

I do not. I think the minds and complexity of humans make great stories that you are attracted to. I’m not using the word creative. I’m using stories. I think AI enables you to be able to get to those concepts faster or to get to higher production more efficiently.

I went to an event and it was fantastic. It was a case for Virgin Cruises and they were talking about building these beautiful long-form content stories. The creatives were talking about how they used AI to be able to get to the concepts faster, curate from static art to digital art to really make these beautiful, multimodal boards, to bring this to life faster. But it wasn’t the core of the idea. They had thought, they had used their brains to come up with the idea.

It’s interesting. In our newsroom, we do a lot of first drafts now with AI — of art, not writing — because it lowers the stakes on saying things are good ideas or bad ideas and just that has empowered people to have more ideas. I see a lot of that. I would challenge you: I think at some point the bottom end of the creative, the UGC world, that stuff gets taken over.

For everyday creative, I’ll give you an example: product display pages that are when you search on Walmart or Amazon and that page gives you all this information. If it’s not done with robots, you’re not doing it right because that’s like looking at the back of a packaging.

If you really do it right, you can make it dynamic. It knows what you bought the last time and it can be magical, but that’s like everyday creative that we need in order to get to personalization. We should be using those tools to be able to do that versus someone making those in an analog world, which is where we all started on that. For some things. When I think about big creative and storytelling, it still needs humans.

I’ve asked you about GEO several times. I’m asking you now in lightning round format. Is this just another wild goose chase?

No, I think for certain platforms, it’s a recast of search. We’re seeing this happening, where you need to be, instead of being found, you need to be known for, so that’s the GEO response. But I also think it’s a better curation of information.

I think it’s harder work for marketers because you need to be figuring out how you are leveraging Reddit or all these other different places so that you are showing up in a positive way with the LLMs. But I do think, if we’ve all been talking about the concept of personalization at scale, we should be delivering on that. GEO allows us to do that.

Have you had to change your SEO practice to account for that?

Talk about a practice that you thought was not going to be big. When you look at the capabilities you have and what you don’t think is going to be a high growth engine, our SEO team has exploded. In addition to new websites, this is a big one and that is really a conversation around influence. It’s how you’re showing up and how you’re also influencing LLMs when you actually get into that.

Is that something you can measure? I worry that LLMs are inherently non-deterministic. Every result is different. Do you have a system of measuring those?

We do. We do have an audit format to see what is your share of voice within that, so you can identify where you are in the moment, but it’s a movable feast in comparison to share of voice for search. It is something that we audit so we can start to see what the different levers we need to be pulling are.

By the way, the LLMs are also fed by different things. Recently, one of the new ones is longform YouTube content. That makes sense — that tracks.

Why do you think we all make podcasts now?

Everyone’s just trying to get to three hours of YouTube content. I don’t know if you can have a lightning round question that starts from a place of existential dread, but here we go. I feel like we’re all here.

Did you all see Adam Mosseri’s post about customizing the Instagram algorithm two weeks ago? This is maybe one of the most important posts in the history of media. I’m not kidding. And of course, he issued it as an Instagram carousel, one of the least consumable forms of media that you can possibly have. It’s also on Threads. If your brain is broken and you use Threads, you can go look at it there.

It’s very long and it’s about how offering the ability to customize the algorithm was previously impossible because no one could describe the series of matrix multiplication equations that led to your algorithm showing you a piece of content, except LLMs can now. It can tell you why you’re seeing content in a very direct way. And then you can talk to the LLM and it can customize the algorithm for you.

He said, “This is great and we think this will give people a sense of agency.” And then he went on and said, “Soon I will customize all of Instagram for you and the app will be different for every single person. Some of the experiences will be interactive. Some of the UI controls will literally be different. How we organize everything will be different. And you can see how this is the obvious future for us.”

He went on again and he said, “This might be bad. This might ruin the shared fabric of truth that we all rely on every single day,” which for a Meta executive is remarkably prescient, but he didn’t say he was going to do anything about it. He was just like, “Here’s what I’m going to do to you.”

I look at all of Cannes and I look at all of our talk about the creator economy, about measuring the data and I think, “Adam Mosseri is going to make Instagram different for every single person. That is his stated goal. He said it two weeks ago and he knows that this might distort literally our shared sense of reality, the media experiences we all have.” This is the lightning round question. It was a big buildup to a lightning round question. Is that good?

Okay. Is there anything we can do about that?

How do you not engage with Instagram?

If I told you about my oldest stepson, who is absolutely analog, that’s why. I think we are seeing more and more consumers [like that]. Think about it. If you sat down in front of your TV and it was like, “We’ve made all these decisions about you and you’re only allowed to do this based on…” It feels like Black Mirror.

If that happens to you, you would say, “I’m not going to engage in this anymore. I’m going to go read a book.” It’s also counter to the value proposition of a social platform around community. It was designed to make this connected world. So that’s the most disconnected world.

People are lonely and isolated. The more you’re just in your doomscroll, it will be terrible. We need shared community. One of the things that’s really interesting and a bright spot about creators is creators show up and they have a community, and they listen to their communities. They’re not influencing, they’re talking to their communities, in a way. Brands that do this right are really bringing the community along. It’s why we’re seeing the rise of IRL. People want to get together and talk about things.

Is it just going to be that competitive pressure that keeps them honest? I look at the creator economy and Instagram is a load-bearing part of the creator economy. Disrupting that, maybe you break the whole thing. What keeps them honest? What keeps them from going down this path that they’ve very clearly said they want to go down? Is it just people switching away?

I think it’s people switching away. When brands stop putting media dollars there, and all of those things come together, you might realize, “Wow, they’re not buying what I’m selling and this is not going to work for me.” There are plenty of platforms that are no longer [prominent], if you think about it.

Big blue Facebook. It’s there for my parents. It’s even on their own platform. Okay. Last question for you and then I think we’re going to take some questions. This one’s really easy. What’s next for Digitas? What should we be looking for this year?

What’s next for Digitas? We are very much focused on marketing systems and how we are helping brands build out their growth engines so that it’s not just about your media needs, your creative needs, your CRM needs. It’s how we can really make all of those different capabilities work together so that it’s easier for our brands to grow.

Great. Well, thank you so much, Amy.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Nilay PatelClose

  Nilay Patel

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Nilay Patel

• BusinessClose

  Business

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Business

• CreatorsClose

  Creators

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Creators

• DecoderClose

  Decoder

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Decoder

• PodcastsClose

  Podcasts

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Podcasts

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

• WebClose

  Web

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Web

The company said that it produced a total of 451,758 vehicles between April and June of this year, including 442,936 Model 3 and Model Y vehicles, as well as 8,822 “other vehicles” like the Cybertruck and the Tesla Semi. (The company discontinued the Model S and X earlier this year.) That represents about a 10 percent increase compared to the second quarter of 2025, when the company produced 410,244 vehicles.

Tesla also said that it delivered a total of 480,126 vehicles, including 467,762 Model 3 and Model Y vehicles, as well as 12,364 other vehicles — about a 25 percent increase compared to the second quarter of 2025, when it delivered 384,122 vehicles. (For a direct-to-consumer company like Tesla, deliveries are a proxy for sales.)

The report comes as Tesla is battling a string of negative headlines surrounding the safety of its partially automated driving technology. Earlier this month, a woman was killed in her home after a Tesla driver using Full Self-Driving crashed into it. Tesla blamed the driver, but the crash prompted the National Transportation Safety Board to open an investigation. Several days later, it was reported that Tesla quietly settled a lawsuit stemming from another fatal crash involving FSD.

Meanwhile, Tesla’s robotaxi service is operating at a much slower and smaller scale than Elon Musk originally predicted. The service relies on a fleet of approximately 60 to 70 Model Y vehicles operating in a geofenced area in Austin, Houston, and Dallas.

If our recent Decoder interview with Weber Blackstone CEO Roger Dahle has you craving freshly grilled meats or veggies, Weber just so happens to have a variety of grills, smokers, griddles, and accessories selling at big discounts ahead of the July 4th holiday. It makes a bunch of models, but we’ve landed on a range of grills and griddles at varying price points. Weber is cheaper than other retailers by about $50 on most of these picks, but we’ve noted where you may be able to find them at a comparable price from local hardware stores like Home Depot or Ace.

For occasional cookouts in the backyard, the Spirit E-325 is a basic but reliable option with three burners. It’s available in either propane for $499 (usually $549) or natural gas for $549 (usually $599), a signifcant markdown considering how rarely they go on sale. The grilling surface is 360 square inches, enough space for a few burgers for you and your friends, and there’s a 90 square inch warming rack for toasting buns or for more delicate items that you don’t want to be as charred. This version features two searing zones with higher heat output for finishing steaks, a nice addition over the more basic $449 E-310.

Meta CEO Mark Zuckerberg is all in on AI as the new social media, and he’s previously described a vision of how users could use AI to make interactive experiences and share them with people. The launch of Pocket appears to be one manifestation of that idea, and it follows Meta hiring engineers from a company called Atma Sciences Inc., which made an app called Gizmo, as Business Insider reported in March. Meta got a non-exclusive license to use the company’s technology, and based on screenshots of Pocket on Google Play, the app looks to be pretty similar to Gizmo.

“Scroll a feed of gizmos from people around the world,” Meta says in Pocket’s Google Play description. “Gizmos respond to your touch and the tilt of your phone. They play sound effects and your favorite songs. They can use your camera or pull in photos from your camera roll. Some can even reason about the world around them.” On a help center page, Meta also describes a gizmo as a “playable AI-generated experience,” and when you post one, Meta says you can choose to let other people remix them.

However, if you’re in the US, it doesn’t seem like you can download Pocket just yet. Two US-based Verge staffers saw a note on the Google Play listing that said the app “isn’t available in your country.” I can’t find the app on Apple’s US App Store. Meta also notes in its help center that the app is “not yet available everywhere.” The company didn’t immediately reply to a request for comment.

While Microsoft focuses on the productivity suite, Salesforce is positioning Slack as the “agentic operating system” for the enterprise. As of April 2026, Slack AI has moved beyond summarizing to orchestrating agentic workflows. This is designed let you trigger complex, multi-step actions across non-Microsoft systems directly from a Slack thread.

Salesforce’s Agentforce platform uses the Atlas Reasoning Engine, which is designed to offer autonomous front-office automation (sales, service, and marketing). For organizations where CRM data is more critical than Word documents, Agentforce is emerging as a formidable, high-ROI alternative to Copilot.

In March 2026, Apple launched Apple Business, a platform designed to integrate Apple Intelligence directly into macOS and iOS. Apple claims its competitive edge is its on-screen awareness. Unlike cloud-heavy competitors, Apple Intelligence is built to act across apps locally, appealing to regulated industries concerned about data leakage.

Q: What sort of threats are you seeing right now? “Infostealer malware remains the single biggest threat to macOS right now. These stealers trick users into running them through convincing fake websites and social engineering, then exfiltrate as many credentials and secrets as possible for the attacker to use, sell, or trade on the dark web. 

“Apple regularly ships new protections, and attackers just as regularly adapt their social engineering to stay ahead of them. Techniques like ClickFix, where users are tricked into pasting and running malicious commands themselves, have become especially effective because they bypass many protections entirely by getting the user to do the work. Beyond that, supply chain attacks are growing at an alarming rate, with attackers compromising developer libraries that get pulled into internal or production projects, quietly introducing backdoors without the creator’s knowledge.”

Q: What about the manufacturing sector? Is there any excuse to use legacy kit at all in an AI threat age? “AI generated threats won’t necessarily be different than traditional ones, but they will stress the seams of traditional security programs that will need to have improved visibility at scale and be able to work at a new kind of speed and agility from start to finish.”

Beyond the potential political ramifications, any deal would have immediate implications for enterprise IT buyers.

“CIOs should focus on the risk that this strategy could introduce. Will Apple be able to thoroughly assess those chips to completely rule out the possibility of trojan horses, backdoors, and hidden functionality such as dead man switches?” asked Flavio Villanustre, CISO for the LexisNexis Risk Solutions Group. “If Apple says that they will do, to what degree of certainty? There have been rumors about hidden backdoors in chips before, such as Supermicro in 2018, ESP32 microcontroller hidden functionality in 2025, and Microsemi backdoor in 2012, to name a few.”

On the naughty list?

This issue gets complicated based on what the US government ultimately does. The two Chinese manufacturers figure on the Pentagon’s so-called 1260H list of “entities identified as Chinese Military Companies,” which also includes Chinese internet giants Alibaba, Baidu, and Tencent; router maker TP-Link Technologies; and drone maker DJI. Being on that list has no real consequences for the companies concerned, but the government could move them to the Department of Commerce’s Entity List, subjecting them to export licensing requirements, or make them the subject of a Section 889 clause, barring them from government procurement deals. That could sharply change the dynamics for Apple and other technology vendors seeking cheaper RAM supplies — and for their customers.

For a few days, it seemed like Universal decided that there would be no advanced screenings of Christopher Nolan’s The Odyssey for influencers. But on Monday, influencers sat alongside traditional critics and journalists at special showings of The Odyssey specifically for the associated press junket. Despite what it may have looked like, Universal was not going back on its word. The studio was simply embracing an undeniable reality of the modern entertainment press.

Much as it may pain some to admit, influencers and content creators have become a significant part of our media ecosystem — one that offers film studios an invaluable way to reach larger, more diverse, and younger audiences. Whether it’s a YouTuber’s video essay, a TikToker’s reaction video, or a glowing review on Instagram, influencers’ content can go viral if it resonates with (or enrages) a rabid fan base. And the way that some influencers build massive followings of their own highly engaged fans makes them useful for studios hoping to engineer the next Barbenheimer moment.

People’s reactions to the idea of Universal snubbing creators ahead of The Odyssey’s theatrical premiere highlights another important facet of our media landscape. Influencers have become a big part of how studios promote their films. Speaking to Yahoo Entertainment, digital marketing strategist Gigi Robinson explained that influencer screenings are “one of the smartest moves in entertainment marketing right now” because of how they “create a full content arc.”

“It’s not just about showing up and watching a movie,” Robinson said. “Every creator makes it their own. These screenings make people feel like they’re part of something exclusive. And that anticipation? That’s what gets people into theaters.”

But some people have grown weary of the (often overly enthusiastic or cartoonishly negative) energy that they bring to the table. It’s clear that studios are aware of this on some level, otherwise trades wouldn’t be running stories about who’s getting invited to screenings.

What Universal probably meant when the no-influencers-allowed news broke was that there won’t be any screenings of The Odyssey that are put on solely for content creators. Those types of screenings aren’t unheard of, but it’s far more common for studios to screen their unreleased projects for small audiences that consist of critics, journalists, regular theatergoers (who have usually won a contest of some sort), and a handful of people best known for their social media presences. Having that mix of people from different walks of life makes for a viewing experience that feels much more like what it might be like to see the movie opening night. And that feeling can — but doesn’t always — inform a reviewer/writer/video essayist’s thoughts as they work on pieces about what they’ve just seen.

One of the more frustrating things about these screenings is the way their scheduling tends to leave members of the press with very little to finish their work before embargos (agreements barring journalists from publishing until a certain time) are up. It can be difficult to put a piece of insightful film criticism together in just a few hours before you have to post it and hope that Google Zero won’t make your work almost impossible to find (it will.) This is particularly challenging for writers who try to hit deadlines in order maximize their chances of catching readers when they’re intentionally looking for news about a movie. And while some studios probably see reviewers having less time to sit with their thoughts about a movie — some of which can be negative — as a good thing, it also creates an environment that discourages people from producing quality criticism.

These hurdles aren’t always an issue for influencers whose contributions to the discourse can be as simple as a hyperbolic X post telling their followers that [insert movie] is the best or worst thing they’ve ever seen. The relative speed with which creators are able to broadcast their thoughts to their audiences is part of what makes them so useful to studios. And because there are quite a few movie influencers who have built brands around hyping up movies rather than discussing them thoughtfully, studios have relied on them to generate positive buzz.

Though there are plenty of creators whose work is much more than that, the algorithmically driven social media platforms we use every day often reward content that’s attention-grabbing and short enough to be consumed quickly. This is why you still see a sea of soyface thumbnails whenever you search through YouTube for basically anything, and it’s why so many online movie influencers only seem capable of speaking in extremes. These tactics have helped influencers build dedicated followings of people who see them as entertainers and news sources that are more trustworthy than traditional media outlets.

Influencers have given people an easy way to access content about art in a moment when newsrooms across the board are shrinking. Social media personalities might not be as well-equipped to go deep with a director about their process or influences, but that doesn’t necessarily matter to a subscriber who sees them as a friend who conveniently lives in their phones. Parasocial relationships are also a key part of the creator economy that studios are now trying to capitalize on through things like influencer screenings. It takes more energy and focused attention to get through a written movie review than it does to half-watch a reactionary TikTok that pops up on a curated For You page. But making the effort to engage with nuanced, studied criticism is really the only way we can encourage studios to help foster it.

This doesn’t mean that everyone should just unfollow every creator who pops up in their feeds or that studios should only work with press from legacy media outlets. But it does mean that we should be willing to work a little harder to find and share substantive criticism that isn’t just a bit of promotion meant to get us into theaters.

Yet without the right foundations, many of those investments may not fully deliver on their potential. Companies that already operate with discipline have an edge. They can channel new tools into proven systems rather than bolting them onto shaky foundations. Organizations with mature process disciplines are also better positioned to translate AI ambition into real outcomes, as they are already accustomed to data-driven decision-making and process discipline—precisely the cultural foundation AI systems need to deliver value.

Simply put: AI can accelerate process excellence, but existing process excellence is what makes AI truly impactful. Technology and process are no longer separate levers, and only organizations that pull them together stand to realize the full value of both.

Download the full report.

This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff. It was researched, designed, and written by human writers, editors, analysts, and illustrators. This includes the writing of surveys and collection of data for surveys. AI tools that may have been used were limited to secondary production processes that passed thorough human review.

If you don’t want to spend $180 on the Elgato Key Light, the Godox ES45 Desktop LED Key Light is a more affordable option that offers nearly as much brightness and plenty of features, including adjustable brightness and color temperature. Right now, it’s cheaper than usual for the first time this year, selling for $119 ($20 off) at Amazon and B&H Photo.

The ES45 is a worthwhile purchase for streamers, or really anyone who wants to look their best on camera. You can position the light to shine directly on you, or aim it at your wall for some even bias lighting that creates a glow effect in your room that’s easier on the eyes. It can go between a warm 2,800K to a cooler, daylight-like color temperature.

One feature it has over Elgato’s popular Key Light is an included wireless 2.4GHz remote, which lets you easily adjust the lighting without having to fiddle with desktop or mobile apps. The remote automatically recharges when you reconnect it magnetically to the light. The ES45 also comes with a desk clamp that requires no tools to install, letting you attach it to a desk that’s up to two inches thick.

Three more Verge-approved deals

Andrew: Well, Megan, we’ve had a philosophy for a long time in Woodside from an innovation perspective, where we really want to think big, we want to prototype small, and we want to scale fast. We want to find big opportunities that we can go after, but we want to ensure that we look at how we deploy those on a small scale first, and then provide the right learning and insight that then can scale it everywhere. Something like maintenance intelligence is a good example of that, or our Startup Advisor, where we know that we’ve got multiple plants that we need to start up. We know that we’ve got multiple assets that need to do maintenance, so we have a big, bold ambition about how we can improve and optimize that. We start with a small prototype; it might be one subsystem, it might be just a part of an asset, and then we scale it out, we learn, and we scale faster.

I think from an AI learning perspective, one of the key things we’ve learned is really the transition from moving from isolated AI solutions to a more coordinated enterprise-wide capability. If you look back maybe 18 months, two years, in our generative AI journey, we rarely started by deploying AI as broadly as we could in the organization from a personal productivity perspective. And probably being quite open in terms of the problems that we will solve, the business problems that we’ll solve with AI. That had a lot of benefits for us in terms of allowing our organization to get to know AI, get to know the capabilities, to build the trust in it.

What we’ve learned though is that we’ve needed to pivot from that to being a little bit tighter in terms of where we are going to invest our time and resources and more higher value solutions. How do we then enable and empower the rest of the organization so that they can actually effectively problem solve with technology in their domain or in their personal productivity without having to come to a central team?

When we think about that, think big, prototype small, scale fast, has been something really important for us. The transition from a more broader approach to use case development and solution development to now a narrower focus on the high value priorities. We’ve seen that paying dividends to us and allowing us to go after solutions and opportunities, things like Startup Advisor.

And so our Startup Advisor is a agentic AI solution that really aims to optimize and empower and better support our operators that sit in front of a panel and have to start up LNG plants, which are incredibly technical facilities and require really specialist skills to start up. And so our Startup Advisor is almost like a copilot that sits alongside those operators, and it gives them the ability to be able to play back previous startups. It gives them the ability to look at how the current startup is progressing, and it provides them better insights to optimize how they start up that facility. And again, starting up an LNG facility is incredibly complex.

Megan: I can imagine.

Andrew: When we think about opportunities like Startup Advisor, again, it goes back to that think big, prototype small, and scale fast. We started with a very bold vision of, how do we start up all of our LNG plants in a much more structured and optimized fashion? How do we better support our panel operators? How do we make, say, a more junior panel operator have a copilot that can help them almost like an experienced panel operator sitting next to them? And when we think about that vision and the ability then to prototype on a small scale and then scale fast, I think it’s been really successful for us.

As we scale, we’ve just naturally expanded into more agent-based solutions. Today, we’ve got around 50 AI agents in production, supporting both our operating assets and our enterprise workflows. These tools have been proven in live environments, and we have really seen the benefit of being able to shift from point solutions that maybe solve small scale problems in specific areas, to AI and agentic solutions with agency that can really work across our workflows.

We’re able to do this because we’ve standardized on the platform that we build on and we’ve got repeatable patterns. That’s been another really important learning for us, is that we don’t want to build 50 solutions in 50 different ways. We really want to be empowering our organization and our technical teams and the users of our solutions to roll them out quickly, to roll them out safely, and to do it in a patternized and platform manner.

But the last point I’ll make, Megan, from a learning perspective is that we’ve really understood that a strong governance around how AI is deployed and developed is critical for us, and it’s critical for us to go fast as well. The traditional ways of governing how we roll out different solutions or digital systems isn’t going to scale to the breadth that we need when we are thinking about AI. Being able to have a clear philosophy around how we innovate, transitioning from isolated solutions to that enterprise-wide capability, and making sure that we’ve got strong platforms with strong patterns and clear governance are the three really critical things that we’ve learned.

Megan: Such important pillars, all of them. And you’ve been working with Infosys on this journey. How has that partnership helped accelerate scaling and embedding AI across the business?

Andrew: Well, Infosys is our managed service provider, and so they play a really critical role in the operations of our core business. One of the things that I like to say is that our license to innovate is based on our license to operate. And so, for my team to be able to turn up to an operating asset or a corporate function and have the trust that’s needed to be able to innovate and reimagine and redesign how work gets done, to be able to do that, we need to make sure that our core platforms, our core systems, our applications are running really reliably, safely, and consistently every day. Having an experienced partner like Infosys looking after those core operations in partnership with our internal teams is really, really important to us.

As we move from pilots to enterprise-wide deployment, the ability to partner with someone like Infosys also gives us the ability to scale. And so being from Perth and Western Australia, while we’ve got a really strong local team in Western Australia, and we’ve also got a very strong team in some of our other operating locations, like everyone, we’re struggling to find people that can fill AI roles. Being able to partner with Infosys and have a number of different operating models at our disposal becomes really important for us. Having co-mingled teams where they are staff, they are Infosys staff, Woodside staff, and some of our other partners, really just brings diversity of thought and experience to how we solve problems.

Fundamentally, the partnership has allowed us to operate and innovate with more confidence. While Woodside always retains ownership of the strategy and where we’re going and the governance and my teams remain accountable for the outcomes, we can’t do what we do without strong partnerships like the one we have with Infosys.

Megan: Fantastic. And as AI adoption scales, you mentioned yourself, governance becomes increasingly important. How challenging has that been, and what guardrails have you put in place at Woodside?

Andrew: So, Megan, governance is really important to us, and we operate in a well-regulated environment. That means we’ve got to make really deliberate and well-reasoned decisions when we’re thinking about how we deploy technology into our organization, whether it’s artificial intelligence or anything else, for that matter. And so, governance is really central to how we approach the execution of our AI strategy at Woodside.

We’ve got maybe two or three really key things that we’ve put in place. The first one is just making sure that every AI use case goes through a structured assessment, and that’s making sure it meets our privacy controls, our cyber controls. We’re also asking the question, not just, could we do this, but should we do this? We’ve really got to bring together safety, ethics, transparency, accountability, and make sure that we make an informed decision. When an AI solution is going through that structured assessment, if there are concerns about how we might use that solution, it then goes to an AI council that’s made up of senior leaders across the organization. That council and that group really oversee some of the prioritization and risk management. That’s where we can have really strong, robust debates around, again, could we do something, should we do it, and how do we mitigate any of the risks that we might introduce here?

I think the last one, Megan, is really around lifecycle management. When you start thinking about, we’ve got 50 at the moment, but if we had 500 agents working in our organization, really amplifying the experience and the decision-making and the value creation of our staff, we really want to have an ability to manage the lifecycle of how those agents operate. We want to know, how many people are using them? What’s the efficacy and the outcome? Is there model drift? Do we need to retune or retrain? I think that’s an area where many organizations, including Woodside, are still leaning into and still figuring out the best way to do this. We can do it quite easily with 50 agents, but 500, 5,000, 50,000 becomes an opportunity for us. Again, thinking about how we partner with others, solving problems like that really present an opportunity to co-create and to co-solve with some of our partners, like with Infosys.

Megan: Fantastic. Just to close, what’s your long-term vision for AI at Woodside? How do you see this evolving over the years ahead, and what could it unlock for the sector in your view?

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology.

LLMs are stuck in a groupthink groove. This startup is trying to get them out.

Open up your chatbot of choice—Claude, ChatGPT, Gemini—and type “Give me a random number between 1 and 10.” You’re going to get 7. Almost always. 

That won’t work every time—but if it did for you, you may wonder if I have superpowers. I don’t.

The truth is that most large language models are stuck in a rut. They are far more predictable and far less creative in their responses than you might expect. That’s fine for tasks like coding or research, but groupthink is a problem when you’re brainstorming or planning your next vacation.

The Australian startup Springboards has a solution. It built an LLM called Flint, which has been trained to come up with a wider variety of responses than mainstream LLMs to open-ended questions such as “Where should I go in Europe?”

Meet the company pushing chatbots away from the obvious.

—Will Douglas Heaven

The must-reads

I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.

1 Scientists say they have built a cell from scratch for the first time
Built with lab-made DNA, it can feed, grow, and multiply. (CNN)
+ It brings us closer to creating synthetic life. (Quanta)
+ And is arguably the greatest feat of bioengineering yet. (New Scientist $)
+ But also raises concerns over the dangers of synthetic biology. (NYT $)
+ Mirror organisms could threaten life on Earth. (MIT Technology Review)

2 OpenAI has proposed giving the Trump administration a 5% stake
Talks over a public ownership deal come amid rising political pressure.(FT $)
+ OpenAI also proposed other US AI giants providing a 5% stake. (CNBC)
+ That could include Anthropic, Google, and Meta. (Bloomberg $)
+ President Trump says he wants the public to have a stake in AI. (BBC)

3 Singapore has seized a $42 million mansion tied to Nvidia chip smuggling
It was seized as part of an investigation into alleged illegal trading. (BBC)
+ Days earlier, Supermicro’s Taiwan offices were raided in the probe. (FT $)

4 Anthropic’s Fable 5 is back online
But queries posing security risks may be routed to less powerful models. (Axios)
+ Anthropic restored access yesterday after the US lifted an export ban. (BBC)
+ But the battle over how to tame AI has just begun. (WSJ $)
+ Anthropic has launched a new AI science product. (MIT Technology Review)

5 Meta is building its own cloud infrastructure business
It’s exploring two ways of monetizing AI compute and models. (Bloomberg $)
+ One is selling access to models hosted on Meta’s infrastructure. (CNBC)
+ The other is selling “raw” computing power. (TechCrunch)

6 PlayStation will stop releasing games on discs in 2028
Future PS5 games will be digital-only releases. (Verge)
+ The news comes days after reports that GTA VI will have no disc. (BBC)
+ It’s put a nail in physical media’s coffin. (Wired $)

7 A low-cost Chinese AI model is catching up with US giants on their home turf
Western customers are drawn to GLM-5.2’s cheap but powerful model. (Reuters $)
+ Chinese open-source models are spreading fast. (MIT Technology Review)
8 Google has lost its fight against a record €4.1 billion EU antitrust fine
It was charged in 2018 for using Android to ‌block rivals. (CNBC)

9 The UN has launched an “AI for Good” commission
Salesforce CEO Benioff and Rwandan President Kagame will co-chair it. (Axios)

10 People prefer AI impersonators over politicians
The study’s findings raise alarm bells around potential public deception. (404 Media)

Quote of the day

“If AI overdelivers, it will impact financial stability. If AI underdelivers, it will impact financial stability.”

—Torsten Slok from Apollo Global Management shares common concerns about AI at the European Central Bank’s annual conference, Reuters reports.

One More Thing

America was winning the race to find Martian life. Then China jumped in.
In July 2024, after more than three years on Mars, the Perseverance rover came across a peculiar rocky outcrop. Instead of the usual crystals or sedimentary layers, this one had spots. Those specks were the best hint yet of alien life.  

NASA began a new mission to bring the rocks back to Earth to study. But now, just over a year and a half later, the project is on life support. As a result, those oh-so-promising rocks may be stuck out there forever. 

This also means that, in the race to find evidence of alien life, America has effectively ceded its pole position to its greatest geopolitical rival: China. Beijing is now moving full steam ahead with its own version of NASA’s mission. 

Here’s how the search for Martian life has become a contest between two superpowers.

—Robin George Andrews

We can still have nice things

A place for comfort, fun, and distraction to brighten up your day. (Got any ideas? Drop me a line.)

+ The classic arcade game Crazy Taxi is returning.
+ Thom Yorke’s live set from the Sydney Opera House is a reminder of what an extraordinary performer he is.
+ Peer into 1,000 gloriously illuminated New York apartment windows at night in this generative photography project.
+ The Orion constellation dazzlingly displays every stage of star formation in this image from the James Webb Space Telescope.

Top image credit: Sarah Rogers/MITTR | Photos Getty

Please send gloriously illuminated New York apartments to hi@technologyreview.com. 

You can follow me on LinkedIn. Thanks for reading!

—Thomas

Twenty-five years ago today, a young, little-known programmer by the name of Bram Cohen fired off a short message to a mailing list for peer-to-peer enthusiasts. “My new app, BitTorrent, is now in working order, check it out here,” Cohen wrote, followed by a link to his personal website.

“What’s BitTorrent, Bram?” the founder of the list asked in response.

Cohen never bothered to reply. The world would find out soon enough.

In the following years, BitTorrent quickly became the world’s most popular file-sharing app, unleashing a massive wave of piracy that upended Hollywood forever. At one point, BitTorrent was said to be responsible for a huge amount of internet traffic — some widely cited metrics peg it at half of P2P and one-third of all internet traffic in 2004. And while the entertainment industry succeeded in shutting down file-sharing systems like Napster and Kazaa, it largely failed to curtail the massive flood of BitTorrent piracy.

What stymied movie studios and record labels alike is exactly what makes the story of BitTorrent’s 25th anniversary a gripping one. There’s BitTorrent, the app that Cohen unveiled in July 2001 and which continues to attract tens of millions of monthly users to this day. There’s BitTorrent, the file-sharing protocol, which has been adopted and advanced by hackers and developers from around the world, and which helped birth an entire cottage industry of piracy websites. And there’s BitTorrent, the company cofounded by Cohen, which struggled for years to make money with its technology.

“Going into it, my plan was not to start a business,” Cohen recalls. “My plan was to start a revolution.”

This is the story of that revolution, and all that remains of it today.

Cohen started working on BitTorrent soon after leaving Mojo Nation, a startup that had ambitious plans to combine file sharing, distributed computing, and micropayments, only to shut down in 2002 without ever really getting off the ground.

Cohen helped Mojo Nation improve the efficiency of its file-sharing technology. His approach was to use something called swarming distribution: Instead of swapping large files between a few users at a time, the app would divide each file into lots of little chunks, and then allow large numbers (or swarms) of users to trade those chunks among themselves. The ethos of sharing was built into the software, insisting that users would help upload files instead of just downloading them selfishly.

Cohen was disillusioned by the failure of Mojo Nation but not ready to give up on the underlying technology just yet. “I decided to make a tool that very narrowly just did swarming distribution and nothing else, because that was a thing I knew how to do,” he says.

“My plan was not to start a business. My plan was to start a revolution.”

All this happened as P2P and file-sharing companies were facing massive headwinds. The music industry sued Napster at the end of 1999 and ultimately forced it to shut down around the time Cohen first released BitTorrent. A few months after that, the entertainment industry sued Kazaa and Grokster, followed by legal claims against Audiogalaxy, LimeWire, eDonkey, and other file-sharing systems.

Despite some significant differences in their P2P architecture, all these apps ultimately lost in court or folded ahead of a likely legal defeat. Their Achilles’ heel: They offered users a way to search for copyrighted files and then helped them find others to trade those files with. Courts deemed this contributory infringement, pointing to the fact that the owners of these services knew exactly which files their users were trading.

BitTorrent was different. From the start, Cohen built the app without search functionality, instead relying on third-party websites for content discovery. He also didn’t run a centralized server to help users discover other BitTorrent clients to trade content with, and instead outsourced that piece of the puzzle to so-called tracker servers. Websites dedicated to BitTorrent file sharing would offer a metadata file (the torrent file) for every download, which in turn pointed users to tracker servers that matched up users to facilitate the sharing. Simply put, BitTorrent was an app to help people trade data, but its developer had no way of knowing what that data was.

That structure effectively shielded Cohen and the company he would eventually found from legal liability. To this day, BitTorrent remains the only major publisher of file-sharing software that hasn’t gotten sued by Hollywood and the major labels. However, Cohen insists that the decision to outsource content search and client discovery wasn’t primarily a shrewd legal maneuver, but a simple necessity.

After all, he was just one guy. “My hands were somewhat tied,” he says. “I was writing stuff on my own.”

Despite lacking the resources of much bigger P2P companies, Cohen’s BitTorrent app quickly gained traction in tight-knit communities looking to trade large files. Among the first adopters was Etree, an online community of jam band enthusiasts. Bands like Phish and the Grateful Dead had long encouraged their fans to trade bootleg recordings of their shows with each other in the offline world. Etree brought this idea online with the help of BitTorrent, which allowed fans to share show recordings using high-resolution audio files that were simply too big to distribute any other way.

But pirated videos were like rocket fuel. “When it started getting used for anime was when the numbers took off,” Cohen says. Granted, that early success didn’t quite measure up to the tens of millions of users apps like Kazaa were attracting at the time. “It was like, Oh, look, we’re getting a thousand downloads a day, that’s massive,” he recalls.

Things changed significantly with the emergence of major torrent sites, which allowed anyone to quickly find and download movies and TV shows. In 2002, a Slovenian high school student launched Suprnova.org, which reportedly attracted up to a million users a day before its closure in 2004. And in 2003, a ragtag group of hackers founded The Pirate Bay in Sweden. With its open embrace of illicit content and an F-you attitude toward Hollywood, the site turned into a kind of lightning rod for self-described pirates and copyright enforcers alike.

“It was so fast that we initially thought something was broken.”

While these sites quickly became household names, countless others sprung up all across the globe in their shadows. Some catered to specific interests like anime, porn, or ’70s B movies, while others simply welcomed everyone and everything. And while copyright holders were able to shutter some sites, others survived, or reemerged under different names or domains. Plus, all those lawsuits generated countless headlines, effectively promoting BitTorrent to new audiences. By 2004, BitTorrent traffic exploded.

The surge of interest didn’t just alarm rightsholders. It also put some significant strain on BitTorrent’s infrastructure, including the tracker servers that helped users discover each other. One popular tracker, Denis.Stalker.H3Q.com, was run by a trio of German hackers associated with the local Chaos Computer Club.

Initially, their tracker ran server software Cohen had written in conjunction with the BitTorrent client app. But as BitTorrent gained in popularity, their server, which was just a single computer humming along in a kitchen somewhere in Berlin, was bombarded with hundreds of requests per second, slowing it to a crawl.

That’s when Dirk Engling, a member of the trio who is also known as Erdgeist in German hacker circles, decided to build a better-performing tracker software from scratch. He dubbed his server software Opentracker, and the difference was like night and day: While the Denis.Stalker server previously always ran hot, Engling’s code put a sudden stop to the near-constant fan noise. “It was so fast that we initially thought something was broken,” he says.

Engling is not as well known as Cohen. But back then, he made a fateful decision that ultimately helped BitTorrent scale to surpass all other file-sharing networks in popularity. Instead of just using his code on the Denis.Stalker tracker, Engling published it online, free to download for anyone.

Opentracker quickly took off and was adopted by major torrent sites like The Pirate Bay as their tracker solution of choice. Many of these big sites relied on third-party tracker servers like Denis.Stalker, but for years also ran their own trackers as fallback solutions. Every torrent file can point to multiple tracker servers, so sharing continued when one server went down. (Eventually, BitTorrent added further decentralization to allow trackerless trading, but tracker servers remain to this day an important part of the ecosystem.)

“Back then, it was a radical political idea,” Engling says. The thought was to allow anyone to distribute large amounts of data without having to rely on expensive servers — and if some of that was used for piracy, so be it. “It was a movement to democratize [file sharing],” he says.

Engling, for one, wasn’t in it for the money. He distributed Opentracker as “beerware,” an open-source license guided by the notion that anyone who found it useful should buy him a beer, as he half-jokingly explained in a 2007 talk. “I thought: A lot of people are using BitTorrent, so…”

Back in the United States, Cohen quickly realized that relying on the goodwill of others — especially people who used his software to download content for free — didn’t pay the bills. In 2004, he founded BitTorrent Inc. to transform the revolution he had started into an actual business. “That didn’t work so well, unfortunately,” he now concedes.

BitTorrent raised $8.75 million in funding in 2005, followed by another $20 million round in 2006. The money helped the company to staff up, but it also put pressure on Cohen and his teammates to somehow turn an app that was beloved by pirates into a legitimate business.

As part of these efforts, the company struck deals with Hollywood to launch a movie download store called the BitTorrent Entertainment Network, which launched with thousands of titles from Warner Bros., Paramount, and MGM in early 2007. However, the company quickly realized that selling DRM-protected movies to an audience that already had access to the same titles for free was a fool’s errand.

“Back then, it was a radical political idea.”

“That didn’t work at all,” says a former senior BitTorrent employee who declined to be named in this story because of fear of repercussions from his current employer. Not only did people not want to buy movies from BitTorrent, but minimum guarantee provisions in its contracts with Hollywood also forced the company to fork over millions of dollars for content no one wanted to buy. “It kind of worked out for the studios,” the former employee says.

“There were gotchas [in those] deals that made them essentially unworkable,” Cohen agrees. “The big copyright holders were perfectly happy with where things [were], and basically didn’t want to create the streaming systems that we have today.”

BitTorrent closed the download store less than two years after its launch while also laying off half of its staff amid a recapitalization. In the years that followed, the company kept trying to figure out other business models, from P2P CDNs to decentralized chat clients — only to hit roadblocks at every turn.

The big issue was always piracy. BitTorrent successfully avoided legal liability, but it couldn’t get rid of its outlaw image. That made it impossible to strike deals that were open to others, Cohen says. “When you’re doing slightly sketchy things, the world doesn’t want to give you a nice off-ramp.”

“I never should have brought on anybody else at all. I could have simply gotten enough money off of doing the slightly sketchy stuff just on my own.”

One thing that worked was advertising. In addition to including ads directly in its products, BitTorrent would also bundle third-party tools with its installers and get a small cut when someone installed a browser toolbar or a similar app alongside the BitTorrent client. But while the ad money was paying the bills, it wasn’t meeting the expectations of BitTorrent’s investors looking for a lucrative exit.

“The [app download] numbers kept growing, but a lot of the use cases were never going to be the kind that you could legitimately either IPO or sell to another company,” the former employee says.

BitTorrent Inc.’s struggles were in stark contrast to the success of Cohen’s technology. “We were moving an exabyte of data every month,” the former senior employee estimates. BitTorrent owed much of its popularity to the wars around file sharing, which got a new poster child when Swedish authorities raided The Pirate Bay in 2006.

The battle between rightsholders and The Pirate Bay is perhaps the best-known part of the BitTorrent saga. The site was up and running again days after the raid, thanks in part to BitTorrent’s distributed architecture. The Pirate Bay didn’t run a file-sharing network, and taking its servers didn’t stop people from downloading BitTorrent’s client, or finding each other via tracker servers. All the raid took offline was a public list of downloadable torrent files, and all it took to get that back online was a recent server backup.

Undeterred, rightsholders and law enforcement continued their crusade against BitTorrent pirates and their alleged enablers, including the three cofounders of The Pirate Bay. After getting charged with criminal copyright infringement in early 2008, in April 2009 they were each sentenced to a one-year prison term and ordered to pay financial damages. Much to Hollywood’s chagrin, their site nonetheless stayed online. The exact identity of The Pirate Bay’s new operators remains a mystery to this day.

Hollywood’s fight against pirates also had repercussions for the German hackers responsible for much of BitTorrent’s tracker infrastructure. Used by a number of big torrent sites, the Denis.Stalker tracker attracted around 100 million daily peers at its peak, which were trading a million files every single day, according to Engling.

“When you’re doing slightly sketchy things, the world doesn’t want to give you a nice off-ramp.”

One day, the Berlin police’s economic crimes unit raided the apartment of one of the three people involved in running the tracker server, confiscating every piece of equipment they could find, according to Engling. “They couldn’t imagine that anyone was running something like this without making any money with it,” he says. However, money was the last thing on the trio’s mind. In the same aforementioned talk, held during the Chaos Computer Club’s annual conference between Christmas and New Year’s, they explained their motivation with three simple words: “Because we can.”

Charging them for knowingly facilitating copyright infringement wasn’t possible either, as tracker servers simply match users by hash values and never see the actual files being traded. Ultimately, the case was dismissed, according to Engling.

“Those were the wild early days,” he says.

Things changed rapidly in subsequent years. Unable to put an end to torrent sites and trackers, rightsholders shifted their focus to end users. In some jurisdictions, including Germany, this can be extremely lucrative. Law offices can effectively send cease and desist letters that double as invoices for hundreds of euros per traded file to alleged offenders, threatening them with full-blown lawsuits if they don’t pay up. Germany’s Stern magazine estimated that local file sharers received claims for around €120 million in 2010 alone.

That led to end users embracing VPNs to protect themselves — and torrent sites cashing in on that trend by adding ads for privacy services to their sites. With that, torrent sites were starting to become a major moneymaker. One 2014 study estimated that torrent sites generated around $114 million in ad revenue that year. In addition to VPNs, torrent sites have long been hawking anything from porn to dating to online casinos, with the occasional malware download thrown in for good measure.

These days, some sites also use elaborate schemes to hijack otherwise legitimate advertising and affiliate revenues. One such trick: When someone uses a torrent site, their browser opens an Amazon page for an extremely random product in the background. Toothpaste, maybe, or perhaps a totally unremarkable sweater. The goal isn’t to sell those specific products, but to plant a cookie on the visitor’s machine. If they order anything from Amazon in the following 24 hours, the pirates earn some affiliate revenue — a tactic known as cookie stuffing.

“It’s whatever the latest scam is,” says Engling, sounding disillusioned about what has become of the BitTorrent world he helped create. “The idea of democratization got lost,” he says.

But while Torrent sites were making bank, BitTorrent Inc. continued to struggle. “I did a good enough job of keeping the wheels from falling off the bus, keeping our core product working,” Cohen says. Efforts to move beyond file sharing, however, kept failing.

That included Cohen’s next big bet: For years, he worked on adapting the company’s tech for low-latency livestreaming. BitTorrent Live, as the initiative was called, even shipped consumer apps, and the company was briefly working on building a kind of live online TV service. In 2017, the company pulled the plug on those efforts and laid off almost the entire BitTorrent Live team. “This is part of the world being out to get you,” Cohen says, clearly still bitter that the project wasn’t given a chance. “That was a product that worked great and was ready to go, and was just outright fucking killed.”

“They couldn’t imagine that anyone was running something like this without making any money with it.”

Soon after, BitTorrent Inc. got sold to controversial crypto entrepreneur Justin Sun, who went on to shutter the company’s SF office after a few very tumultuous years and now operates the company from overseas. (BitTorrent did not respond to multiple requests for comment.) “It eventually got an even sketchier exit than it had been all along,” Cohen quips.

Looking back, Cohen believes that many of BitTorrent’s troubles could have been avoided if he hadn’t gone down the typical path of a VC-backed startup. Perhaps he should have just done his own thing — not unlike the path Engling chose, except more focused on modest revenue and less on free beer.

“I never should have brought on anybody else at all,” Cohen says. “I could have simply gotten enough money off of doing the slightly sketchy stuff just on my own.”

How many people still use BitTorrent today is unknown. The days of it making up a third of all internet traffic are clearly over, with one estimate suggesting that it now contributes less to residential upstream traffic than iCloud and FaceTime. BitTorrent Inc., the company cofounded by Cohen and now owned by Sun, claims that its clients still have 54 million monthly users. Popular third-party clients like Transmission, BiglyBT, and qBittorrent likely attract millions more.

I know what you download, a website that tracks usage across multiple torrent sites, estimates that about 0.25 percent of all internet users download torrents on any given day. “The number of torrent users [has been] pretty stable over the last eight years,” says the site’s admin, Andrey Rogov. Data Rogov shared with The Verge shows Russia, where Netflix and other streaming services aren’t legally available due to sanctions related to the Ukraine war, as BitTorrent’s biggest market. The United States ranks second.

Engling doesn’t have any data on the total usage of Opentracker, but he does at times look at the publicly available reporting dashboards of popular torrent trackers running his software. That data suggests we may be seeing a resurgence. “In absolute numbers, it has gotten more,” he says.

“It’s whatever the latest scam is. The idea of democratization got lost.”

One reason? Streaming is getting awfully expensive these days. For a while, people were fine paying for Netflix and Disney Plus instead of downloading movies and shows for free. “At some point, it flipped,” Engling says. “An increasing number of consumers who previously paid are dusting off their pirate hats.”

Amid that resurgence, he still sometimes tinkers with Opentracker’s code, reenergized by new technical challenges. And just last month, a Spanish anime fan took him out for beers in Berlin to thank him for his contribution to BitTorrent’s global infrastructure.

Cohen, on the other hand, is more than happy to leave that chapter of his life in the past, and instead focus on his blockchain startup Chia Network. “I don’t really care to revisit BitTorrent too much,” he says. Except, if you start a revolution, some of it always stays with you, whether you like it or not. What haunts him to this day are the struggles that came with turning BitTorrent Inc. into a venture-funded startup, and ultimately a profitable business.

“You know how people say they wake up their whole lives with nightmares that it’s the morning of a final exam in grad school? I never had those,” Cohen says. “I do get nightmares about board meetings.”

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Janko RoettgersClose

  Janko Roettgers

  Lowpass author, Verge contributor

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Janko Roettgers

• EntertainmentClose

  Entertainment

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Entertainment

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

OpenAI has floated giving the US government a 5 percent ownership stake as a way of easing tensions with the Trump administration and blunting mounting public backlash against AI, according to the Financial Times.

CEO Sam Altman argued that giving the public a financial interest in the company would be the best way to share the upside of AI, the FT reported, citing two unnamed people familiar with the talks. He’s said to have first pitched the idea to Trump early last year.

Altman reportedly suggested the 5 percent figure. Based on OpenAI’s latest funding round, which ended with the company valued at $852 billion, that stake would be worth roughly $42.6 billion.

The discussions are reportedly still in their early stages, and the proposal would involve other US AI companies giving the government similar stakes. It’s unclear whether they would agree to such a deal.

Public officials have shown growing interest in using policy to capture and redistribute some of the wealth generated by AI. Under Trump, the US government has already taken a 10 percent stake in chipmaker Intel and reportedly demanded Nvidia and AMD give the federal government a 15 percent cut of their revenue from AI chip sales to China. Others, such as Senator Bernie Sanders, have argued that AI is a public resource and suggested a one-time 50 percent tax on their stock value to establish a sovereign wealth fund.

Here’s how the system works: The state’s climate regulations require the transportation fuels industry to lower the carbon dioxide levels in its products over time—or purchase credits from other parties that cut fuel emissions, including cattle farmers.

Dairies generally spray cattle manure into giant open lagoons, where microbes gobble up organic matter and produce methane as a by-product. But if farmers set up what are known as anaerobic digesters, the sludge is redirected into covered vessels that capture the biogas, which can be converted into natural gas and injected into a pipeline. It can then be used to fuel certain vehicles or generate electricity in a power plant. Either way, petroleum companies can pay those farmers for Low Carbon Fuel Standard (LCFS) credits, to meet regulatory requirements in lieu of reducing the emissions from their own fuels.

Burning biogas in a bus or turbine still releases carbon dioxide, but the idea is that this process reduces market demand to extract natural gas from the ground and avoids the release of methane, which is a far more powerful greenhouse gas (at least initially). In fact, methane is so much more powerful that under California’s program, “adding one average biogas-powered vehicle to the fleet would produce enough LCFS credits to cover the deficits incurred by 26 similar gasoline-powered vehicles,” according to Aaron Smith, a UC Berkeley economist.

But there’s a problem with this carbon math. California assumes that methane exerts about 25 times the warming effect of carbon dioxide over a 100-year period. That’s not how it really works in the atmosphere, though.

Methane is very powerful, but it also breaks down quickly, generally within a couple of decades. Meanwhile, carbon dioxide builds up cumulatively in the atmosphere—and much of whatever we emit will continue heating up the planet for hundreds to thousands of years.

Dbrand apparently knows how to really whiff a product launch, as we recently saw with the Steam Machine Companion Cube shell. But it also makes the best case out there for the Nintendo Switch 2, which happens to be discounted right now. The Killswitch 2 is versatile, with better build quality than other similar models, snapping around the back and the edges of the console, with Joy-Con slip covers that add a bulkier, more ergonomically-friendly feel to the controllers. The “Travel” bundle that also includes a hard bungee strap-secured cover to protect the screen (that can hold 10 cartridges) plus stick grips is $69.99 at Dbrand, or $71.99 at Best Buy (usually $80).

The Killswitch 2 is a great pick if you want to retain the console’s hybrid use cases, swapping with ease between handheld to docked mode. The case includes a USB-C dock adapter to let you beam the console’s video feed to a TV; it’s a slightly cumbersome solution if you’re aiming to have a tidy entertainment center, but I haven’t seen other brands achieve a more elegant adapter solution yet.

Other good deals to consider

The future of video game preservation just took a major hit. This morning, Sony announced that, starting in January 2028, the company will no longer produce physical PlayStation discs, which means that from that moment on you can only purchase new PS5 games digitally. At the same time, Sony also announced that it’s going to start winding down the digital stores for both the PS3 and PS Vita, helpfully illustrating one of the most pertinent issues with a digital-only future for gaming: Once the stores are gone, so are the games. It’s a terrible blow for preservation of the medium.

In a lot of ways, both announcements seemed inevitable. It’s clear that consumers have largely — though not entirely — shifted to buying games digitally. According to Sony’s most recent financial results from May, around 80 percent of PS5 games sold are digital. Meanwhile, just last week, Grand Theft Auto VI developer Rockstar Games announced that when the next GTA shows up in retail stores in November, it will actually just be a code in a box, with no disc.

Digital games are in some ways more convenient for players. You can prep downloads in advance, keep a big library of games on a single console, and take advantage of frequent sales. There are also some notable immediate drawbacks, like the inability to sell used games or simply share a copy with a friend. From the publisher’s perspective, digital games are more lucrative for Sony and its contemporaries, who no longer have to deal with the costs of producing discs.

But a more hidden cost to the digital shift comes in the form of preserving video games. Given the complex nature of consoles and constantly changing formats, preservation is already a massive challenge. Back in 2023, the Video Game History Foundation claimed that 87 percent of classic games — defined as anything released before 2010 — were “critically endangered.” The reason for the 2010 cutoff? That’s when digital storefronts became prominent, portending an even bigger issue in the future. “Our experiences gathering data for this study suggest that these problems will intensify over time due to a low diversity of reissue sources and the long-term volatility of digital game storefronts,” the report read.

The closure of digital stores isn’t new, of course; just two years ago, Nintendo shut down the storefronts for both the Wii U and the Nintendo 3DS. And while you can still (for now) download games you’ve already purchased, these closures mean that titles that are exclusive to that particular store are essentially inaccessible to everybody else. If you buy a 3DS right now, you have no way to play the classic puzzler BoxBoy.

In the past, this has perhaps seemed like a minor issue, as the number of digital-exclusive games for platforms like the Wii U and PS Vita was relatively small. But with the entire medium moving toward digital, including industry-shaking giants like GTA VI, this problem becomes much more clear. Starting in 2028, every PlayStation game comes with an expiry date; as soon as the PS5’s digital store shuts down, a huge swath of games will simply become inaccessible. That means not just small digital games, but also major blockbusters.

There have been some steps made to alleviate these issues. Most modern game consoles now let you carry your digital library across devices — Xbox has made some notable efforts on this front — and on PC the storefront GOG has an entire program dedicated to keeping old games playable on modern hardware. These are great and necessary efforts, but they aren’t enough on their own. For one thing, plenty still slips through the cracks; mobile games, for instance, are massively popular, but are rarely subject to preservation efforts outside of fan projects. Just as important is the fact that these efforts rely on the goodwill of platform holders. Once they shut down a store, or stop putting efforts into keeping games playable across multiple generations of hardware, the games are once again impossible to legally play. (This isn’t just a gaming problem, by the way; the shift to streaming is having a similar impact on film.)

Physical media isn’t a perfect solution to these problems. Discs and cartridges will degrade over time, and they usually require specialized hardware to even use. But they at least give players and preservationists a greater degree of control over how they’re able to collect, share, and preserve these experiences, without having to deal with the whims of console makers like Sony. It’s already a nightmare trying to keep old games alive — in 2028, the problem is only going to get worse.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Andrew WebsterClose

  Andrew Webster

  Senior entertainment editor

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Andrew Webster

• AnalysisClose

  Analysis

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Analysis

• EntertainmentClose

  Entertainment

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Entertainment

• GamingClose

  Gaming

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Gaming

• PlayStationClose

  PlayStation

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All PlayStation

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

NBCUniversal executives are about to find out whether Peacock will sink or swim in the streaming industry. Now that Comcast is planning to split NBCUniversal, Peacock, and Sky from its broadband and wireless businesses, Peacock will be forced to stand on its own — without the backing of a combined company that pulled in more than $123 billion last year.

Peacock grew by just five million subscribers between March 2025 and March 2026, bringing it up to 46 million. Netflix’s more than 325 million subscribers easily eclipse Peacock’s user base. Even Disney Plus’s 132 million subscribers and HBO Max’s more than 140 million viewers make Peacock seem small in comparison. Part of that is because, unlike other major streamers, Peacock is only available in the US. Comcast co-CEO Mike Cavanagh said in March that the company doesn’t have plans for a global rollout of Peacock, but that may change as the soon-to-be standalone service scrambles for scale.

It’s also taking longer for Peacock to hop the hurdle of profitability — one of the biggest challenges for streamers. Peacock reported $2 billion in revenue in the first quarter of 2026. However, it experienced $432 million in losses, an increase from the $215 million it reported losing at the same time last year. But NBCUniversal media chairman Matt Strauss claims Peacock will become profitable in the current quarter, according to Deadline. “There’s not one way to approach a streaming strategy or market,” Strauss said during the Evercore Global TMT Conference last month. “Sometimes you have to play to your strengths, which is what we’ve been doing.”

It’s not clear how long Peacock can rely on live sports and reality TV to keep its service afloat. The service canceled its hit series Poker Face last year, leaving it without a tentpole series that makes Peacock worth subscribing to, like Severance on Apple TV or White Lotus on HBO Max. Though Comcast co-CEO Brian Roberts and Cavanagh told investors that the company’s split isn’t a setup for a merger or acquisition, it still seems like a possibility.

Peter Supino, a Wolfe Research analyst, said that he expects “one or both Comcast units to merge with peers or competitors,” according to The Hollywood Reporter. Media executives who spoke to Oliver Darcy for his Status newsletter are similarly doubtful about Roberts’ and Cavanagh’s M&A denials, with some insiders speculating that Netflix could make a bid for NBCUniversal’s assets. Either way, Peacock will need to do something more than just tread water, or else a competitor may just have to keep it from sinking.

Update, July 1st: Clarified that Peacock was offered for free to Xfinity X1 and Flex subscribers.

“AI summary does not merely create a record. It creates an authoritative-looking one that is often wrong and, in doing so, it inverts the burden of proof. Once a summary exists, the question shifts from proving what was said to disproving what the machine wrote,” Gogia said. “A tentative ‘we should look at acquiring them’ can harden into ‘we agreed to acquire them’ and that version becomes the default until someone corrects it.”

And, noted Justin Greis, CEO of consulting firm Acceligence, the problem will only get worse as AI summary generators morph into agentic systems, with action-taking autonomous agents.

“Over the next few years, we’ll see AI agents that summarize, extract decisions, assign work, update business systems, prepare follow-up documents, and collaborate with other AI systems after the meeting ends,” he said. “In fact, we are already seeing that integration happen, and it is simultaneously incredibly valuable and outrageously risky. The real question isn’t whether to allow an AI notetaker. It’s how organizations will govern an increasingly machine-readable workplace.”

If you’re not heavily invested in Amazon’s ecosystem, I think the Kobo Libra Colour is a better buy than its closest Amazon rival, the $249.99 Kindle Colorsoft. It offers many of the same core features, including a sharp seven-inch color display that makes book covers, comics, and highlights pop with only slightly less vibrancy than the color Kindle. It’s also waterproof so you can rest easy at the beach, and features adjustable warm lighting for more comfortable nighttime reading.

The Kobo also has a few advantages over Amazon’s e-reader. It supports EPUB and a broader range of file formats, making it easier to read books from a variety of sources, while offering twice as much storage (32GB). Unlike the Colorsoft, the Libra Colour includes physical page-turning buttons, which I find make for a more intuitive reading experience. It also supports Kobo’s optional Stylus 2, letting you annotate ebooks, jot down handwritten notes, use built-in notebook templates, and even convert your handwriting into typed text. Due to its size, I wouldn’t recommend it as a dedicated digital notebook over something bigger like the Kobo Elipsa 2E or Kindle Scribe, but it’s far more practical for quick notes than using your phone. It also works with Instapaper, making it easy to save web articles for offline reading.

Microsoft will likely soon follow Sony and stop the production of physical discs for Xbox games. But instead of leaving physical discs behind entirely, sources familiar with Microsoft’s plans tell me the company has quietly been working on a disc-to-digital feature that will allow Xbox owners to digitize their existing physical game collections.

Xbox employees recently started testing this new feature, after references to “enable Disc2Digital” appeared in the Xbox PC app code in May. I’m told that Microsoft’s disc-to-digital feature will work on Xbox One and Xbox Series X discs only, and not those for the Xbox 360 or original Xbox console.

Getting a digital copy of a game works simply by inserting a compatible disc and installing and playing the game. This will require a Microsoft account on an Xbox console and will grant a digital entitlement for physical games. This digital entitlement is tied to the specific disc, and it will move from account to account if you swap the physical game with a friend or log in to a different Xbox profile and try to play a disc-based game.

The digital entitlement for a physical Xbox game is similar to buying the title from Microsoft’s digital store. If the game is available on Xbox Cloud Gaming and you have a Game Pass subscription, you’ll be able to stream it. If it’s an Xbox Play Anywhere title, you’ll also be able to access it on PCs and handhelds.

The Xbox disc-to-digital feature will also work with discs that come bundled with a console and multi-disc titles, providing access to everything the disc usually offers, including downloadable content. Some Xbox One discs might not work with this new feature, though. “It all depends on how and when the disc was manufactured and it may not have the features we need for this program,” warns Microsoft to its internal Xbox testers.

Discs will still work after they’re digitized, and you’ll only lose a digital entitlement if you loan the disc to a friend or sell the disc to someone else. Microsoft is currently testing this new feature internally, and I expect we’ll hear more about it in the coming months.

I understand Microsoft hasn’t fully finalized whether the next-generation Xbox, codenamed Project Helix, will ship with a built-in disc drive, but if Helix is disc-free, the feature could be essential for existing Xbox owners wanting to digitize their game collections in time for future Xbox consoles.

After a lengthy legal dispute, Krafton has settled with its subsidiary Unknown Worlds Entertainment, which is developing Subnautica 2, and will pay bonuses to the studio’s staff, Bloomberg reports.

The dispute began last year after Krafton pushed out Unknown Worlds’ cofounders, Charlie Cleveland and Max McGuire, and its CEO, Ted Gill, ahead of a potential $250 million bonus for hitting certain financial goals that would have been shared with the studio. After the executives sued, a judge reinstated Gill as CEO in March, and Subnautica 2 finally launched in early access in May. The game passed four million copies sold in five days.

Details of the settlement haven’t been disclosed. “Krafton, Inc., Unknown Worlds Entertainment, and Ted Gill, Charlie Cleveland, and Max McGuire have reached a mutual settlement and agreed to dismiss all pending legal proceedings,” Krafton said in a statement to Game Developer.

Bonuses will be paid to all Unknown Worlds staffers instead of only those who were there when Krafton acquired the studio in 2021, and they’ll be paid out in three yearly installments, according to Bloomberg. However, Gill will be leaving the studio. “We mutually agreed to part ways,” Gill tells Bloomberg. “New leadership is the best way for the studio to move forward.”

Krafton didn’t immediately reply to a request for comment from The Verge.

Cody Spencer, the co-owner of the small games retail chain Pink Gorilla Games, put it well when I asked about the impact of Sony’s recent announcement that it will stop making discs for new games starting January 2028. “It’s sad to see. This decision is only a negative for gamers. We’re losing the ability to sell games, to share games, and to own games.”

Sony’s announcement has been devastating news for many in the games industry. Not just players, but also groups like independent retail stores and preservationists that try and make gaming more accessible.

“This is unfortunate news for those who still prefer buying games on physical media, and is certainly a significant hit to consumer rights, the resale market, and game creators whose businesses rely on the physical market,” Frank Cifaldi, executive director of the Video Game History Foundation, says in a statement.

Boutique publishers are also lamenting PlayStation’s announcement. “We are profoundly disappointed by Sony’s decision to suspend physical games production in 2028,” says iam8bit in a statement. “Physical games are vital to games preservation, ownership, and consumer choice, values that have guided iam8bit since our first physical release in 2016. Our commitment to these values remains unchanged. Long live physical media.” Lost in Cult, in its own statement, says that it aims to “do everything in our power to preserve video games to the best of our ability and will continue to do so for as long as we can.”

But the move isn’t exactly unexpected. For a long time now, video game sales have primarily been digital — just look at Capcom saying that 93 percent of its game sales were digital over its last fiscal year. Still, it’s disappointing for people who want to have games in a more tangible form than as data on a hard drive.

The reality is that, despite the outcry, things aren’t going to be different for most people right away. “Physical sales of new PlayStation 5 games have been declining for some time,” Spencer says. “So immediately after the switch to digital only I don’t think we’ll see much of a change.”

In five to 10 years, Spencer expects to see “increased prices for the physical titles printed before 2028 and a niche but strong demand for our products.” While that’s good for business, “I’d personally rather not have [that] be the case.”

Even further down the road, “the very idea of physical video games will be foreign and more seen as a novelty, which will not be good for us at all,” Spencer says. “Our type of store may be seen more like a record store. A place for largely the most passionate fans of the medium rather than a spot everyone goes.”

Sony has already been inching toward a potential digital-only future for a while now. The PS5, after all, launched in 2020 with a cheaper version without a disc drive, and the PS5 Pro requires a separate disc drive purchase if you want to play physical games at all. And this generation isn’t even the first time Sony has launched digital-only hardware; 2009’s PSP Go handheld didn’t have a UMD drive, Andrew Borman, director of digital preservation at The Strong National Museum of Play, reminded me.

“The challenges of digital preservation aren’t new”

“The challenges of digital preservation aren’t new, nor are they exclusive to the video game industry,” Borman says, pointing to things like required online connectivity, frequent game patches, and “much of the game development process happening only using digital tools.” But he says that “it is important that we act now to preserve the history of the industry.”

Borman says there will always be a market for used and new physical games, pointing to the resurgence of vinyl records. But losing the option of physical PlayStation games still stings. “From a consumer perspective, choice matters, and losing that choice is unfortunate, especially for those who may not have reliable or fast Internet connections — or just like to feel ownership over their purchases,” Borman says.

Cifaldi says that the shift won’t have “as much of an impact as you might expect” to the work of professional preservationists. “The reality is that this continues to be a trend,” Cifaldi says in a slightly different version of the statement sent to The Verge. “Sony PlayStation isn’t the first ones doing this nor will they be the last, as the vast majority of video games produced over the last two decades were not made for dedicated home video game consoles, let alone pressed to physical media. And even when they were released on physical media, a day-one digital patch was all but guaranteed, meaning that even though a disc is preserving data in an accessible way, it may not represent the game that people actually played. Museums and archives like ours have been preparing for this future for a while, with the expectation that putting discs on a shelf isn’t going to be a long-term solution for preserving new games.”

It is worth noting that platform holders are doing some preservation to make their older games available to play. Sony has an IP Preservation team, Microsoft has made a bunch of older games playable on modern Xbox hardware thanks to the backward compatibility program, and the Nintendo Switch Online retro catalog is jam-packed with classics, including titles from as recent as the GameCube era.

GTA VI, which will be sold in physical stores but only as a download code in a box, is likely a preview of what’s to come. Since it won’t be sold on a disc, you can’t sell the game, borrow it from a friend, or come across a cheap used copy for a fraction of the $79.99 price. People might put up with that for GTA VI, one of the biggest games of all time, but it won’t feel so good when it’s the experience for every video game box on the shelf in a store.

Cifaldi is calling on trade groups such as the Entertainment Software Association, which has opposed preservation efforts in the past, to “offer meaningful solutions for archives and museums to legally preserve digital-only content and make it accessible for research.” Cifaldi says that the industry “needs to meaningfully come to the table on this issue, because expecting museums to download a copy of Grand Theft Auto VI and hope it’ll run in 50 years is not a preservation solution.”

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Jay PetersClose

  Jay Peters

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Jay Peters

• EntertainmentClose

  Entertainment

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Entertainment

• GamingClose

  Gaming

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Gaming

• PlayStationClose

  PlayStation

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All PlayStation

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

Apple is working on a “revamped” version of its entry-level MacBook Pro that it could launch as soon as the first half of 2027, Bloomberg reports. The company is also testing four new iPad Pros that are set to launch in the spring with a focus on “internal improvements.”

The updated MacBook Pro, which will keep the 14-inch screen size, will have a design that’s “in line” with what Apple is planning for the touch screen MacBooks it also has in the works, Bloomberg says. Those new touch screen laptops are set to be released between “the end of this year and early next year,” and Bloomberg has previously reported that they will get a Dynamic Island-like pill at the top of the screen.

Apple last updated the base MacBook Pro in October with an M5 chip bump. The company is working on an M6 processor, and Bloomberg says that Apple “finished work months ago” a different base MacBook Pro upgrade that keeps the laptop’s present design and is scheduled to launch this year. Apple will quickly move to the M7 line in 2027, including new Pro and Max chips, Bloomberg previously reported.

As for the iPad Pros, Bloomberg says that they’ll retain 11-inch and 13-inch screens. Apple last updated the iPad Pro line last October with the M5 chip.

Xbox is making some big changes — again. On June 10th, a few months after Asha Sharma took over as CEO, she and newly-promoted chief content officer Matt Booty sent a memo to staff warning of an “Xbox reset.” The business, they said, is facing significant challenges, including a 3 percent “accountability margin,” massively higher component prices for consoles due to the memory and storage shortage, and an “over extended” studio system.

That same day, The Verge and Bloomberg reported that the Xbox division was planning layoffs in July. The cuts, expected to be announced on July 6th, could include studio closures or spinoffs. The Verge’s Tom Warren reported on June 30th that Microsoft is looking at closing at least five studios and potentially canceling games like Blade.

Read on for all of our coverage of the shifts at Xbox.

The device was “slimmer than an iPhone,” and they were told it would run on a Qualcomm Snapdragon chip, according to the WSJ. It would also reportedly use its own AI-enabled operating system with features powered by the SpaceX-owned xAI.

Last week, the Financial Times reported that SpaceX chief operating officer Gwynne Shotwell told investors that the company is considering launching a mobile service in the US that would connect to its Starlink satellite network. Meanwhile, some analysts speculate SpaceX could acquire T-Mobile. Starlink is currently SpaceX’s only profitable business.

Musk has brushed off the idea of creating a phone in the past, but didn’t rule it out completely. During an event in Pennsylvania last year, Musk said, “the idea of making a phone makes me want to die,” but added, “if we have to make a phone, we will, but we will aspire not to make a phone.” In February, Musk said SpaceX is “not developing a phone” after Reuters reported that the company is working on a phone for Starlink.

1 The US has lifted restrictions on Anthropic’s Mythos and Fable models
Anthropic said it would begin restoring access today. (NYT $)
+ The US had imposed controls over security concerns. (Bloomberg $)
+ It lifted the restrictions after lengthy talks with Anthropic. (BBC)
+ But the crackdown has already opened doors for Chinese AI rivals. (CNBC)

2 The most detailed survey of the universe ever is now underway
It’s using the largest digital camera on Earth. (New Scientist $) 
+ The project is based at the Vera C. Rubin Observatory in Chile. (NYT $)
+ It aims to transform our view of the cosmos. (MIT Technology Review)
 
3 Tech talent is fleeing the US due to H1-B visa chaos
They’re eyeing relocation to Canada, the UK, or the Gulf. (Rest of World)
+ While China is poaching AI talent from the US. (CNBC)
+ Visa rules are also affecting young scientists. (MIT Technology Review)
 
4 Trump raked in more than $1 billion from crypto businesses in 2025
He reported $635 million in royalties from a Trump meme coin. (BBC)
+ The rest largely came from his World Liberty Financial venture. (The Hill)
 
5 The UN warns that the rapid spread of AI may worsen global inequality
It’s proposed a shared framework for responsible AI development. (Guardian)

6 Companies are making LLMs talk like a caveman to curb AI spending
A senior OpenAI employee contributed to the “caveman” project. (404 Media)
 
7 Babies are born with the neural foundations for math
Brain recordings have identified the mechanisms. (New Scientist $)

8 An independent studio has bought the OpenAI movie Amazon dropped
Neon has purchased “Artificial,” which focuses on Sam Altman. (NYT $)
+ Amazon had dumped it after investing in OpenAI. (Gizmodo)
+ The depiction of Altman is reportedly unsympathetic. (Variety)

9 AI has re-created Gene Wilder’s voice for a new “Willy Wonka” series
Wilder’s wife said his estate is “delighted” with the new show. (NBC News)
+ Netflix partnered with AI company ElevenLabs on the project. (The Verge)

10 NASA aims to send a spare Mars rover—and soccer ball—to the moon
The nuclear-powered “Promise” may help establish a lunar base. (NYT $)

Quote of the day

“Caveman save you token, save you money.” 

—The GitHub repository for the “caveman” plugin explains how the project curbs AI spending by turning verbose LLM outputs into concise text.

One More Thing

SELMAN DESIGN

AI is dreaming up drugs that no one has ever seen. Now we’ve got to see if they work.

On average, it takes more than 10 years and billions of dollars to develop a new drug. A growing number of startups are betting that AI can make the process faster and cheaper. 

By predicting how potential drugs might behave in the body and discarding dead-end compounds before they leave the computer, machine-learning models can cut down on the need for painstaking lab work. 

Yet it is still early days for AI drug discovery. A lot of AI companies are making claims they can’t back up—and the technology is not a panacea. But the technology is beginning to move from promise to practice.

Find out how AI is speeding up drug discovery.

Even if a catastrophic blackout doesn’t occur, other problems could still hinder access. Political scientist Amélie Férey explained on France Culture how license prices could gradually rise and drive up costs. Or access to certain features could be gradually restricted. As García de Viedma put it, a moratorium on model access would create a “temporal asymmetry.” Disruptions can occur in “the complex supply chain behind all our technology,” through export controls (something now happening in the chip market) or through the degradation of essential services (such as what could happen with Starlink coverage).

In recent years, the European Union has entered a race in that arena, one that involves symbolic measures well as practical legislative moves. The European Parliament has dropped Google as its default search engine and replaced it with the French Qwant. And in early June, the Commission presented its European Technology Sovereignty Package, which addresses, among other issues, AI. They aim to ensure that Europe becomes “a continent of AI, strengthen its digital autonomy, and help build a more sustainable digital future,” while also acknowledging Europe’s technological dependence.

“We cannot afford to depend on others for the technologies that keep our hospitals running, our energy grids stable, and our services secure,” said Commission President Ursula von der Leyen. Specifically, Europe aims to triple the capacity of its data centers over the next five to seven years, boost the adoption of AI, enhance research and innovation, and work on its own development and deployment efforts. The package will now have to go through an approval process to become law and take effect.

The logic is simple. If an AI system can find a flaw for one user, it can identify vulnerabilities for other users; that’s a benefit for well-resourced attackers prepared to move fast once a vulnerability becomes public. Hackers are, after all, one group in tech that really doesn’t worry much about moving fast and breaking things.

Apple said it has no evidence any of the patched vulnerabilities were actively exploited before the fix shipped, but that’s not a reason for complacency — especially in business and education, where deployment tends to lag.

Caution has become risk

That lag is a vulnerability. Traditionally, some industries — particularly, regulated ones — have delayed the installation of system upgrades. That was sensible because it enabled them to identify compatibility problems before a company-wide deployment took place. 

All sorts of helpful Docs options are accessible right from your home screen.

JR Raphael, Foundry

You can always get to those by long-pressing the Docs icon, but if you find yourself using the functions often, you can make it even easier by pressing and holding one of ’em within that pop-up menu and then dragging it directly onto your home screen for one-touch access.

You could even build yourself a nifty little Docs command center for super-fast access to all the stuff you use the most:

Docs, Docs, everywhere — so many options, never more than a tap away.

JR Raphael, Foundry

And while we’ve got easy access on our minds…

Google Docs Android feature #10: The offline on switch

By default, the Docs Android app will make any files you actively work within the app available for offline use for a while — but if you’re getting ready to travel or expecting any other connectivity-challenged moments, you don’t have to rely on its judgment to make sure your stuff is accessible even without internet access.

From the main Docs screen, tap the three-dot icon alongside any document name and then look for the “Make available offline” option within the menu that pops up.

Pro tip: Turn offline access on before the need actually arises.

JR Raphael, Foundry

Tap that for any document that you expect to need and then rest easy knowing it’ll be there and available for you — no matter your current connection status.

Google Docs Android feature #11: Wordless reactions

Sometimes, a picture really is worth a thousand words. Or at least a couple hundred.

That’s especially true when collaborating on a document and expressing your opinions — which, let’s be honest, often come down to simple reactions like 👍 or maybe 💩.

Docs has allowed emoji reactions as a part of its editing process for a while now, and at some point along the way, the Android app gained the same ability. It’s just weirdly tucked away in a place where few word-minded mammals would ever find it.

So do this: The next time you’re working on a shared doc, try pressing and holding your finger onto any word to highlight it. (You can then use the selector icons that pop up to expand or shift your selection, if needed.)

Now for the tricky part: In the menu that appears alongside your selection — the one that contains “Copy” and other such commands — look for the three-line icon at its far right side.

See that little three-line icon within the text actions pop-up?

JR Raphael, Foundry

Tap that — and lookie what we have here: the awkwardly hidden option to add an emoji reaction! 🥳

Emojis for everyone — hip, hip, hoorah!

JR Raphael, Foundry

Hit that sneaky little thing with all your might, then select the most appropriate reaction and move on with a satisfied 😊 in your mind.

Google Docs Android feature #12: Your in-doc AI

Generative AI these days is a bit of a mixed bag, to put it politely. Google’s Gemini and other such services are arguably causing more harm than good, on numerous levels, and also just creating paths for lazy, low-quality and accuracy-challenged work.

But in the right scenario and with the right sort of framing, Gemini-style AI can actually be useful. The onus just falls squarely on you to determine how to most effectively use it and avoid falling into the traps of unoriginality or, worse, inaccuracy.

The Docs Android app now offers a direct shortcut to Gemini within its editing interface — via the starburst-shaped icon in the toolbar at the top of the screen — and with some careful considering, it might just end up being a helpful reading or editing tool for you.

A few suggestions that notably don’t involve having AI write lazy, uninspired copy on your behalf:

• You can use the Gemini in Docs system as a quick ‘n’ easy way to get a definition or list of synonyms for any word in front of you.
• You can also use it to ask for context or related information — like an integrated research aide. (Just remember that AI doesn’t always get things right, so treat it as more of a starting point than a final quote-ready answer.)
• And you can lean on it to perform tasks like summarizing or outlining a long document or helping you reorganize a document into a more logical state.

Gemini is now available directly within Docs. Please, use it wisely.

JR Raphael, Foundry

You may still end up spending a ton of time in Docs, but at least now you’ll make the most of every second there and avoid wasting your effort on piddly little tasks that can be made more efficient. And that, as far as I’m concerned, warrants an enthusiastic 🥂 reaction — maybe even followed by a well-earned 🍪.

Get six full days of advanced Android knowledge with my free Android Shortcut Supercourse. You’ll learn tons of time-saving tricks for your phone!

Springboards realized that parameters were blunt instruments for what it wanted to do. It does not make sense to dial up the randomness across the board; you only want to boost it at specific points in its output, he says.

For example, when you ask a chatbot “Where should I go in Europe?” the model only needs to tweak the randomness just before it names a destination, not for every word in its response.

To make Flint do this, Springboards trained its version of Qwen 3 to identify the points in its output where more variety was possible and fill those spots with words or phrases that were a little more random.

“Flint’s programmed to throw an oddball in. It’s more of an invitation to think wider,” says Maximilian Weigl, cofounder and chief strategy officer at Uncommon, a marketing firm. “That’s super interesting.”

Weigl’s team uses Flint alongside ChatGPT, Claude, and Gemini. “You can’t really create something boundary-breaking with tools that pull you back to the average,” he says. 

And yet Weigl notes that nine times out of 10 the average is fine. You don’t always need to reach for extremes with something like Flint, he says: “Most people are fine with good enough. They want to see mass-market familiar things.”

Weigl also cautions against using any LLM too much. “I have a big problem when people rely on the output from any AI, including Flint,” he says. “If I saw people on my team copy-pasting something from AI, I’d be like, ‘That’s not your job! Think, talk to other people, use your own voice.’”

For now, Flint is aimed at advertisers and marketers because those are Springboards’s customers. But Bingemann and Browne insist that a lack of variety is a problem for anyone using chatbots.

The idea is to give people the choice and leave it to them to decide if the result is good or not, says Bingemann. “Variety is great when you’re trying to spark ideas,” he says. “Let’s go down this route instead of letting the machines do it all and ending up in a gray, boring world.”

Anthropic proposes a common jailbreak framework

Anthropic used the announcement to call for an industry-wide framework to assess AI jailbreaks, saying developers and governments currently lack a common standard for evaluating newly discovered techniques.

“There’s currently no consensus in the AI industry on how to describe, in objective terms, the severity of an AI jailbreak,” the company said. Anthropic said it is working with Amazon, Microsoft, Google, and other Project Glasswing partners on a framework for evaluating jailbreaks, while also expanding collaboration with the US government through pre-release testing of future frontier models, information sharing, and joint AI security research.

“Government involvement in AI releases requires a durable, transparent process that gives cyber defenders and others the certainty they need about access to powerful models,” Anthropic said. “These rules should be codified in strong regulation and applied equally across frontier model developers.”

It’s the end of an era for the PlayStation store on PS3 and PS Vita, with Sony now planning to shut down its digital distribution service on both consoles. The PlayStation store on PS3 will close in select markets later this year, including Mexico, Honduras, and Nicaragua starting in August, with “additional Latin American and Middle Eastern countries” to follow in late 2026.

The shutdown across both PS3 and PS Vita “in all other countries” will take effect in July 2027. After this date, Sony says that console owners will be unable to purchase new content, but can continue downloading games they’ve previously purchased “for the foreseeable future” after the deadline passes. That means if you haven’t already purchased specific titles for the aging consoles — the PS3 and PS Vita turning 20 years old and 15 years old respectively this year — you still have some time to do so.

“We know this news may be disappointing to PS3 and PS Vita players who hold a special place in their hearts for this generation of gaming,” Sony said in its announcement. “As we continue to expand the PlayStation experience on newer devices that most of our users are playing on today, we need to focus more resources on delivering the best gaming experiences on these platforms as we look ahead towards the future.”

This announcement comes as Sony is also gearing up to cease all production of physical PlayStation discs for games released after January 2028 — which begs the question of how releases will be preserved in future given the finite lifespan of digital stores.

TheThe storm developed quickly over west central Illinois on April 17th, first as a single high-intensity system called a supercell, and then later that evening transforming into a long squall line of thunderstorms. Tucked inside a wall of wind were several smaller, quick-forming tornadoes, one of which landed directly on Rivian’s electric vehicle factory on the outskirts of the college town of Normal. The ferocious storm knocked down one of the walls and tore through the plant’s roof like a can opener peeling back the lid.

Bobby Dean Parker, vice president of manufacturing at Rivian, was at home in Normal when he got the call from someone at the factory. Parker, an affable Southerner, had already experienced his fair share of powerful Midwestern thunderstorms despite only being on the job for six months. But he quickly realized that this one was different. “Hey, it’s not good,” the person on the phone told him. “Oh shit,” Parker thought, before dashing out into the still-stormy night.

Thousands of miles away at an event in Southern California, Rivian CEO RJ Scaringe’s phone suddenly lit up with a flurry of texts. At first, information was scarce. Was anyone hurt? How bad was it? Scaringe was reluctant to send anyone inside to assess the damage without knowing the structural integrity of the building. Then someone sent him a video.

Even in the dark, you can immediately see a giant hole in the factory’s roof. The power is out, so floodlights illuminate the wreckage. Water can be heard rushing in the background; the tornado triggered the plant’s sprinkler system, causing the pits under the assembly line to flood. The concrete floor is also slick with water. The damage appears quite extensive.

Scaringe was stunned; his factory suddenly had what seemed to be its own sunroof.

April 22nd, just days after a tornado ripped a hole in the side of the Rivian factory.Photo: Anthony Tahlier / The Verge

For Rivian, the tornado couldn’t have hit at a worse time, or in a worse place. Not only did it strike right as the company was poised to roll out its all-important, more affordable R2 vehicle, the storm landed directly on the section of the factory where production for the midsize SUV was set to begin in a matter of days.

Since bursting on the scene eight years ago with a pair of exciting, luxury-priced EVs, Rivian has cultivated an impassioned fan base and earned a lot of goodwill in the industry. And while the company appears to have done many things right in its short history, all of its hopes for the future are currently pinned to the R2 model. These next few months are especially crucial. Rivian needs to ramp up production for the R2, hit all of its targets, and get the vehicle to customers in what is arguably the harshest climate for electric vehicles since the first battery-powered models arrived nearly two decades ago.

Unlike many of the forces currently bearing down on Rivian, threatening its very existence, all the tornado did was blow a hole in its factory. The roof can be repaired, the debris swept away, the pits pumped out. But as for the myriad of existential challenges currently facing Rivian, the fix won’t be so easy.

The Rivian factory in Normal, Illinois.Photo: Anthony Tahlier / The Verge

TheThe first attempts at modern, commercial electric vehicles in the US date back to the ’90s. These were largely compliance cars, built by major automakers to satisfy California’s strict emissions regulations but not intended for wide release. The compliance cars were followed by a string of uninspiring mass-market models like the Nissan Leaf and Chevy Volt, a pair of egg-shaped economy sedans that projected an electrified future that was inoffensive and characterless — a car for people who don’t want to think about cars.

But Tesla took an entirely different approach with its Roadster, followed by the Model S and Model X, which would become the first legitimate and commercially appealing EVs. The S was a luxury sedan, and though it had to convince buyers that the quiet hum of an electric engine was more powerful than the roar of an internal combustion engine, Tesla had smartly targeted a consumer that actually wanted a car that looked like a computer — and would spend a lot of money on it.

Things really got rolling after Tesla released the Model 3 in 2017, and then the Model Y in 2019. Four years later, the Model Y would become the best-selling car in the world, a title it reportedly retained for the next two consecutive years.

EVs were originally thought of as a niche product, something for virtue-signaling bevested Silicon Valley types high on their own supply about “changing the world,” but hardly as a solution for the perils associated with climate change. It wasn’t until Elon Musk started framing Tesla’s mission around accelerating the world’s transition to sustainable energy that EVs came to be viewed as a critical piece of the puzzle. But Tesla’s success was never a foregone conclusion. After all, no new American car company had ever emerged to challenge Detroit’s Big Three (Ford, GM, and Chrysler/Stellantis) in achieving sustained, profitable volume production in nearly a century.

Musk’s unprecedented success kicked off a global race for electrification. Legacy automakers were pushed to pivot their long-term strategies, restructure their production lines, and build ground-up EV platforms to survive. In China, companies like BYD, Nio, Xpeng, and Li Auto quickly adopted Tesla’s EV- and software-first business models. And a new class of EV startups emerged, aiming to copy Musk’s luxury-to-mass-market rollout strategy.

One of those companies was Rivian. Founded as Mainstream Motors in 2009, the company would go on to great acclaim for its adventure-themed R1T and R1S electric truck and SUV, respectively. If Tesla had produced the first appealing EV sedan and crossover, Rivian had done the same for the pickup and full-size SUV. An early $700 million investment from Amazon alongside an impressive order for 100,000 electric delivery vans helped solidify the company’s reputation as one of the more serious players. Unlike most of the other EV startups to emerge at the time, Rivian seemed to have a bright future. While most companies, Tesla included, focused on sedans and hatchbacks, Rivian targeted the booming truck and SUV market. Branding itself as the “Patagonia of pickups,” Rivian sought to appeal directly to affluent, outdoorsy types.

But in the years since Rivian first emerged, the mood around EVs has soured dramatically. Charging woes, range anxiety, and sky-high luxury prices kept most people at arm’s length from the EV market. Long before Donald Trump was elected the second time on a promise to eliminate the fictitious “EV mandate,” clouds were looming on the horizon.

The Rivian R1T seen at the New York International Auto Show at the Jacob K. Javits Convention Center in New York City, April 17th, 2019.Photo: Michael Brochstein/SOPA Images/LightRocket via Getty Images

It’s undoubtedly a weird moment. On one side, Rivian is unveiling the new R2 to massive anticipation, Tesla is still pumping out Model Ys at scale, and EV adoption numbers are continuing to grow globally — thanks in part to a tidal wave of affordable Chinese exports. On the other side, you’ve got an avalanche of cancellations, pauses, and write-downs that makes the whole thing feel like it’s falling apart.

Tesla’s Model Y continues to be a certified global phenomenon. Despite Musk’s sky-high unpopularity and the company’s various missteps, Tesla has outperformed better-capitalized competitors like Volkswagen and Toyota to sell over 1 million Model Ys in 2025 alone, proving that there is a real appetite for well-designed, spacious, zero-tailpipe-emission vehicles.

While Tesla soaked up the spotlight, Rivian was struggling to maintain its foothold, losing billions of dollars every year while trying to keep up production of its R1 vehicles. Then, earlier this year, the company rolled up to SXSW with its first pre-production R2s, a genuinely attainable family car that starts at under $50,000, built by a company that a lot of people seem to genuinely love and that has figured out how to make desirable EVs at scale, even if it’s still figuring out how to do so sustainably. The buzz around Rivian began to swell. Fans were convinced this would be the company’s Model Y moment as they flocked to get one for themselves. Rivian quickly racked up an estimated 200,000 reservations for the R2 — falling short of Tesla’s 450,000 reservations for the Model 3, but still an indication of widespread excitement.

A Rivian R2 electric vehicle drives over a dirt obstacle course set up on Congress Avenue during SXSW in Austin, Texas, on March 12th. Festivalgoers could ride in the vehicles as part of Rivian’s Electric Joyride demonstration.Photo: Jay Janner/The Austin American-Statesman via Getty Images

The R2 will arrive at a particularly brutal moment for EVs. Over a dozen automakers are walking away from their EV plans. Ford canceled the F-150 Lightning. Honda killed its Zero series of electric vehicles, as well as its Afeela collaboration with Sony. Mercedes reined in its EV output. VW idled the ID.4 and ID.Buzz. Nissan axed the Ariya. Many of those aforementioned EV-only startups, like Fisker, Canoo, Lordstown, Proterra, Arrival, and Sono Motors, ended up stillborn. It would seem that the road to the EV future is lined with sinkholes.

Not even Musk and his indelible brand have been immune. Tesla’s annual sales have dropped two years in a row — but Rivian’s decline was even steeper. In 2025, the company only sold 42,247 vehicles, an 18 percent decrease compared to 2024. It’s never earned an annual profit, and continues to lose upwards of $8,000 in gross profit for every vehicle it builds and delivers, according to its earnings reports.

Globally, the auto industry has booked over $70 billion in write-downs related to their EV investments in the past year. The reasons are a mix of things converging at the worst time. President Donald Trump, a longtime friend and beneficiary of the fossil fuel industry, eliminated the $7,500 federal EV tax credit just as EV demand in the US was really taking off. Tariffs made the import of foreign EVs, especially affordable Chinese models, nearly impossible. Regulatory credits, which were a lifeline for pure EV companies like Tesla and Rivian, netting them hundreds of millions of dollars in revenue each quarter, are also kaput. We’re in an era of extreme price sensitivity for consumers, and many electric models are still just too expensive. And a lot of automakers built EVs that were rushed, not very good, or gas models retrofitted to run as electric.

The stakes for Rivian are huge. Starting at around $45,000, the R2 is designed to catapult Rivian to the next level of automotive success. If it’s a hit, then Rivian would be well positioned to challenge Tesla’s status as the most important EV company in America. A successful R2 would beget the R3, and more importantly the R3X, a rally-inspired off-roader with the playful, nostalgic design of an ’80s hot hatch. The R3X has been getting nothing but praise since it was first revealed, helping the company earn an enormous amount of goodwill that it’s hoping to sustain long enough to make it to production. But will it ever come out? Everything depends on the success of the R2.

R2 models unveiled on the floor of the Rivian factory in Normal, Illinois. Photo: Anthony Tahlier / The Verge

“This is arguably their biggest make-or-break moment yet,” Ed Kim, president and chief analyst at AutoPacific, tells me. “This is where Rivian goes from offering a small lineup of very expensive, large EVs that can get up to the six-figure range, down to a truly mainstream-priced EV that starts in the $40,000 range — which is actually well under the median price of a new vehicle these days, which is about $50,000. So here we have Rivian’s really first big volume play happening here.”

If the R2 flops, the company’s struggles would deepen. Rivian is coasting on its buzz and fan love, even though a wealth of technical issues and hardware faults have labeled its R1S SUV as one of the least reliable vehicles on the market. Like every other car company, Rivian is attempting to navigate an increasingly tumultuous and unpredictable global supply chain. The company is burning nearly hundreds of millions of dollars a quarter as it stands up its R2 production line, while breaking ground on a new factory in Georgia. It’s pushed back its expectations for profitability to account for massive investments in AI and autonomous driving technology. And the R2 is about to enter an auto market beset by price anxiety and an outwardly hostile regulatory environment.

Rivian’s staunch supporters can’t help but inject more hype in the moment. “I am really excited about the R2!” John Krafcik, the former CEO of Waymo who now sits on Rivian’s board of directors, tells me. “I think it can reset the entire auto industry the way the XJ Cherokee did in the ’80s.”

Image 1 of 4

The Rivian Factory in Normal, Illinois, on April 22nd.Photo: Anthony Tahlier / The Verge

The Rivian Factory in Normal, Illinois, on April 22nd.Photo: Anthony Tahlier / The Verge

The Rivian Factory in Normal, Illinois, on April 22nd.Photo: Anthony Tahlier / The Verge

The Rivian Factory in Normal, Illinois, on April 22nd.Photo: Anthony Tahlier / The Verge

More than just this EV, the promise of American industry is that innovation breeds successful businesses. There is a basic logic: Make a good product and people will buy it. It’s an idea that is easy to grasp but hard to execute. Rivian has demonstrated its prowess as a low-volume, premium company. But the headwinds — policy shifts, regulatory overhang, tariffs, rising costs, competition from China — are proving to be strong, perhaps even more powerful than a tornado.

FiveFive days after the tornado hit, I’m standing on a catwalk surveying the wreckage of Rivian’s factory. The hole is still there, though now there’s a makeshift plywood fence surrounding the damaged area. The extent of the destruction is immense and staggering to see in person. A huge section of the roof has collapsed. Much of the outer wall is gone, giving me a full view of the road outside the factory, green grass, and beyond that, cornfields. A remaining piece of roof is dangling precariously dozens of feet in the air. Other pieces lie twisted in obscene shapes on the ground.

A yellow excavator works slowly outside to clear the area. It’s very much still an active cleanup scene. Rivian plans to install an autonomous system in the excavator to help speed along the cleanup process, Tony Sanger, VP of production facilities, tells me.

The tornado tore a hole right above the logistics section, where trucks drop off equipment and various supplies — basically a loading dock. After photos of the damage surfaced on Reddit, commenters obtusely noted the absence of any completed vehicles, as if that somehow proved that Rivian was behind schedule on the R2.

But Scaringe insists that the company is still on schedule and has no plans to lower its guidance on the number of vehicles it expects to sell this year. Which is not to say that nothing important was damaged. They’ll still need to check on nearly 300 conveyor drives that were flooded by the sprinkler system to see if any need to be swapped out.

Despite these setbacks, Scaringe feels like they dodged a bullet. He says he feels extremely lucky the tornado landed when and where it did, and not, say, on top of the actual R2 assembly line during work hours. While the damage is significant, it’s not insurmountable. Rivian will bounce back — is presently bouncing back, he argues. The factory lights came back on April 21st for the first time since the storm. The pits have been pumped out. The floor has been dried and swept. The R2 is ready. Full steam ahead.

“We could have come in and been somber about it,” Scaringe says. “But everyone’s just like, ‘Okay, let’s go fix it.’”

Hours earlier, I was standing on the factory floor next to one of the first fully assembled R2s, presented in a sophisticated blue color that the company calls Catalina Cove. For such an important vehicle, the R2 is surprisingly unassuming. It’s smaller and lighter than Rivian’s flagship R1 vehicles, since it was designed to be significantly more efficient and cost-effective to manufacture. The R1T and R1S both start in the low-$80,000 range, but can be optioned up much higher. The R2 starts at $45,000, though that version won’t be available until late 2027. Still, it’s definitely a Rivian, right down to the friendly oval headlights and flashlight tucked in the door frame.

Much of Rivian’s factory floor was unharmed by the tornado.Photo: Anthony Tahlier / The Verge

A few years ago, a meme circulated featuring two dozen or so white SUVs, all facing the same way and looking practically indistinguishable from each other. The implication was that all modern cars look alike, and as a result, individuality and design in automobiles was dead. You can argue with the veracity of that claim, but one of the reasons Tesla made such a splash in the mid-2010s was that it went in a different direction, designing its cars for sleek, aerodynamic efficiency, especially at a time when vehicle profiles were aggressively boxy. Nothing else really looks like a Tesla.

By contrast, Rivian borrowed heavily from other automakers, like Jeep and Land Rover, but with a rounder, softer edge. Overall the company’s design ethos is decidedly friendlier, with the aforementioned oval headlights and a plethora of Easter eggs. On the R2, right next to the windshield wiper fluid cap, there is a small molded engraving of a cross between a frog and a skunk — a “frunk,” also the nickname for the front trunk that is common in EVs. The company’s vehicles do a good job blending midcentury modern design aesthetics with the utilitarian needs of an off-roader and the clean minimalism most often associated with lifestyle brands. But there’s also a bit of Star Wars in there, even some WALL-E and solarpunk. It’s what endears Rivian to so many people.

At 8:30 in the morning, the factory floor is teeming with Rivian workers, most of whom are wearing matching olive green T-shirts that read “Keep Normal Adventurous Forever.” OneRepublic’s “I Ain’t Worried” is blasting over the sound system. Employees are encouraged to take as many selfies with the R2 as they want. The mood is understandably festive. They just got hit by a tornado, but they ain’t worried ’bout it.

Scaringe grabs the microphone and begins by thanking everyone for their hard work, especially in the aftermath of the storm. With his thick frames, neatly parted dark hair, and a tall, athletic build, Scaringe resembles Clark Kent — but Steve-O from Jackass is also a fair comparison. Today he’s traded his traditional military-style khaki shirt for a Rivian-branded dark blue button-down. Dark jeans and a pair of On Cloud waterproof hiking boots complete the look.

RJ Scaringe speaks to workers at the Rivian press event at the factory in Normal, Illinois.Photo: Anthony Tahlier / The Verge

“The resilience that we’ve demonstrated as an organization over the last couple of days is emblematic of the strength of the organization we’ve demonstrated in the development of this product,” he says to polite applause.

He’s not a magnetic public speaker and seems a little stiff at times. Still, he runs through his prepared remarks at a steady clip. There’s an air of “let’s make this quick.” There are cars to build.

Javier Varela, Rivian’s COO, tries to amp the energy levels by repeatedly shouting “Big volumes” and “Go R2!” in his raspy Spanish accent. By contrast, Bobby Dean Parker, the VP of manufacturing, is all Southern charm. Parker, who cut his teeth at BMW, Toyota, Volkswagen, and Tesla, gets arguably the biggest round of applause from the assembled employees. He was the one who stood shoulder to shoulder with them over the last few days during storm cleanup.

“Today, I’m not letting the tornado steal the spotlight from our wonderful R2 and our wonderful team,” Parker says. “But what I will say is praise God that we’re not standing here today in a moment of silence.” He pivots to the R2, which he describes as both “thrill on the road” as well as “luxury on the road.” “So, whaddaya say? Let’s get it on the road.”

Before that, the R2 needs to get off the assembly line. A row of three R2s await Scaringe, Varela, and Parker, one for each. They drive slowly off the line while high-fiving with the workers gathered along the sidelines. Varela honks his horn in celebration, while someone shouts, “Come on Bobby Dean, give it some!” Not that there’s any room for acceleration. Nor would it be safe to floor it while surrounded by hundreds of people.

Rivian COO Javier Varela addresses a crowd of Rivian workers and press at the unveiling event of the model R2.Photo: Anthony Tahlier / The Verge

One of those people is Hemanth Atthipatla, a tall, friendly quality manager in the operations and manufacturing division. He calls himself a “gatekeeper” in charge of policing the R2, making sure it meets all of its quality checks before it reaches customers. Atthipatla was also one of the first Rivian employees to buy his own R2, which is on display at the factory today. His wife loves it, he reports, as does his newborn child, though it’s unclear how a baby expresses approval about an automobile.

“I think if we are able to maintain the good quality and ramp up, and provide the vehicles to the customer, I think it’s going to be a deal-breaker for Rivian,” Atthipatla says.

Brittany Clauss stands out from the crowd with her neon green hair and matching cheek studs. She’s been working at Rivian for four years, first on the R1 line and now over to R2. The main difference, she says, is that the R2 line is more compact, which allows for easier communication between the various stations and workers. She works at the end of the line, checking each vehicle for defects before it goes off to the next stage.

She doesn’t own a Rivian herself, but says she can’t wait to see the R2 out on the road. “I’ll be able to point them out and be like, ‘That’s what I’ve been working on,’” she says.

NearlyNearly 900 miles east of Normal, the center of the Rivian universe is in the parking lot of Campmor, off Route 17 in Paramus, New Jersey. At least for a dozen or so hardcore fans who are gathered there on a brilliantly sunny Saturday in early April.

Romy Vreeland is a nonprofit executive with graying blonde hair and a welcoming smile. She’s also the national secretary of the Rivian Clubs of America and host of today’s meet and greet in the Campmor lot. Rivian Club members go off-roading together, share tips about the best camping equipment, and generally bask in a shared appreciation of the Rivian brand. They like the brand, the cute-but-rugged aesthetic, and the fact that the company’s CEO isn’t a household name, let alone a lightning rod for criticism.

“It’s like, it’s not necessary to know [his name],” she says of RJ Scaringe, a man she’s met only once and found to be lovely. “It’s not a liability.”

Today is one of their semi-regular meetups: a chance for Rivian owners to get together and talk about their favorite electric adventure vehicle, while hopefully converting a few newbies. Jeep owners have their rubber ducks. Tesla fans have their fixation on Elon Musk. Rivian owners… are still trying to come up with their own identity. They love the environment, but also are wealthy enough to afford a brand-new $80,000 electric vehicle. (It’s not dissimilar from early Tesla owners.)

Vreeland is unloading boxes of seltzer and donuts for the meetup. She likes to joke that the Rivian community could at times resemble an enthusiastic social club but admits she frequently wonders how she’d become so immersed in it all. She does love the outdoors, camping, and cooking over an open fire. But she doesn’t consider herself a car person. After all, it isn’t always safe for women in these EV-themed forums online. Early on, she would hide behind anonymity when chatting with other Rivian fans on Reddit or Discord.

But she made plenty of friends, like Rod Gillis, the beanie-wearing finance guy from Mahwah who also owns his own fitness business on the side. Or Steven Santora, who works in real estate, loves to watch EV content on YouTube, and has cash deposits down on three upcoming EVs, including the R2. These are New Jersey’s most unapologetically devoted Rivianiacs.

“It’s a vehicle for someone who is old enough to afford it, but young enough to have fun with it,” Gillis says. “It’s a vehicle that can be a jack of all trades. You can go off-roading with it.

You can take it to the track. You can have fun with it, but if we’re being candid, it’s not the least expensive option out there.”

The R2 could change all that. A $45,000 midsize SUV will naturally bring more people into the world of Rivian ownership. Not everyone will join the club or seek out camaraderie with other Rivian owners, and that’s okay, Vreeland says. Rivian owners used to wave when they would see other Rivians on the road. She’s not expecting that tradition will last in the company’s new era. “Normal car buyers,” she calls them.

Whether those normal car buyers will actually consider the R2, much in the way they did the Tesla Model Y, will be closely watched by members of the Rivian Clubs of America — as well as investors, industry stakeholders, and everyone with a desire to see whether this country can still compete with China. The federal EV tax credit is gone, but the rise in fuel prices is leading some people to give EVs another look.

Vreeland knows Rivian isn’t out of the woods yet. She reads the news, hears the stories about EV cancellations and financial headwinds, and hopes for the best. The message boards are split over various predictions for the future: Maybe some combination of R2 success plus investment from Volkswagen and now Uber will be enough to put the company on the path to a brighter future. She’s not convinced that autonomy and AI will be the saving grace — it just needs to be a great car.

“You read these dire headlines, like the EV industry is dying, whatever, that can make you feel pessimistic,” Vreeland says. “But then when you get into the car and you understand it better and you understand, ‘Wow, this is an American company, like how did they do this?’”

WhenWhen Rivian was created over a decade and a half ago, it looked almost nothing like it does today. In fact, it didn’t even intend to make electric vehicles. Scaringe founded Mainstream Motors in 2009, naming the company after his father’s engineering firm based in his hometown of Rockledge, Florida, a bedroom community for NASA workers. The elder Scaringe started his own company in his 30s. His son, fresh out of MIT with a PhD in mechanical engineering, founded his at 26. He was the sole employee.

Mainstream Motors, which soon became Avera Motors, was unwittingly following the Tesla playbook by setting its sights on a fast, four-seat coupe with rear-wheel drive and extreme fuel efficiency. If they could find success with a fun-to-drive sports car, perhaps mainstream automotive success would follow. With money from friends and family, as well as $2 million seed investment from the Florida Energy and Climate Commission, Scaringe hired about 15 designers and engineers from a raft of other car companies, including Mitsubishi, Chrysler, and Mercedes-Benz, and set to work on their first prototype to show to potential investors, according to a 2010 interview with Scaringe in Hagerty. The company was so cash-strapped that the weekly snack budget was only $35, and employees frequently housed interns to save money, according to a 2021 article in Bloomberg.

In late 2010, the team unveiled their first prototype, a hatchback dubbed “the Blue Thing.” To save costs and provide a proof of concept for investors, the vehicle was essentially a rebodied Mini Cooper that used the Mini’s engine and drivetrain. A photo marking the occasion shows Scaringe and his team proudly posing behind the unfinished-looking Blue Thing with wrapping tape still covering the front bumper.

“It was like anything else,” Scaringe said on a 2010 podcast about his company’s origins. “The only way to start was to start. Lots of U-turns, lots of twists, lots of turns, lots of gut punches. Overall, just brute force built it.”

Scaringe was able to secure critical early funding through a creative lease-back agreement with Space Florida, the state’s aerospace economic development agency. According to TechCrunch, Space Florida invested nearly $2 million, partially because it was interested in the vehicle for the Moon or Mars. The agency bought the Blue Thing for $500,000, and leased it back to Avera for just $100 a year. It later bought three additional prototypes for $1 million under a similar agreement. (The story doesn’t end well for Space Florida. The agency thought it had a 3 percent stake in the company that would become Rivian, which could have translated to a $3 billion windfall after the company’s 2021 IPO. But thanks to an unexplained recapitalization, it only netted about $7 million.)

Eventually, Avera changed its name after Hyundai threatened to sue, claiming the name was too similar to its Azera sedan. The new name, Rivian, was a play on the Indian River lagoon near Scaringe’s hometown.

During this time, the product goals kept shifting between an affordable sports car, an expensive supercar, and a spartan race car to woo various investors. They even considered a dune buggy after a brush with insolvency necessitated bringing in investment from Saudi Arabia. It was there that Scaringe was challenged to design a rugged but efficient pickup for the desert, a mashup between an F-150 and a Toyota Prius, according to the Bloomberg article.

The desert vehicle aspect fell by the wayside, but the truck part stuck. Scaringe knew that to succeed, it needed to be different from all the big, bloated pickups coming out of Detroit. According to Bloomberg, he made a unilateral decision during a board meeting: The truck would be electric. Tesla was about to release the Model 3 to much fanfare. Trucks were some of the best-selling vehicles in America, yet no one was thinking about how to make an electric one. The old compliance cars from the 1990s were fading from memory. Tesla, and electrification, seemed like the future.

RJ Scaringe, founder and CEO of Rivian, unveils the R1T electric pickup truck, left, and R1S electric SUV during a reveal event at AutoMobility LA ahead of the Los Angeles Auto Show on November 27th, 2018.Photo: Patrick T. Fallon/Bloomberg via Getty Images

Rivian first revealed the R1T in 2018; the R1S SUV came out a day later. Deliveries were delayed several times, with customers having to wait until 2021 and 2022, respectively, to receive their vehicles. But eventually Rivian worked out the kinks; since production began in September 2021, the company has delivered over 175,000 vehicles — a respectable number for a brand-new startup. Tesla didn’t sell its 200,000th car until 2018, about 10 years after selling its first.

But the real coup was securing a massive contract with Amazon for 100,000 electric delivery vans, leading to its blockbuster IPO in 2021, briefly valuing the company higher than Ford. The retail giant invested a total of $1.4 billion in Rivian, becoming a major shareholder in the process.

This would become a running theme for Rivian: hitching its wagon to much larger, better-financed companies to ensure long-term viability and survival. Before the Amazon deal went through, Rivian was in talks to build EVs for GM and then Ford, though neither deal would pan out. (In muscling out GM, Ford invested $500 million in the company in 2019, only to sell its majority stake in a $7.3 billion write-down three years later.)

The Amazon deal would eventually get eclipsed by the 2024 partnership with Volkswagen, which offered $5.8 billion to form a software joint venture — a life preserver for Rivian during a precarious moment. The company closed that year with over $4 billion in losses.

Amazon delivery trucks built by Rivian sit in a lot at a Rivian facility on February 22nd, 2024, in Chicago.Photo: Scott Olson/Getty Images

Rivian is now heavily tied to Volkswagen, which is dealing with its own problems around plunging profits, declining market share, and software reliability issues. Rivian could help fix the lattermost of those problems, designing software for brands across the German auto giant’s portfolio. Sure, making software for internal combustion engine vehicles may seem like a bad fit for an EV-only company like Rivian, but the deal was too good to pass up. VW now owns a nearly 16 percent stake in Rivian, surpassing Amazon as the largest shareholder.

In some ways, it’s a hedge against a more uncertain and unpredictable future. In their public comments about the partnership, VW executives have seemed to imply that Rivian could one day become a subsidiary under the German auto giant’s global brand. The company didn’t start as an electric vehicle maker, and there’s a possibility it may not survive as one either.

“There’s a lot of moving pieces that they have to juggle,” said Sam Abuelsamid, an automotive tech analyst at Telemetry. “If they drop any of them, if they drop any of those balls that they’re juggling, that could be an existential crisis for them.”

The most prominent ball is the R2. If Rivian drops it, and the R2 is just a modest hit, will Vreeland and the rest of those fans stick around?

Scaringe, a longtime vegan, is just finishing up a veggie breakfast before getting ready for an all-hands meeting at the Normal factory. The continued post-tornado cleanup will be a major topic, as will the impact on the R2 ramp. His message will be clear.

“We’re trying to compete with the best in the world. We have to have a very large engineering budget [and] we’re spending a tremendous amount of money to build truly world-class technology,” Scaringe tells me. “If you qualify the same with the tornado hitting the plant at the start of production, it’s definitely the fastest. I don’t think Tesla or anyone else had to deal with that as well.”

Rivian founder and CEO RJ Scaringe.Photo: Anthony Tahlier / The Verge

The company desperately needs this vehicle to be a hit. Last year was a tough one for Rivian, with sales dropping 18 percent year over year. Expectations of profitability are being pushed further into the future thanks to heavy R&D spending on autonomy. Still, Rivian projects it will sell upwards of 67,000 vehicles in 2026, anchored primarily by the R2. With R1 sales expected to remain flat, that would mean Rivian expects to sell as many as 25,000 R2s before the end of the year. As TechCrunch noted, that would be one of the fastest launches of a new EV ever recorded.

“R2 is the most important thing we have because it’s what takes us from being subscale to being a scaled manufacturer,” Scaringe tells me.

Ramping up production of a brand-new vehicle, especially one as vital as the R2, is rife with pitfalls. Ask Elon Musk. Tesla became a giant in the auto industry based on the success of the Model 3, an affordable EV that seamlessly blended cool tech with an affordable price. But the car very nearly drove Tesla out of business. The company was bleeding cash while being over-leveraged on robots and automation. In a desperate bid to hit a 5,000-vehicle weekly milestone, Tesla set up a giant, makeshift assembly tent in the parking lot to build the Model 3s by hand and ease bottleneck issues inside the main plant. Musk famously referred to the rollout as “production hell.”

Scaringe could hardly be confused with Musk — one’s an egghead from Florida, the other’s a pro-natalist white supremacist who recently (and briefly) became the world’s first trillionaire — but he knows that dangers lurk around every corner. (Scaringe himself is estimated to be worth $200 million, despite a recent divorce that cost him millions of dollars and some of his voting stake in the company. That said, Scaringe did succeed recently in getting Rivian’s board to approve a Musk-like, performance-based pay package worth over $4 billion. If the R2 is a hit, Rivian lives on — and Scaringe becomes the world’s newest billionaire.)

Like Tesla, Rivian relies on some automation — during my tour, I watch as hundreds of industrial robot arms assemble the trays that will eventually contain the R2’s lithium-ion battery — but won’t be overreliant on it. For Scaringe, the real threat lies further downstream, lurking in the company’s supply chain. That’s why he’s putting such an emphasis on making things in-house, like writing its own operating system software. While many automakers rely on suppliers like Bosch for electric motors, Rivian designs and manufactures its own. It sources its battery cells from LG Energy, but makes its own packs, as well as the cooling systems for fast charging. Soon, it will design its own AI chips for autonomous driving. It’s even considering making its own lidar sensors. Anything less, Scaringe believes, would open the company up to an unacceptable level of risk.

“The supply chain environment has not been stable in the last five years. It’s just very, very dynamic,” Scaringe says, citing tariffs, the war in Iran, and high gas prices. (Yes, even an EV company can feel the sting, due to the downstream effects of mining for battery minerals, for example.) “The biggest risk for ramp-up is the supply chain.”

The Rivian Factory in Normal, Illinois, on April 22nd.Photo: Anthony Tahlier / The Verge

Vertical integration at this level is enormously costly. But Rivian isn’t like Fisker or any of these other EV startups that have already gone out of business, Scaringe insists. He explains that the company is spending significantly on a world-class engineering team, producing its own silicon, motors, power electronics, and computers. Yes, Rivian loses money every year, but it also creates a variable cost advantage at scale. The R2 is the essential vehicle to unlock that scale, enabling the company to cover these fixed costs. But the process will start slowly, Scaringe says. “We’re not pressing a button to have hundreds of thousands of cars produced the day we start production.”

Still, it quickly becomes clear that what really animates Scaringe isn’t talk of supply chain woes or the anguishes of early production — it’s autonomy. A conversation about Rivian’s entry into self-driving cars is when he is at his most Muskian, projecting that the company — which has never publicly demonstrated a fully autonomous vehicle — will eventually develop a Level 4 system, meaning no human supervision required within certain limits.

“It’s our biggest focus; it’s the largest area of investment,” Scaringe says. “It’s my biggest focus in terms of my activities within the business — this is what I’m spending the most time on.”

This surprises me, almost as much as it did when I was in Palo Alto last year to watch Rivian announce its plans to design its own AI chips. It seems counterintuitive, like a pivot from what the company had previously represented. Rivian is supposed to be an outdoor lover’s alternative to Tesla: rugged and off-road-worthy, while still being fully electric. Now it seems like it’s trying to chase the mercurial billionaire down an AI rabbit hole.

Scaringe subscribes to Musk’s outlook that, eventually, “everything will be autonomous … and that’s not 10 years away,” he says. “That’s in the next couple of years. You’re going to start to see ever-increasing levels of autonomy.”

It’s not just the cars that are getting an AI makeover. Over 70 percent of Rivian’s software team is using Cursor, an AI coding agent owned by SpaceX, to code new software features, said James Philbin, senior VP for autonomy and AI at the company. “I think it definitely changes the job,” Philbin told me last April. “But potentially I think will accelerate a lot of software development.”

I asked whether Philbin, who previously worked at Waymo and Zoox, felt like he was swimming against the tide by developing an autonomous vehicle at an adventure- and off-roading-themed EV company. “I mean, I would flip the question back and I would say, ‘Why are these legacy OEMs swimming against what is clearly the tide of the future?’” he said. “In 10 years, the vast majority of cars will be EVs and I think the vast majority of cars will be increasingly autonomous.”

The Rivian R2 reveal event held at Rivian’s factory in Illinois.Photo: Anthony Tahlier / The Verge

While some may see the move as savvy, others could construe it as desperate. As EV demand stalls out, autonomous vehicles have been surging, fueled by the incredible growth of robotaxi startup Waymo over the last year. The company now operates in 11 cities in the US. Meanwhile, Uber’s hot on Waymo’s heels. Last March, the ridehail company said it would buy 10,000 robotaxis from Rivian — specifically, autonomous R2 vehicles — as part of a deal that could net Rivian $1.25 billion in new investment. Like VW, Uber has become another well-capitalized benefactor during a pivotal moment. And before the R2 has even rolled out, Rivian has an order for 10,000 of them.

To lay the groundwork, Rivian is offering more AI-powered products to its customers. Last year, the company rolled out its so-called Universal Hands-Free driving feature, covering 3.1 million miles of road in the US and Canada, for second-gen R1 vehicles. And later this year, it expects to release point-to-point hands-free driving, which includes navigating turns, intersections, and on/off-ramps. Think of it as Rivian’s version of Tesla’s controversial Full Self-Driving feature.

But early tests have been mixed. Edmunds called it “boring” but impressive. In his review of the R2, Verge contributor Lawrence Ulrich called the Universal Hands-Free system “overconfident to the point of obliviousness.” My own experience in the passenger seat of an R1S using hands-free driving was marked by several disengagements and a couple of close calls with cyclists. And now, Rivian is facing a lawsuit from R1 owners over allegations of false claims about its autonomous driving capabilities.

But while Tesla’s approach to full autonomy relies exclusively on cameras, Rivian plans on adding lidar sensors to its R2 vehicles later in 2026 for extra safety. That was enough to get Uber through the door and into a lucrative deal. And Rivian could certainly use the cash as it wades through the sticky terrain of early R2 production. Scaringe argues that Rivian can outpace Tesla, and even Waymo, given its ability to feed camera data from its customer fleet into its “Large Driving Model.”

“We’ll have a lot more data than Waymo,” he says. “Tesla, by far, has the most data but the most constrained sensor sets … Waymo has the most sensors but a very small fleet. We have something in the middle.”

Of course, the money won’t come unless Rivian hits a series of crucial milestones, Uber CEO Dara Khosrowshahi said on the Decoder podcast recently, including testing without a safety driver behind the wheel. “They’ve got to deliver,” Khosrowshahi said, “and based on everything that we’ve seen from RJ and team, putting together a first-class AI team, we’re confident that they can deliver on those R2s.”

Scaringe seems equally confident in Rivian’s success. Of course, he has to be. Throughout our interview, he projects confidence. Rivian won’t be acquired by VW, but the partnership allows for cost sharing of parts and content between brands — much in the way a subsidiary would.

It isn’t until later, when we’re walking the factory floor together, that the mask slips a little.

“The R2 is how all that [vertical integration] comes to a realization,” he says. “It’s why R2 for us — if R2 is not successful, then we’ll need to resize how we’ve set up the business.”

Which is to say: If the R2 doesn’t work, Rivian won’t look like Rivian much longer.

When the R2 comes out, it won’t be available in the world’s largest market for electric vehicles: China.

Unlike most American car companies, Rivian only sells its cars in North America. Tesla’s decision to open a factory in Shanghai is widely seen as a pivotal moment in the company’s history and a key waypoint on the road to its current $1.24 trillion market cap.

Aerial view of vehicles destined for overseas markets at Taicang Port on June 29th in Suzhou, Jiangsu province, China.Photo: Ji Haixin/VCG via Getty Images

But China’s domestic brands, not Tesla, are the ones that are currently devouring the world. Backed by heavy state support and massive domestic overcapacity, China currently manufactures roughly 75 percent of all EVs worldwide. Many of them sell for less than $20,000 — half the price of any comparable vehicle in the States. Tariffs and other protectionist policies have kept China from selling its cars to US customers, though experts wonder how long that will hold. Ford CEO Jim Farley has warned that allowing China to sell its cars in the States would be a death blow for Detroit.

Rivian has imported its own fleet of Chinese cars for benchmarking purposes. It tests them on a private course, and then tears them to shreds to examine every bolt and computer chip. Xiaomi’s SU7, in particular, has impressed Scaringe, with its clean-sheet approach to software and superior quality execution. But the Rivian CEO doesn’t ascribe the same industry-killing abilities to China as some of his automotive counterparts.

“We’ve taken the cars apart … there’s no magic,” he says. “They’re using stampings, they’re using castings.”

The country’s actual competitive advantage, according to Scaringe, is that labor is so much cheaper over there. Even if the cars are built here in the US, that advantage vanishes, and China has to compete on a more level playing field.

Still, Washington seems to have taken Detroit’s apocalyptic warnings to heart, throwing up seemingly insurmountable trade barriers to prevent China from exporting its cars to the US. Tariffs make it prohibitively expensive for any Chinese car to be sold here. And cars with any connected vehicle software that originates from China are barred from sale under a Biden-era law.

But many auto experts think it’s inevitable that Chinese cars will eventually find their way to US dealerships. The US could require Chinese OEMs to create a joint venture with a domestic brand, much in the same way that China does for foreign automakers that want to sell to its residents. Even President Trump has expressed willingness to allow China to sell in the US, as long as the cars are manufactured here.

If they do, Scaringe is confident Rivian won’t get snowed under. “If the Chinese companies decide to produce domestically in the United States, those advantages disappear,” he argues. “The cost of labor for a Chinese company operating in the United States and the cost of labor for a US company operating in the United States is going to be the same.”

But expecting a level playing field in this day and age seems a bit foolhardy, if not outright unrealistic. Detroit is trying to insulate itself from a Chinese onslaught by going back to the drawing board on EVs. Can Rivian engineer its own solution?

TheThe relationship between small towns and the big-name employers who come in promising lots of jobs and investment is always a bit fraught. There’s concerns about putting all your eggs in one basket when relying heavily on a single employer, especially one deeply rooted in a volatile market environment. Normal certainly knows this better than most, having endured the “changed fortune” headlines after Rivian’s purchase of the former Mitsubishi plant in 2017.

But those fortunes could reverse if the EV market fails to climb out of its current slump. Like most automakers, Rivian has gone through several rounds of layoffs in its short history, most recently in June, with cuts from its sales and marketing divisions. The company’s salaried factory workers were excluded; next time, they might not be so lucky.

The entrance to the Rivian factory in Normal, Illinois. Photo: Anthony Tahlier / The Verge

“It’s been kind of a culture change for the Bloomington-Normal community,” said Adam Campbell, dean of career and technical education at Heartland Community College, where Rivian runs an apprenticeship program.

Campbell said that aside from its vast cornfields, Normal has always had a sizable white collar community, with State Farm’s international headquarters and two universities. “With Rivian coming in as a major manufacturer within the state, that’s kind of shifting the culture a little bit for the better. So now we are moving more into advanced manufacturing.”

But even Heartland realizes that the winds are changing. The future, at least the American future, won’t necessarily be “EV only,” as some had previously predicted. Hybrid vehicles are now the new growth product of the industry, which has spurred the college to add a new hybrid diesel equipment technology program to its curriculum.

“We’re not stuck in the old way of doing things,” Campbell said. “We do try to be innovative.”

It may not be enough. Rivian has already experienced significant turnover in its short history, shedding thousands of employees. Campbell knows he needs to structure his curriculum to respond to a rapidly changing economic environment.

“Hypothetically, if the electric vehicle industry ebbs and there’s a pause in that, the students that come through our programming, what they learn is still 100 percent viable,” he says. “It can work across multiple industries.”

It seems like everywhere you go in Normal, there’s some reminder of Rivian’s outsize role in the community. The Parkview Inn, a restaurant in Bloomington known for its broasted chicken, slot machines, and decor that leans heavily on license plates as art, is a typical haunt for Rivian workers, current and former. Joe (who declined to give his last name) worked at the factory for a year and a half before deciding that spending hours indoors every day on the assembly line was not the best use of his time. He now works in landscaping and is much happier.

Joe, who is still wearing his “Keep Normal Adventurous Forever” T-shirt, says he keeps in touch with his friends at the plant, who report that the pace has picked up considerably for the R2 production. “All my friends say it’s, like, at least 10s if not 12s right now,” he says, referring to 10-hour and 12-hour shifts. Some workers are on four 12-hour shifts a week on a three-week rotating schedule.

He has a lot of opinions about Rivian, like about the managers (“Everybody’s chill”), the high-end e-bike the company says it’s making (“In my opinion, they were late to the game when they dropped it”), the skilled workers that Rivian brought in from out of the country (“They bought houses for people to come here”), and the rumors of a secret hydrogen project Rivian is allegedly working on (“I don’t know if they’ll ever speak about that again”). But ultimately, he wishes Rivian well and is actively cheering for the R2 to be a success.

“I hope so,” Joe says. “Just for the sake of my town.”

An event for the R2 on the Rivian factory floor.Photo: Anthony Tahlier / The Verge

Joe may not know it yet, but Normal isn’t the only town relying on Rivian’s future success. The new Normal is a town called Stanton Springs North, about 45 miles east of Atlanta, Georgia, where Rivian recently broke ground on a new $5 billion factory to build hundreds of thousands of R2s. While the Normal plant is a retrofitted facility with a capacity of only 200,000 vehicles a year, the upcoming Stanton Springs campus is designed to be the platonic ideal of a Rivian factory, built from the ground up to manufacture up to 400,000 EVs annually.

It represents Rivian’s dream for its future. But the company recently was forced to scale back the plan after Trump’s Energy Department slashed a loan agreement with the company from $6.5 billion to $4.4 billion. It’s still getting a big loan, but even its dreams are getting downsized. Its own government appears to be actively working against it, first by eliminating pro-EV policies, and now by renegotiating a settled loan deal.

The myth Americans tell themselves about their country is that it’s a place that breeds innovation and nurtures the entrepreneurial spirit. A well-regarded company that makes interesting, thoughtfully designed products that are also environmentally friendly, one that employs thousands of people, with the potential to employ thousands more — you’d think a company like this would be on a glide path to success. And yet Rivian is struggling to secure its future in the face of a volatile market, a hostile government, and fatigued consumers. If Rivian stumbles, or worse, fails, that myth of American innovation may suffer an irreparable blow.

The R2 had its official launch June 9th. The reviews were broadly positive, and fans are already crowding into the Rivian subreddits and other forums to yap about specs and test drives. Will the non-fans, the hundreds of thousands of regular people that Rivian desperately needs to dig its way out of this rut, show up too? The company is riding high on the positivity of loyal customers like Romy Vreeland and her club cohorts — even as its stock never seems to rise above $20 a share. Clouds loom on the horizon: Days after the R2 deliveries started, Rivian announced it was laying off hundreds of employees.

The market headwinds are blowing, gusty and gale-sized, but not yet spinning fast enough to classify as a tornado. Rivian can only batten down the hatches, and face thestorm.

Fujifilm is expanding its QuickSnap lineup with a new disposable camera focused on taking monochrome photos and another built to survive harsh outdoor environments. The $22.90 QuickSnap Black and White and the $24.75 QuickSnap Active are expected to launch sometime later this fall, to the delight of Gen Z snappers driving the current resurgence in disposable and retro film cameras.

As the name suggests, the QuickSnap Black and White comes with black and white film installed that’s designed to capture “rich contrasts, tones, and textured grains,” according to Fujifilm’s press release. The camera is also compatible with standard color film processing, meaning you won’t have to hunt down a specialized black and white service when you get the camera developed.

The QuickSnap Active is a rugged disposable camera that caters to “unpredictable, all-weather, all-terrain adventures.” It’s waterproof up to a depth of 10 meters (35 feet) and features an oversized green film advance knob and lever shutter, which make the camera easier to operate while underwater. Fujifilm says the QuickSnap Active is replacing its current waterproof QuickSnap model, which is still available to buy at retailers like Walmart and Target.

Getty is planning to axe its $3.7 billion merger agreement with Shutterstock after a UK regulator imposed restrictions that would prevent part of Shutterstock’s business from being included in the deal. The move comes despite the US Department of Justice granting the deal “unconditional antitrust clearance” in February.

In an SEC filing published on Tuesday in the US, Getty said it is “not required to accept” approval conditions outlined by the UK Competition and Markets Authority in May that require Shutterstock to sell its global editorial business, including the Backgrid and Splash paparazzi agencies.

Those conditions have proved unappealing enough for Getty to walk away from the deal, which aimed to combine the companies’ stock photo libraries. Both companies face competition from AI image generators that provide fast and cheap media content on demand. The company’s board of directors “unanimously” voted to terminate the merger agreement on July 6th, “assuming no material change in the aforementioned circumstances” occurs before July 7th. That essentially leaves the Getty/Shutterstock merger dead in the water.

Announced at CES 2026, the Moto Tag 2 has finally arrived in the US. The Bluetooth tracker with improved battery life over its predecessor, plus UWB sensing for more accurate tracking, is available through Motorola’s site at a special $19.99 launch price (originally $29.99). There are plenty of trackers for Android owners to consider, but the Tag 2 seems tough to beat at this price. It supports Google’s Find Hub network, has an IP68 rating, and runs off a replaceable CR2032 coin battery that can last up to 600 days.

The Moto Tag 2 has a built-in speaker that can sound off at up to 77 decibels. The tracker itself has a button that can be set to do several things, and is configurable within the Moto Tag app. Pressing the button can help you find your phone nearby. It can also trigger your phone’s camera shutter. A Bluetooth tracker doesn’t need most of these features to be good, but the Moto Tag 2 has them anyway.

Dear gaming community,

For a good while, it has been all positive news from IO Interactive. We remain humbled and honored by the response to our latest outing with a young, unproven Bond. A bold new story and a take on one of the most famous characters in entertainment, which the world has embraced with love.

However, today we have to share some more downbeat news. A relationship with an external partner on our own IP, Project Fantasy, has come to an end. This means we have to adapt to this new reality and its short-term consequences, including staffing decisions, which is what is happening as we write this update, and we are fully committed to supporting those affected through this challenging transition.

Project Fantasy is a game, a world, and an IP that we absolutely love and remain 100% committed to, now and in the future. This wonderful universe will see the light of day.

The price of LG’s UltraGear 27GX700A-B 1440p gaming monitor that came out last August has dropped to nearly its best price yet. You can grab it at Amazon for $484.99 or at LG for $499.99 (originally $599.99). That’s a great deal considering this is one of LG’s Primary RGB Tandem panels built with fourth-generation WOLED tech that delivers higher brightness and punchier contrast than its predecessors, rivaling QD-OLED.

This model could satisfy gamers with PCs built to deliver smooth 1440p gameplay. It features a 280Hz refresh rate, twin HDMI 2.1 and one DisplayPort 1.4 jack, a near-instant .03-millisecond response time, not to mention a thin design. The monitor also has some USB-A ports for connecting accessories, plus a headphone jack for private listening.

The 27GX700A-B supports AMD FreeSync Premium Pro and is Nvidia G-Sync-compatible, meaning it supports variable refresh rates and can keep the image running smoothly even with variations in frame rate. We haven’t tested this model, but reviews site Rtings does say it has “distracting VRR flicker,” though that likely depends on the game.

While costs have risen in recent months on computer components like RAM and storage, it’s been delightful to watch prices go in the opposite direction for OLED gaming monitors. Some people may prefer to spend less on fast, feature-packed IPS and VA panels, but it’s nice that there are more opportunities to get cutting-edge panel tech for your gaming PC or console.

Would you pay $20 a month for access to AI hardware you already own? That appears to be one of Meta’s next bets. This week, it quietly announced that your glasses’ Conversation Focus feature will soon be limited to three hours of use per month, unless you pay for a $19.99 Meta One Premium subscription.

In a help article, the company insists that it won’t require a subscription to use your glasses, period; it’s merely erecting a “rate limit” for certain AI features. Even premium subscribers will only get 15 hours of Conversation Focus per month under that “rate limit,” it claims.

Problem is, Meta’s rate limit is ridiculous. The Conversation Focus feature, which amplifies the voice of the person you’re speaking to so you can hear better in noisy environments, is not something that should plausibly be rate-limited, because it doesn’t use Meta’s servers. It runs on-device, using the chips inside the glasses that you’ve already purchased. I turned off my internet, and it kept working.

Here’s how the company introduced it last year: “[C]onversation focus uses your AI glasses’ open-ear speakers, beamforming technology, and real-time spatial processing to dynamically amplify the voice of the person you’re talking to.”

Not only does it avoid Meta’s servers, but Conversation Focus doesn’t technically require an internet connection at all. I double-checked by turning off my phone’s Wi-Fi and cellular, turning on Airplane Mode, and I was still able to use Conversation Focus just fine by tapping a button on my phone.

Does Meta have some secret licensing deal with another company that costs it money every time a person uses Conversation Focus? Failing that, the rate limit sounds utterly bogus.

We’ve asked if Meta can explain the move, and whether the company plans to put other on-device features behind a subscription. Meta didn’t immediately respond to a request for comment.

After weeks of negotiating with the Trump administration, Anthropic is finally going to be able to bring Claude Fable 5 back online. In a post on X, Anthropic said it plans to begin restoring access tomorrow.

We’ve received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5.

We’ll begin restoring access tomorrow, and will share an update soon.

We’re grateful to our users for their patience, and to everyone who worked with us on redeploying the models.

In early June, Anthropic sidelined Fable 5 — its consumer-facing model built the same underlying technology as Anthropic’s Mythos 5, but with more safeguards — after a Friday evening ultimatum from the Trump administration. Due to concerns over potential jailbreaks of the technology, the government had slapped Anthropic with an export control directive, disallowing any foreign national (including non-US members of enterprise client companies and even many of Anthropic’s own employees) from using either Mythos 5 or Fable 5, i.e., both the models that Anthropic had spent the last week hyping.

The Trump administration recently greenlit the return of Mythos 5, but to a pre-approved list of organizations only. Non-US members of those organizations, as well as Anthropic’s own foreign national employees, are now allowed to access the model. The decision came shortly after OpenAI unveiled GPT-5.6, which the Trump administration only allowed to debut under similar rules: a staggered rollout, initially to only a pre-approved list of organizations and government departments.

The initial export control directive by the Trump administration came at an inopportune time for Anthropic, as it’s preparing for an IPO and has been feuding with the government for months over a supply chain risk designation.

“It represents how important this is to our mission that this is right up there with Claude Code and Claude Cowork as the next really significant product that we’re releasing,” says Eric Kauderer-Abrams, Anthropic’s head of life sciences. “Our mission is to develop AI that serves humanity’s long-term well-being, and we believe that by far the greatest opportunity to do that is in the life sciences.”

For the past decade, one company—Google DeepMind—has been at the vanguard of AI for science. CEO Demis Hassabis and researcher John Jumper won the Nobel Prize in chemistry for their work on the company’s AlphaFold model, and DeepMind has also made major contributions to meteorology, materials science, and a variety of other disciplines. But in the past several months, the fast-advancing frontier of AI progress seems to have left DeepMind in the dust. When it comes to coding, which has become the most lucrative use case for LLMs, DeepMind is stuck playing catch-up.

Anthropic is well positioned to take up DeepMind’s scientific mantle. Like Hassabis, Anthropic CEO Dario Amodei is a PhD scientist—unlike OpenAI CEO Sam Altman, who’s a businessman through and through. Many scientists are already avid users of tools such as Claude Code. These days, a lot of scientific research involves some amount of coding, but not all scientists are expert software engineers, and so tools like Claude Code can make a huge difference for their productivity. And the company has recently earned a major scientific vote of confidence: Earlier this month, Jumper announced that he is leaving DeepMind for Anthropic.

Since agents powered by LLMs, including Anthropic’s Opus model series, became capable of useful, independent work in late 2025, scientists have been seeing just how much they can do. In a blog post published on Anthropic’s website, the Harvard physicist Matthew Schwartz estimated, on the basis of his work with Claude Code and other Anthropic tools, that the company’s Opus 4.5 model is about as capable of executing scientific projects as a second-year graduate student.

According to Kauderer-Abrams, Claude Science isn’t intended to displace Claude Code and Claude Cowork in scientists’ workflows. Instead, it’s designed to build on what scientists already find useful about Anthropic’s products. For instance, it not only writes code but also helps scientists run their code on powerful computer clusters, which many many scientists need for their work but can be difficult to manage. And it prioritizes reproducibility, so that scientists can trace back the source of any figure or result and check it for accuracy and validity.

Google’s NotebookLM is adding a new way to catch up on your notes: TikTok-style AI videos. The new feature is rolling out to Google AI Ultra and Pro subscribers, allowing NotebookLM to generate 60-second vertical AI clips based on the sources you upload to the app.

The example shared by Google details Australia’s unsuccessful war on emus, pairing paper cutout-style AI art of emus with narration. It adds to some of the other ways NotebookLM lets you interact with your research, including by generating AI podcasts, cinematic videos, and visual explainers.

To generate a 60-second clip, head to NotebookLM on the web or app, select a notebook, and then choose “Video” from the Studio column on the right side of the screen. From there, select “Short,” choose the topic you’d like NotebookLM to focus on (or enter your own), and then hit the “Generate” button.

The feature is rolling out in English only for now, with support for free users coming “soon.”

The Federal Trade Commission fined Amazon $2.25 million to settle claims that the company failed to help customers who fell victim to identity theft, as reported earlier by Bloomberg. In its complaint, the FTC accuses Amazon of refusing to provide customers with information about purchases made with fraudulent accounts, in violation of the Fair Credit Reporting Act (FCRA).

As alleged in the filing, identity theft victims who contacted Amazon “would often enter a Kafkaesque sequence” where a support agent wouldn’t provide records related to a fraudulent account unless they could name the person who opened it.

In one instance, a victim attempted to guess the fraudulent account owner’s name over 30 times, but Amazon allegedly wouldn’t remove the victim’s credit card information from the thief’s account. The FTC claims Amazon also failed to respond to identity theft victims’ requests for records within the 30 days required by the FCRA.

An Amazon spokesperson tells Bloomberg that the company has “resolved this matter with the FTC” and “implemented process improvements for customers who believe they may be victims of identity theft.”

As high memory and storage prices have driven up the cost of everything from consoles to computers, finding a competent laptop for under $1,000 has become a challenge. Thankfully, the Acer Swift Go 16 AI is on sale for $899.99 at Best Buy, a steep discount from its usual list price of $1,549.99. The well-equipped laptop has an Intel Core Ultra 7 355 processor, 32GB of LPDDR5X memory, and a 1TB SSD. It has a great 16-inch OLED (1,920 x 1,200) touchscreen, too. Its tall 16:10 aspect ratio lets you see more, and it has a 180-degree hinge that lets you find your preferred viewing angle.

The generously-sized multi-touch touchpad has a Corning Gorilla Glass top layer, and its backlit keyboard even squeezed in a numpad. Sure, its integrated Intel graphics aren’t great for serious gaming, but it should be more than powerful enough for Balatro, and the system can run for over 20 hours on a single charge, depending on brightness settings. It has quite a few ports, particularly when compared to other slim laptops, including two USB-A 3.2 ports, two USB-C with Thunderbolt 4 support, HDMI 2.1, a microSD card slot, and a 3.5mm headphone jack.

The current design sees the RAM placed on top of the SoC, which is slightly less efficient and runs hotter. This informative image helps explain the difference between the two designs; essentially, the new architecture should reduce heat dissipation and speed up communication between the two components. 

That results in better performance and power efficiency and also means the vapor cooling system inside Apple’s upcoming pro iPhones can work more efficiently to reduce heat dissipation. While no one knows for sure, some estimates claim this new WMCM packaging should enable 15-20% performance boost, even before we consider the improved efficiency inherent in the new A20 chip.

A huge data heist

All of this information is carried within the 200,000+ files (630GB) World Leaks published on its dark web site. The data also includes confidential Apple supplier list info, detailed information concerning the circuit board, battery parts, and camera modules – even confidential information about which suppliers are competing to supply specific components. These are all confidential trade secrets the company is unlikely to want public, as they give rivals rare insight into how the company’s supply chain is structured.

For example, Google engineers teach at ETH Zurich. ETH graduates join companies such as Anthropic. Researchers launch startups, while former employees of global technology firms go on to found new ventures of their own. Investors, founders, academics, and corporate teams encounter each other repeatedly through shared networks, industry events, and professional circles. In a region of this size, collaboration often happens less through formal introductions than through proximity. While talent flows freely, it rarely leaves the ecosystem.

One indicator of the region’s maturity is its ability to convene. Events such as the Zurich AI Festival will bring together more than 6,500 guests this September 28 to October 3. With more than 35 confirmed events across AI and the arts, AI literacy, health, technology, and policy, it is designed as a platform for cross-sector exchange. Its flagship events, the AI + X Summit, AI + Environment, and the AI + Policy Summit, will bring together internationally recognized leaders alongside researchers, policymakers, venture capitalists, and entrepreneurs, convening international voices and fostering dialogue across sectors.

Research, talent, and company creation

At the center of the country’s AI capabilities are institutions such as ETH Zurich, the University of Zurich, École Polytechnique Fédérale de Lausanne (EPFL), Scuola Universitaria Professionale della Svizzera Italiana (SUPSI), and Zürcher Hochschule für Angewandte Wissenschaften (ZHAW).

ETH Zurich ranks among Europe’s leading universities for deep tech commercialization, generating more than 40 spin-offs and startups in 2025 alone, helping create some of the continent’s most valuable technology companies.

The Stanford AI Index 2026 reinforces that picture: Switzerland ranks first globally for AI researchers and inventors per capita, with 110.5 per 100,000 inhabitants—ahead of Singapore (109.5), Sweden (80.6), and the United States (64.8). And the IMD World Talent Ranking ranked Switzerland as number 1 for the 10th consecutive year, leading globally in investment, development, and talent appeal.

Engineers, researchers, and founders move frequently between universities, startups, and established technology firms, creating strong knowledge flows across organizations. That density is increasingly attracting companies from outside the region too. Even before formally announcing their Zurich office, Exa.ai received a strong pipeline of candidate applications. ‘To assemble the greatest search team in the world, you’ve got to meet people where they are,’ says Will Bryk, the company’s CEO and co-founder. ‘And many are in Greater Zurich.’

Former Google Switzerland employees alone have founded approximately 210 companies and created around 2,600 jobs over the past two decades. For a country of around nine million inhabitants, the multiplier effect is significant. Large technology firms contribute not only through direct employment, but also through the creation of new companies and the transfer of expertise.

Why the Greater Zurich Area complements Silicon Valley

For many technology companies, Switzerland is not a substitute for Silicon Valley. The two serve different functions within the AI value chain.

Workers at Grand Theft Auto VI developer Rockstar Games have submitted a request for their union, the IWGB Game Workers Union, to be voluntarily recognized, according to a press release. The request follows Rockstar firing more than 30 staffers last year in a move accused of being “union busting.”

According to the release, IWGB members have been “actively” organizing since 2019 and “claim they now represent a significant proportion of the workforce across each of the studio’s sites in Edinburgh, Dundee, Lincoln, Leeds and London.” As the union has grown, “Rockstar has already made improvements to workers’ conditions including unprecedented average pay rises and financial incentives for crunch for the first time ever.” The union wants to improve pay transparency, strengthen “flexible working arrangements,” and address crunch, which is something Rockstar has come under scrutiny for in the past.

Rockstar owner Take-Two didn’t immediately reply to a request for comment. A final hearing in an employment tribunal over last year’s firings is scheduled to start in September — shortly before the planned launch of Grand Theft Auto VI on November 19th. Rockstar announced a $79.99 starting price for the game and opened preorders last week.

The push from Rockstar workers is part of a bigger movement by game developers to unionize. Double Fine staffers filed a petition to unionize in May, and workers at the Activision studio Raven Software secured their first union contract last year.

Agricultural AI also needs to understand more than just customer attributes; it needs to understand the land: GPS coordinates, farm boundaries, field blocks, and soil variation across a single property. Where do you apply fertilizer, and at what rate, and in which specific area of the farm? Not all parts of a field are the same, and an AI system that treats them as if they are will produce recommendations that are at best imprecise and at worst damaging.

There is also a compliance dimension due to the chemicals and the responsibility involved. Operational AI in agriculture needs significantly more checks and governance than it might in a lower-stakes environment. When a flawed recommendation gets acted upon in the field, the consequences can be severe. 

What data readiness means in practice 

Data readiness is the difference between AI delivering on its promise vs. a “garbage in, garbage out” scenario. Fundamentally, being ready for AI means having a data model that accurately reflects how the business operates. 

For a company like Wilbur-Ellis, a 104-year-old, family-owned agricultural distributor, that means understanding who your customers are, which fields they farm, which inputs they need, which suppliers those inputs come from, what they paid last season, and how all of that connects to margin. That information needs to be current, consistent, and accessible across the organization, rather than locked in separate systems that were never designed to talk to each other.

Similarly, for farming operations themselves, data readiness means having a reliable, connected picture of what is happening across every field: soil health records, input application histories, yield data from previous seasons, equipment performance, and real-time sensor readings from irrigation systems.

Governance matters just as much as structure. Prices change, relationships evolve, and suppliers come and go. An AI system drawing on data that was accurate six months ago but has not been maintained will make recommendations based on a version of the business that no longer exists. 

Building the foundation that makes AI trustworthy

The good news is that the path to data readiness is feasible. It starts with a strong data model: a single, governed source of truth that connects customers, suppliers, products, pricing, orders, and margins in a way that reflects how the organization operates. 

From there, it requires data pipelines fast enough to deliver insights when decisions need to be made, governance frameworks that keep that data trustworthy over time, and security controls that ensure sensitive commercial information is accessible to the right people under the right conditions.

The Supreme Court upheld birthright citizenship, ruling 6–3 against President Donald Trump’s effort to end the longstanding constitutional right via executive order.

Birthright citizenship dates back to Reconstruction. Under the 14th Amendment, which was ratified in 1868 to guarantee citizenship and equal protection to the children of formerly enslaved people, anyone born in the United States “and subject to the jurisdiction thereof” is a citizen.

Trump issued the executive order, titled “Protecting the Meaning and Value of American Citizenship,” just hours after being sworn back into office in early 2025 — and the administration was almost immediately sued in response. Defending the executive order before the Supreme Court, Solicitor General D. Sauer argued that noncitizens and their children aren’t “subject to the jurisdiction” of the United States, since their loyalty lies with a foreign power.

An estimated 250,000 children are born on US soil to noncitizen parents each year. The birthright citizenship ban would not have gone into effect retroactively, but would have applied to anyone born 30 days after Trump’s executive order was issued.

Even conservative justices were skeptical of the administration’s argument. After Sauer claimed that the 14th Amendment needs to be revisited because we live in a different world than that of Reconstruction, Justice Neil Gorsuch countered: “it’s the same Constitution.” Still, the fact that such a fundamental constitutional right was up for debate shows just how much Trump’s nativist vision has affected American politics.

Gorsuch ultimately caved, not only signing onto Justice Clarence Thomas’s dissenting opinion but also writing one of his own. Justice Samuel Alito also dissented.

Videos purporting to show the iPhone 18 Pro undergoing a drop test were removed from X shortly after they surfaced, as spotted earlier by MacRumors. An account imitating leaker EvLeaks was suspended from X after sharing the clips, while the post was removed as well, with X saying it “violated” the platform’s rules.

Another leaker, IceUniverse, also shared the clips. In a post on Weibo, IceUniverse showed a screenshot from X indicating the post had been taken down, alongside the machine-translated caption: “Apple has already started blocking the leaked data on Twitter.”

The videos line up with Reuters’ reporting, which said the leaked data included photos of the iPhone 18 Pro’s drop test and parts lists. In the clips that surfaced on X, a gray device featuring a three-camera layout and an Apple logo falls a short distance against a checkered background. A source told Reuters that Apple had been “concerned” about the breach. Apple didn’t immediately respond to The Verge’s request for comment.

If you’re anything like the managers studied by Boston University professor Emma Wiles, treating that AI as a “coworker” would lead you to do a worse job. They caught 18% fewer errors when the work was attributed to an agentic “AI employee” rather than a chatbot.

This is an alarming glimpse of the future Silicon Valley is hurling us toward. Microsoft, OpenAI, Anthropic, and Google have all released tools for managing teams of AI agents, many of which are advertised as digital colleagues. Find out why that’s a losing proposition for workers.

—James O’Donnell

This story is from The Algorithm, our weekly AI newsletter. Sign up to receive it in your inbox every Monday.

This flying solar-powered platform could deliver better internet from the air

As soon as August, a giant silver bullet will cut its way through the dry air of the southwestern US and cross the Pacific to reach the coast of Japan.

Once there, the roughly 200-foot-long craft, built by the New Mexico–based company Sceye, will park some 18 kilometers above the ocean’s surface in the stratosphere, then use a custom-built antenna to supplement a 5G network, in a test that includes beaming data straight to devices.

Sceye (pronounced “sky”) is one of several firms building these high-altitude platform stations, or HAPS. Find out why they plan to connect us from the stratosphere.

—Rachel Courtland

This story is from the latest edition of our magazine, which is all about engineering. Subscribe now to get a copy, plus all our other issues and a range of subscriber-only content.

Longevity’s next frontier: “reprogramming” your body

Billions of dollars are flooding into efforts to reverse aging as scientists explore ways to return cells to a younger state. But how far off are these experimental treatments? Will they really work? At a virtual Roundtables event today, MIT Technology Review will examine the science behind the hype.

There are a variety of security concerns about artificial intelligence (AI), especially when it comes to the behavior of agentic AI. But until recently, the concept of locking down the models to prevent tampering hasn’t gotten a lot of attention.

Now, a security technology called “confidential computing” has emerged that could help solve that problem: it protects AI models from hackers by restricting models to authorized users. (It also protects data wherever it is — in storage, when moving between systems, and when it is accessed.)

With many top cloud and hardware providers championing confidential computing for AI, Computerworld talked with Dion Harris, Nvidia’s senior director of high-performance computing and AI factory solutions, about what the technology does and how it works.

width=”1024″ height=”722″ sizes=”auto, (max-width: 1024px) 100vw, 1024px”>

Dion Harris, Nvidia’s senior director of high-performance computing and AI factory solutions.

Nvidia

Why should organizations care about confidential computing now? “Seventy percent of data exists outside the cloud, in on-premise data centers and data lakes. To deliver AI on this data while maintaining security, you need confidential computing to unlock that use case for enterprise AI.”

Why is confidential computing suddenly important in the AI age? “Enterprises want AI on sensitive data — customer records, medical information, financial data — without exposing it in cloud environments where they lose control.

“Traditional encryption protects data at rest and in transit, but not during computation. When you run an AI model, you must decrypt data. It sits in plain text in memory, accessible to administrators and cloud operators. Confidential computing creates a hardware-rooted trusted zone where data is decrypted only when computation needs it, then immediately re-encrypted.

“This lets enterprises get AI value without compromising security. Financial services, healthcare, government, and regulated industries are adopting it.”

How is confidential computing changing with agentic AI? “We’ve gone from generative to agentic AI being deployed and used to solve real business problems. To deliver agentic capabilities with required privacy, security and performance in enterprise, confidential computing provides the unlock.

“We can spawn agents, access data, leverage tools and generate real useful work. With agentic AI, confidential computing helps in two ways: protecting the data and helping design implementation and workloads. It’s deployment, implementation, and use combined.”

How does confidential computing work? “Encryption at rest protects stored data. Encryption in transit protects data moving over networks. Encryption in use is the problem — when you compute on data, you must decrypt it in memory. 

“Confidential computing encrypts data in memory and between CPUs and GPUs. A dedicated element in the GPU decrypts information only when needed for computation, 100% inline, with minimal performance impact.”

Can you walk through a real-world example? “Apple’s standard operating policy for years has been to keep private information on device to avoid having access to private data. However, to leverage advanced AI models, they no longer fit on the device. Apple instituted their Private Compute Cloud, and now they’re extending that to Google Cloud.

“Let me give a hypothetical and hopefully describe how it works. If you have a user who wants to upload a medical transcript from their doctor, their system creates a secure, attested environment. It sends a request to the server: validate yourself. That attestation says: ‘I am a Nvidia GPU. This environment is secure. It hasn’t been tampered with.’ Now, it’s okay for you to send your information over that secure line. Remote attestation allows the edge device to ensure it’s sending to a trusted environment.

“The medical data comes over encrypted and remains encrypted until it lands in the GPU memory. Specific compute engines in the processor decrypt that information only to use it. The LLM dissects and summarizes it. Then it re-encrypts and sends it back over the wire.

“They get the capabilities of data center AI mode — larger, more advanced, delivering more services. But they also get the security and privacy from Apple’s [Private Cloud Compute]  platform. It’s the best of both worlds: efficiencies and intelligence from data center AI with Apple’s PCC security.

What prevented confidential computing adoption? “The main challenge in the past was significant performance impact. When you adopted confidential computing, you had to trade off performance for privacy. A 30% to 40% throughput reduction undermined the economic viability of the entire solution. If you sacrifice that much performance, it reduces the ability to get full utilization out of the hardware you’re deploying to deliver tokens or deliver a service in an economical fashion.”

How was that solved? “With Blackwell and newer GPU architecture, you can deploy confidential computing without impact to performance. You get both privacy and performance, a win-win scenario.

“Performance now translates directly into economics. You get full utilization of the hardware, which matters for delivering tokens or services at scale. When we built Blackwell, we made confidential computing a first-class system feature to deliver not just the security, but also the performance the market requires.”

Where is adoption accelerating? “Companies are thinking about the cloud. Model builders can’t expose APIs, enterprises customize for their business. Hybrid on-prem and cloud models — where builders deliver services behind enterprise walls — require zero trust. We’re no longer in fully owned infrastructure.

“By 2030, [billions of dollars] is expected in confidential computing use cases. It’s emerging as essential infrastructure for AI adoption across the industry. For organizations using cloud infrastructure, deploying AI on sensitive data, or operating under regulatory requirements, confidential computing is becoming essential.”

What should organizations do? ”The performance problem is solved. The technology is ready, ecosystems are built, partners enable it. If you want to deploy agentic AI on sensitive data, protect customer privacy, meet regulatory requirements, and maintain security across hybrid environments, you need confidential computing in your strategy.

“Most customers start with a developer license, validate performance and security, then migrate to a commercial license. Partners like Red Hat and Fortanix integrate these mechanisms within their platforms. Google Cloud offers it through their services.

“Start with a proof of concept. Validate it works. Plan deployment. Organizations gain competitive advantage securing AI on sensitive data.”

Memora also introduces a policy-guided retriever that, rather than returning the top-k semantically similar items in a single shot, iteratively refines its query, expands through cue anchors to surface related-but-not-similar memories, and decides when to stop.

“The deepest flaw in current agent memory is that it mistakes retrieval for memory. A vector store is superb at finding text that looks relevant. An enterprise agent needs more than resemblance. It needs to know what has changed, what still holds true, and what should never be recalled in the task at hand,” said Sanchit Vir Gogia, chief analyst at Greyhound Research.

Memora is interesting precisely because it refuses that shortcut, Gogia noted. It separates the rich detail of a memory from the handle used to find it, indexing a stable abstraction and a set of cue anchors while keeping the full content intact beneath them. Retrieval then becomes an act of navigation rather than a single hopeful guess, as the system re-queries, widens its search, or stops once it has enough, he added.

Samsung is expected to unveil its next generation of foldables at a Galaxy Unpacked event next month, but now we know what they might look like, courtesy of some leaked images published by Android Headlines. Images shared by the publication include case designs for two new Galaxy Z Fold 8 models and the Galaxy Z Flip 8, each giving a good look at the phones inside the cases too.

This could be our closest look so far at Samsung’s new wide-style foldable, following a leak of a purported dummy unit in May, and a fresh photo of an apparent dummy unit from leaker Ice Universe this week. There are some similarities between all of these leaks: they show a wide, passport-style form factor, with a selfie camera built into the cover display, and a double-camera setup on the rear.

Recent leaks suggest the wide foldable may in fact be called the Z Fold 8, with Samsung’s more familiar, narrower book-style foldable rebranded the Z Fold 8 Ultra. Images of what appears to be that phone, along with the Z Flip 8, were also shared by Android Headlines, though these look similar to the current Flip 7 and Fold 7 models.

Samsung’s new foldables are expected to get an official launch at the company’s summer Galaxy Unpacked event, which is reportedly taking place on July 22nd. That’s a little later than usual, but should mean that the wider Galaxy Z Fold 8 will launch before Apple releases its own wide-style iPhone foldable competitor, currently set to debut in September.

WhatsApp is introducing a new way to add and chat with contacts, without having to share your phone number. Usernames will launch “later this year,” in a move to make the communications platform “even more private,” allowing you to keep your phone number concealed from people who aren’t already in your contacts.

Usernames are available to reserve starting this week, giving you time to claim a specific handle you may already be using elsewhere before the WhatsApp username feature officially launches. The username launch will be rolling out gradually over the coming months, and users will be notified when the feature is available in their country.

“With over three billion people on WhatsApp a lot of names overlap, which is why we’re opening reservations early so everyone has the opportunity to select the username that matters to them,” WhatsApp said in its announcement.

You can reserve your username by heading to Settings > Account > Username in the latest version of WhatsApp on Android, or by tapping You, selecting your profile, and tapping on Create Username for iOS users. Individuals and organizations will have the option to claim their existing Instagram or Facebook handles, as long as it hasn’t been taken by another user. WhatsApp product head Alice Newton-Rex told The Verge that usernames for famous figures like celebrities and politicians have already been reserved to prevent them from being claimed — so if you share a name with a recognized public figure, you’ll have to create an alternative handle.

You can’t browse for people using their usernames, so you’ll need to know the exact username of a new contact before you can reach out to them. If you’ve already shared your phone number with WhatsApp contacts or group chats, the number will still be visible to them after you enable the username feature, so these privacy protections only apply to new conversations going forward.

WhatsApp is also introducing an “optional username key” that others will need to know before they can message you, to help you control who can reach you with a WhatsApp username if it’s made public without your consent.

Update, June 29th: Clarified how to reserve your WhatsApp username across both iOS and Android, now that the FAQ page is live.

A new proposal would ban the sale of Americans’ health and location information to data brokers — including information people reveal to an AI chatbot like ChatGPT or Claude.

In the coming weeks, Senator Elizabeth Warren (D-MA) and Representative Mary Gay Scanlon (D-PA) are planning to debut a new version of the Health and Location Data Protection Act that’s better suited to the AI era. The former version of the bill, first introduced in June 2022, prohibited data brokers from collecting and selling health and location data. Four years later, it’s expanded to ban other companies from selling such data to brokers, and to specifically cover data entered into AI systems.

AI labs have set their sights on building health and medical products. In January, Elon Musk publicly called for people to upload their medical records, like MRI scans, to Grok, xAI’s chatbot. That same month, OpenAI introduced ChatGPT Health, a sandboxed tab within ChatGPT that it deemed more secure, and encouraged users to upload their medical records and other sensitive information. It also introduced ChatGPT for Healthcare, aimed at medical providers. A few days after that, Anthropic quickly followed up with Claude for Healthcare, a “HIPAA-ready” tool for individuals, health providers, and hospitals.

But when battling data breaches or unauthorized access to data, users are largely at the AI companies’ mercy. The data protection for tools like OpenAI’s and Anthropic’s “largely depends on what companies promise in their privacy policies and terms of use,” Sara Gerke, a law professor at the University of Illinois Urbana-Champaign, told The Verge in January. The US lacks an overall federal framework for data privacy, despite years of attempts.

This bill — which is also sponsored by Senators Ron Wyden (D-OR) and Bernie Sanders (I-VT) — would require the Federal Trade Commission to enact the rules within 180 days and will allow the FTC, state attorneys general, and affected individuals to sue to enforce it. It would also earmark $1 billion to the FTC over the next 10 years for enforcement.

“It’s more important than ever that we crack down on data brokers that are raking in giant profits from selling Americans’ most sensitive information,” said Senator Warren in a statement. “Especially as more people enter their private health data into AI, we need to make sure that information isn’t exploited by the highest bidder.”

OpenAI is releasing some sort of device related to its AI-powered coding tool, Codex, on July 15th. In a video posted to X on Monday, OpenAI shows a square-shaped device with several buttons, alongside the caption, “Your favorite Codex shortcuts are getting an upgrade.”

This isn’t the mysterious AI-powered device OpenAI is working on with former Apple designer Jony Ive, however. As shown in the teaser, OpenAI is launching the device in partnership with Work Louder, a company that sells an array of mechanical keyboards and macro pads with mappable keys, dials, and switches.

The silhouette of the device shown by OpenAI looks a bit like Work Louder’s Creator Micro 2, a macro pad that comes with 13 mechanical switches, a joystick, and a touch sensor. Users can assign shortcuts and custom actions to the toggles on that device when using different apps, like Photoshop. Figma similarly partnered with Work Louder to launch a macro pad with preconfigured shortcuts.

Neither OpenAI nor Work Louder shares any more details about their upcoming collaboration, but with a launch just two weeks away, we won’t have to wait much longer to find out more.

Q: Considering consumer behavior, could you update us on your current thinking regarding hardware pricing and profitability? For the next-generation platform, is it reasonable to assume that your pricing will continue to prioritize profitability of the hardware, as it does today?

A: First, we regard hardware as the base for providing the gaming experience, and by offering products such as the PlayStation Portal Remote Player (PS Portal), we aim to provide experiences tailored to users’ play styles beyond the living room, which has traditionally been considered the primary usage environment. As for pricing, it is not realistic for us to absorb all component cost increases, and we have already implemented some price increases outside Japan. At present, however, sales are proceeding as planned, and we do not believe this has led to a decline in customer demand. As a principle, we do not intend to sell hardware at significant losses. At the same time, we are carefully monitoring the market and continuing to evaluate our approach. We believe it is important for us to make every effort to ensure that customers fully understand the value we provide in relation to pricing.

Leaked iPhone 18 Pro photos and parts lists appeared on the dark web following a data breach affecting one of Apple’s key suppliers, according to a report from Reuters. The leaked images show a drop test of what a source tells Reuters is the iPhone 18 Pro, equipped with a three-camera layout and Apple logo.

In addition to images of the iPhone 18 Pro, Reuters reports that at least six files detail the “hundreds” of components used in the device, along with a list of their suppliers. AppleInsider similarly reported last week that the leaked documents show the purported board layouts for the iPhone 18 Pro and 18 Pro Max, as well as what appears to be data sheets for the rumored A20 Pro chip.

Apple is “concerned about the documents being shared on the dark web,” according to Reuters. The Verge reached out to Apple with a request for comment but didn’t immediately hear back.

Though hopes were high for Supergirl, the movie has turned out to be a bit of a dud. Critics have been rather down on the project, and its lackluster box office performance has it on track to lose WBD somewhere between $100–120 million. Films flop all the time, and Supergirl not resonating with audiences probably wouldn’t be a huge deal if we knew that DC Studios had more exciting things coming down the pike. But Supergirl feels like it could be an early sign that Gunn’s grand plan for the DCU is falling apart before it even really gets off the ground.

Loosely based on Tom King and Bilquis Evely’s Supergirl: Woman of Tomorrow comic miniseries, the new Supergirl film follows Superman’s cousin Kara (Milly Alcock) as she embarks on an interstellar bender that culminates in her dog being poisoned by crew of sex-trafficking pirates. Unlike Superman (David Corenswet), Supergirl doesn’t really have a problem with killing her enemies — especially when they’re trying to stop her from saving Krypto. But with an orphaned girl (Eve Ridley) tagging along for the adventure, Kara tries to set a good (read: no murdering) example.

Supergirl struggles to make its titular heroine feel distinct from Superman

Though Supergirl comes from director Craig Gillespie and writer Ana Nogueira, everything about this movie — from its focus on animals in distress to its needle drops — makes it feel a lot like some of Gunn’s previous work. Supergirl’s drunken brawls in alien bars and scenes of her schlepping around space in a junky starship look like they could have been ripped from any one of Marvel’s Guardians of the Galaxy features. You can hear Nogueira channeling Gunn’s spiky sense of humor as the movie introduces new faces like unhinged bounty hunter Lobo (a distracting Jason Momoa in comics-accurate garb). Momoa’s presence is a constant reminder of how the DCEU fell apart, but Lobo isn’t really what drags Supergirl down.

As refreshing as it was to see Superman gloss over Clark Kent’s oft-repeated tragic backstory, Supergirl spends much of its runtime rehashing the details of Krypton’s destruction. Flashbacks to Kara’s past are meant to help us understand the grief she’s been living with, and to see why her sense of morality is very different from her cousin’s. But rather than unpacking Kara’s emotions in any meaningful way, the movie makes light of her substance abuse while sending her on a by-the-numbers adventure that’s generally lacking when it comes to intrigue or visual spectacle.

One of Supergirl’s more glaring issues is the way it struggles to find organic ways to make its titular heroine feel distinct from Superman. Aside from her relative brutality and moody outlook, she’s just another indestructible alien who periodically needs to recharge her powers by basking in yellow sunlight. The movie tries to give itself some stakes by constantly putting Kara in situations where she’s left without her abilities. But by the second sequence in which Kara’s getting punched out by a bunch of dudes, you get the sense that DC Studios never really locked in on a plan to make this story pop.

That’s somewhat surprising given the way Gunn has previously insisted that DC Studios would “never put a half-assed script in production” simply because the project had already been announced. Half-assed is the perfect description of Supergirl’s entire vibe, and it being the studio’s second major feature doesn’t exactly bode well for the DCU’s future. Supergirl needed to demonstrate that Gunn had a solid plan to build a new universe on the backs of some of DC’s lower profile characters. Though we’ve already seen some of how that could work in HBO’s Peacemaker series, it was less clear whether the studio could pull it off on the big screen. The entire point of rebooting WBD’s superhero movies was to put DC Studios in a better position to compete with Marvel — which is on the verge of its own major reset. But whereas Marvel has a few reliable aces like the X-Men and a new Spider-Man movie up its sleeve, DC is essentially starting from scratch.

Some of Supergirl’s problems might not be so readily apparent if there had been more time before it and Superman’s theatrical debuts. The two movies coming out so close to one another emphasizes their characters’ general similarities, and makes it seem like DC might be a little too comfortable putting out iterative projects. This calls into question Gunn’s decision to prioritize a series about the Green Lanterns and a Clayface film before introducing new versions of more well-known heroes like Batman and Wonder Woman. WBD still plans to put out a sequel to Matt Reeves’ The Batman that won’t technically be part of the DCU, but the Gotham of it all may get audiences primed to see a new Bane / Deathstroke movie that the studio is reportedly prioritizing in the wake of Supergirl’s underperformance.

All of these B-tier projects and alternate realities give the nascent DCU a whiff of the same messiness that has plagued Sony’s universe of Spider-Man spinoffs since its inception. And when you factor in WBD’s impending merger with Paramount Skydance, it seems very possible that the DCU might not come together the way Gunn originally intended. Though it’s possible that next year’s Man of Tomorrow could steer things in a stronger direction, what feels more likely right now is DC putting out another Super-movie that feels a little too similar to what we’ve seen before. It wouldn’t be the first time that WB found itself on the ropes with a comics-related crisis, but it might be the last chance the studio has to get this stuff right.

The ultimate promise of agents is not only to automate tasks but to manage and coordinate entire workflows, pursuing business goals in a way that allows humans and agents to work together. Given the risks involved in automated decision-making, teams cannot delegate the work that agents do without confidence that they are fully capable of performing the task and that it will do so in a safe, reliable, and secure manner.

Among technology experts, our research shows that teams are exceedingly confident about using agentic AI across a significant amount of AI, data, and cloud tasks.

Where agent readiness drops is largely due to a lack of business context being supplied to agentic systems. The more complex the task, the more reasoning capability an agent requires and the greater its need for business context. Such context-generation capabilities for agents are still at an early stage of development, especially in situations where enterprise data is difficult to wrangle and connect into the agent lifecycle at the speed and quality in which developers and executives need it. Human oversight is a key factor of success in deploying agentic AI.

Knowing that tech teams are in a pivotal position to lead this transformation, the experts we interviewed expect agent confidence to accelerate as experience with agents deepens and business environments mature. “As we design agents to operate within the same operational boundaries, identity systems, and governance models that teams already use, they start to behave more like the systems organizations already trust,” says Jeremy Winter, corporate vice president and chief product officer at Microsoft Azure Platform.

This report, based on a survey of 300 global technology experts, ranks 101 tasks across AI, data, and cloud workflows based on respondents’ confidence in agents acting on their behalf. It also examines how technology teams view the opportunities and challenges related to agentic AI, along with the potential for the technology to enhance their careers.

Key findings from the report include:

Confidence in agents is surging for measurable tasks and growing in areas of complex judgment. Technology experts overwhelmingly believe agents help with everyday work including streamlining processes, improving performance, and reducing repetitive tasks. Confidence is highest for processes like generating reports and boilerplate code, and there is clear opportunity where tasks involve multistep workflows and advanced reasoning to make decisions.

Data workflows are the breakthrough domain. Tech teams trust agents most where structure can provide a reliable foundation for decisions. This includes areas such as data quality monitoring, visualization anomaly detection, real-time data stream monitoring, and data profiling. This is where domain experts closest to the point of data generation can provide context to allow agents to act and deliver trusted outcomes.

Tidal shared its new policies regarding AI-generated music today and how the platform plans to “protect artists” and “inform listeners.” Instead of banning it outright, starting on July 15th Tidal will label tracks it has identified as being 100 percent AI-generated with an icon. But starting today those tracks will no longer be monetizable. “Tidal’s priority is ensuring royalties go to original works directly produced, written, and performed by people. We will therefore not knowingly attribute royalties to music we identify as wholly AI-generated,” the company’s announcement reads.

The platform didn’t specify what tools it’s using to identify AI-generated music, but the new policy says that as tools improve and become more reliable it eventually plans to also add the label to any uploads that are “substantially AI-generated.” In addition to its own detection tools, Tidal warns that identifying AI-generated music should not be its responsibility alone and that the platform will “begin to enforce” an expectation that content distributors properly label AI-generated music as well.

In an effort to deter the use of any AI generative tool that “exploits an individual’s or group’s music, name or likeness, deceives listeners, or diminishes the quality of our service,” starting in mid-July Tidal will also remove or block “AI-generated music associated with fraudulent activity.” Those activities can include music designed to deceive listeners or interfere with authentic artists, high-volume uploads, or “unusual streaming activity.”

Tidal’s competitors have previously addressed the influx of AI-generated music with tools and policies of their own. In April, Spotify launched a new verification program where some artists the company has confirmed as being a real person received a green checkmark and a “Verified by Spotify” badge on their profiles, while profiles that primarily upload AI-generated content are not eligible for the distinction. Deezer also developed tools for detecting fully AI-generated music when it’s uploaded in order to reduce its visibility on that platform, and last month it created a website you can use to scan your playlists on other streaming platforms to detect AI-generated tracks.

If you’re anything like the managers recently studied by Emma Wiles, a Boston University business professor, treating Alex as a “coworker” and not a software tool would lead you to do a worse job. Wiles found that people caught 18% fewer errors when the work was said to have come from an agentic “AI employee” rather than a chatbot. It turns out that what’s in a name matters. A lot. 

This is an alarming glimpse of the future Silicon Valley is hurling us toward. Last year Nvidia’s CEO, Jensen Huang, talked about workplaces of “digital humans.” Since April, Microsoft, OpenAI, Anthropic, and Google have all released new tools oriented toward managing teams of AI agents, many of which are explicitly advertised as digital colleagues with the flexibility and cognitive power of actual humans. And nearly a third of the 1,261 managers who participated in Wiles’s study said their companies already frame AI agents as employees (23% even list them on org charts).

The technical progress of agentic AI is not all hot air, of course. Agents, which can effectively be thought of as AI tools programmed to work in a loop until they achieve a goal, have become measurably better at more complicated tasks. But it’s a huge leap to refer to these tools as coworkers or employees, and doing so will set unrealistic expectations for what AI can do while leaving the human employees supposedly responsible for them worse off.

That’s partially because, Wiles’s research suggests, it inverts our sense of who’s in charge. When an AI tool was framed as an employee, participants in the study saw themselves as less responsible for its output. They were also 44% more likely to escalate its questionable work to a manager for further review rather than trusting their own corrections (thus negating the time-saving purpose of using the AI agent in the first place). 

That matters far beyond office culture: As AI agents are embedded into health care, warfare, education, and government, there’s a growing risk they’ll become a convenient place to dump blame for failures that are instead the product of bad human decisions, incentives, and oversight (recall how the bomb strike on a girls’ school in Iran was popularly blamed on Claude, when all signs point to a cascade of human errors).

Most of Apple’s price increases have gone into effect, resulting in iPads and other products costing hundreds more than they did a few days ago. If last week’s Prime Day sale wasn’t a good time to consider buying an iPad, we found a deal worth considering that’s just under $500 right now. The 128GB iPad Air that has the M3 processor, plus both Wi-Fi and 5G cellular support, costs $499 at Best Buy. At just $50 more than the base iPad (now $449, up from $349), it’s a significantly faster and all-around better tablet.

While the last-gen Air matches the base iPad’s 11-inch screen size, the Air has a laminated, antireflective coating, plus a wider color gamut. In other words, it’s the better choice for entertainment or for creating content. In terms of specs, its M3 processor leaps ahead of the A16 with more cores, not to mention Apple Intelligence support (as well as support for iPadOS 27 coming later this year). It also boasts Apple Pencil Pro support.

Google offers a dizzying array of Gemini modes and options and an equally overwhelming series of AI subscription plans that control how much you can use those capabilities, but you don’t need to worry about any of that to create custom Chrome extensions. You might sometimes see better results if you switch your Gemini model to “Pro” or your Gemini thinking level to “Extended” — designations that even Gemini itself has trouble deciphering (believe me, I asked!) — but just using the default Gemini settings with a free Google account will generally work quite well.

Getting going with a custom Chrome extension is as simple as opening up a new Gemini chat and telling the system what you want it to cook up for you. The hardest part is deciding what you want and what’d be helpful for you — something we’ll explore more in a moment, via specific examples and suggestions. Once you’ve got that, you can just ask Gemini to create a Chrome extension that’ll accomplish what you’re envisioning, with as much specificity as possible about what it’ll do and how it’ll look.

Gemini will spit back a series of plain-text code chunks with instructions to copy each cluster and paste it into a new plain text file with a certain specific name — things like “manifest.json,” “content.js,” and “styles.css.” All you’ll do is use the on-screen button to copy each segment, then open up any simple text editor (like Windows Notepad, macOS TextEdit, or any number of simple online text editors) and paste the text in, then save it under the name Gemini gives you.

It’s lazy thinking to see Apple as a villain in this scenario. The company might have been charging more for add-on memory than market rates, but there were real technical reasons to do so. And while critics might be castigating Cupertino for those past practices, they’ll still find themselves now paying more for whatever brand of electronic devices they use to write their screeds on in future.

It’s all about supply and demand. Memory manufacturers see the opportunity to feed AI need, even if it means sacrificing consumer markets as they do.

Cash through chaos

You can argue that the consequences of that decision are unethical. Should memory makers have considered the consequence of curtailed supply on their existing markets? After all, every business, every school, and almost every consumer is now a digital entity, and the massive increase in PC, smartphone, and other consumer electronics prices will have a consequential impact across all layers of society.

Italy’s competition watchdog has opened an investigation into Microsoft over concerns it may not have clearly informed consumers about the integration of Copilot and Designer into Microsoft 365 subscriptions, associated price increases, and automatic upgrades to higher-cost plans.

The Italian Competition Authority (AGCM), in a statement to the press, said it had opened an investigation into Microsoft S.r.l., the vendor’s Italian subsidiary, and Microsoft Ireland Operations Ltd. to assess whether the way the changes to M365 pricing were communicated may have unduly restricted consumers’ freedom of choice.

Although the AGCM’s announcement does not explicitly identify the pricing event under investigation, its description aligns with Microsoft’s January 2025 rollout of Copilot and Designer for Microsoft 365 Personal and Family subscribers.

At the time, Microsoft announced its first price increase for the consumer subscriptions since the launch of Copilot and Designer, stating that existing subscribers would pay the higher price at their next renewal after the AI features were added.

Could scrutiny spread beyond Italy?

The Italian probe is not the first time Microsoft’s communication around AI-related Microsoft 365 pricing has drawn regulatory attention.

The investigation follows earlier scrutiny of Microsoft’s consumer pricing communications in Australia and New Zealand, where Microsoft apologized and revised some of its messaging after regulators raised concerns over how AI-enabled Microsoft 365 subscriptions were presented to customers.

That episode could prove relevant for the current investigation, even though the legal frameworks differ, said Pareekh Jain, principal analyst at Pareekh Consulting.

“Microsoft’s apology and revised communications show similar concerns were raised before,” the analyst said.

“While the legal cases differ, Italian regulators may see it as evidence that clearer customer communication was already known to be necessary,” he added.

Jain also expects regulators elsewhere to watch the outcome of the Italian investigation: “Regulators in the EU, UK, Australia, New Zealand, and Canada are likely to watch closely, especially where AI is bundled into existing subscriptions with higher prices or automatic renewals.”

Enterprise buyers are likely to scrutinize AI pricing more closely

Whether the Italian investigation ultimately results in penalties against Microsoft remains to be seen, but analysts say the probe could have an impact on M365 commercial plans, which are separate and set to take effect on July 1.

The investigation serves as a reminder for enterprises to examine AI-related price changes even if Microsoft’s commercial licensing process seems more transparent than its consumer subscription model, Jain said. “CIOs should still verify what is changing at renewal and whether AI features are optional.”

Enterprise procurement teams should ask tougher questions during negotiations, said Bhupendra Chopra, chief revenue officer at IT consulting firm Kanerika.

“Procurement teams should ask questions, such as what AI am I paying for, can I see it itemized, and can I decline it without losing the rest? Buyers would be right to want AI pricing written into renewal terms with clear opt-outs and price protection,” Chopra said.

These questions, according to Jain, could provide additional leverage during licensing discussions to negotiate flexible terms.

However, Chopra pointed out that the investigation is symptomatic of a broader tension between software vendors’ efforts to integrate AI into existing products and regulators’ expectations around transparency and customer choice.

“Building AI into existing products and pricing it in is becoming standard across software, not unique to one vendor. Regulators are testing an old question against a new feature — whether buyers were given clear information and a real choice,” the analyst said.

For enterprises, the analyst added, the practical takeaway is straightforward: “Expect AI to show up inside the tools you already own, expect it to carry a cost, and review what each renewal actually includes. Treated as routine cost discipline, it stops being a surprise.”

Microsoft said it is committed to complying with Italian consumer law and will cooperate with the Italian Competition Authority in its preliminary investigation.

Like a lot of people bitten by the self-quantifying bug, I started gathering personal data to pursue a nebulous collection of goals and desires. I wanted to feel better physically and emotionally, get outside more, and bring order to the messiness and uncertainty of my daily existence.

But external metrics and data can never capture what’s truly important. Worse, they inevitably redefine your core sense of what’s important, whether you’re aware of the trap or not.

Dive into the dangers of quantifying our lives with metrics.

—Bryan Gardiner

This story is from the next edition of our magazine, which is all about engineering. Subscribe now to get a copy when it lands!

Elephant alert! AI warning systems aim to avoid deadly clashes

India is home to about 60% of the world’s wild Asian elephants, and around 80% of their habitat lies outside protected areas. That brings them into close contact with people, and clashes can turn lethal: there have been some 3,000 human casualties in the last five years and over 1,000 elephant deaths since 2014.

In response, state forest departments, NGOs, and locals are designing, testing, and deploying a range of AI systems that cut response and warning times to minutes—or even seconds. They range from wildlife eyes in Maharashtra to infrared drones in Chhattisgarh.

Find out how they work in our interactive map.

Ending any story is hard, but that’s especially true of mystery-packed TV shows. Series like Lost initially hook viewers with constantly building secrets and questions, to the point that they can often seem incomprehensible. But the promise is that it will all pay off in the end — a feat that few shows ultimately manage. This creates tension as viewers try to put all of the pieces together while hoping that the show’s creators know what they’re doing in the long run. The latest example is From on MGM Plus, a Lost-style horror series that just wrapped up its penultimate season, with a fifth and final one coming in 2027. Naturally, endings are currently top of mind for the team behind the series.

“Making this type of show, you’re making a contract with the audience,” says From’s creator, John Griffin. “You’re saying, ‘Look, if you guys get invested, we promise we’re going to take you somewhere worth going.’”

From tells the story of a group of people in a small town that they’re unable to leave, all while dealing with prowling monsters, a dwindling food supply, and other existential threats. The season 4 finale was perfectly indicative of how these shows often work, mixing intense drama and bizarre mysteries. A beloved character turned into a monster, and there were more than a few unexpected deaths, strange and violent weather patterns, potentially fatal dream sequences, and some upending of the rules that everyone has had to live (and die) by over the last four seasons.

Image: MGM

Tracking everything that has happened so far is a huge challenge with a show like this — both for viewers following along at home and those working on the show. According to showrunner Jeff Pinkner — who previously worked on series like Lost, Alias, and Fringe — the solution is fairly straightforward: remembering what’s most important.

“One of the challenges of making a show this complicated is that we’re expecting an audience to watch a show this complicated,” he explains. “If we had some kind of master document that we were relying on, or that was necessary for us to hold it all in our heads, we would be expecting too much of the audience. The honest truth is, we hold it all in our head because that’s what the audience has to be able to do. We only want it to be as complicated as we can pay attention to.”

Knowing how the series will end from its very beginning helped the creators of From stay focused on reaching that conclusion even as they delved into side mysteries. Unlike many shows, they always had a specific end goal in mind while making narrative decisions. That doesn’t mean that the show hasn’t changed from the initial vision, though. Griffin likens the experience of creating the show to that of a road trip, where there isn’t necessarily one straightforward journey, but the final destination remains the same. “We set out with an intent, and on any journey things change and things evolve along the way,” he says. “Some of what was planned in the beginning was jettisoned in favor of other roads.”

According to Jack Bender — an executive producer who directed many of From’s episodes, and is perhaps best known for directing the emotionally charged Lost finale — one of the strengths of long-form mystery television is its team’s ability to change and adapt. While the creators might have concrete ideas in mind initially, those can shift based on everything from fan feedback to input from other members of the crew, whether it’s the actors or production designers. “That’s one of the great creative things about this kind of storytelling, where you have 50 episodes to tell this story,” Bender says. “It gives you time to go off into the woods and take little detours, and still get back on the path of where you’re going.”

Image: MGM

One of the difficulties, however, is that there are no guarantees in the world of television. From’s creators may have had an ending in mind from early on, and they may have planned for a story that spans five seasons, but getting there depended on the show reaching an audience and getting renewed multiple times. That outcome wasn’t assured. It’s a reality that most TV writers have to contend with, but it becomes especially tricky for a series like From that has definitive plans for its beginning and ending. Contingency plans become a necessity.

“If there came a time when MGM Plus had come to us and said, ‘Hey, listen guys, the numbers are bad, we’re going to have to wrap this up next season,’ could we have done it? Sure,” explains Griffin. “There are 9,000 ways to tell any story. But the fact that we got to let the story breathe, and let it lead us to where it wanted to go, and fulfill the original vision we all had, has been incredibly gratifying.”

All of those factors combined make pulling off a successful ending all the more significant — and rare. As M. Night Shyamalan told me ahead of the finale of his Apple TV thriller Servant, “I’m astonished now when I think of any peers who have done this.” In the case of a mystery show, the finale has a lot of ends to tie up and mysteries to reveal.

But according to the team behind From, checking unresolved storylines off of a list isn’t the goal; it’s to make viewers feel intensely about the world and characters in order to leave a lasting impression. As controversial as the Lost finale was, it’s still something people talk about today. “You only miss characters that you care about,” says Griffin. “You only miss shows that you care about.”

“We want the ending to feel surprising and simultaneously inevitable,” adds Pinkner. “We want the ending to feel like it was set up in the first frame of the first episode.”

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Andrew WebsterClose

  Andrew Webster

  Senior entertainment editor

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Andrew Webster

• EntertainmentClose

  Entertainment

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Entertainment

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

• TV ShowsClose

  TV Shows

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All TV Shows

Comcast has announced plans to separate itself into two publicly traded companies, spinning off its NBCUniversal and Sky broadcasting arms. The shake up aims to protect the media conglomerate’s profitable broadband and wireless brand, which will retain the “Comcast” company name, as its media and entertainment business — now collectively named “NBCUniversal” — faces increasing pressure from industry consolidation and streaming rivals.

The separation is expected to take approximately a year, with Comcast shareholders owning shares in both Comcast and NBCUniversal upon completion. While current Comcast CEO Brian L. Roberts will be “actively involved in the leadership” of both companies, co-CEO Mike Cavanagh has been selected to lead NBCUniversal, and Comcast’s former chief financial officer Michael Angelakis will become the Comcast CEO after the split.

If the agreement goes ahead, Sky Media — the British broadcaster that Comcast acquired in 2018 — alongside the company’s theme parks division, Universal film and television studios, NBC, Peacock, Bravo and Telemundo networks will be housed under the new NBCUniversal entity. The new Comcast, meanwhile, is expected to keep serving customers through its unspecified “broadband, wireless, and entertainment platforms.”

“Both companies begin this next chapter from positions of strength. Comcast will continue to build on its leadership in connectivity, while NBCUniversal, together with Sky, will have the scale, brands, content and financial resources to compete as a premier global media and entertainment company,” Cavanagh said. “I’m personally thrilled to continue leading NBCUniversal into the future.”

China’s Zhipu AI (Z.ai) released its open-weight GLM-5.2, and some researchers have claimed that it matches Mythos in certain bug-finding and cybersecurity scenarios. While GLM lags behind models from Anthropic and OpenAI in other, more general tasks, it seems that China has dramatically reduced the gap in the capabilities between its models and those of the US.

This level of advancement is particularly concerning to the US government, which has worked to restrict China’s access to powerful models like Anthropic’s Mythos and Fable, as well as the hardware necessary to train and run them. The Trump administration views Mythos and other advanced AI models capable of identifying vulnerabilities as serious national security threats. Recently, OpenAI unveiled GPT-5.6, which has also raised concerns about its potential for misuse and has limited access to it.

Because GLM is an open-weight model, it can be downloaded and run by anyone on readily available hardware. That gives it great flexibility and allows power users deep access, but it also makes it ripe for abuse by bad actors who can run it with little oversight.

Suno has ambitions to be more than just a toy to churn out AI slop, it also wants to be a streaming destination and to break new artists. Spark is their new incubator program for independent artists that provides grants, mentorship, and marketing support.

To apply, artists need to be an unsigned singer, songwriter, or producer releasing music under their own name. They also need to agree to some terms and conditions that have raised some eyebrows over on the Suno subreddit. For one, you need to agree to make your songs available on Suno for remixing. That’s not necessarily super concerning, but the broad license it grants Suno to your works, including the ability to create derivative works, is. By agreeing, you also waive your right to a trial and to participate in a class action, and give Suno limited exclusivity to your material. Suno is already facing a proposed class action lawsuit from a group of independent artists.

Perhaps most alarming is the “Good Vibes Only” confidentiality and non-disparagement clause that requires you to promote Suno and give the company the right to request edits and removals of your content. It says the participant “will not at any time make any statements or representations, either directly or indirectly, whether orally or in writing, that portrays Suno, Suno personnel, and/or any Suno products or services in a negative light,” and doing so could mean being removed from the program.

Despite trade restrictions, China has reclaimed the title of the world’s fastest supercomputer for the first time since 2018. LineShine has pushed El Capitan out of number one on the TOP500 ranking. That’s despite strict limits on what high-powered computing components can be sold to China by US firms, which dominate the list, with America holding three of the top five spots. LineShine doesn’t even use any GPUs, which are typically the backbone of modern supercomputers.

While reaching the peak of the Top500 carries obvious bragging rights, it also serves as a message from the Chinese government to the US. The Trump administration has sought to limit China’s access to chips from firms like NVIDIA and placed steep tariffs on products going in and out of the country. China responded by building around more readily available and generalized CPUs. LineShine uses roughly 45,000 LX2 processors, each with 304 cores running at 1.55GHz, connected over a special high-speed, low-latency network called LingQi.

LineShine is the first supercomputer to cross the 2,000 exaflop barrier and is 20 percent faster than the number-two system, El Capitan, on the Top500 list. However, it also uses 42.2 megawatts, dramatically higher and less efficient than El Capitan’s 29.7 megawatts.

I’m sure we’re all familiar with Dark Crystal, so we know that Jim Henson can be weird and tackle slightly more mature subject matter. But there is little in his oeuvre that is quite as mind-bending as the Muppetless The Cube. This 1969 teleplay was produced for an NBC anthology series called Experiment in Television, which featured, appropriately enough, various experimental films, plays, and documentaries. One episode even featured Marshall McLuhan explaining his oft-cited theory that “the medium is the message.”

Even among all these oddities, however, Jim Henson’s The Cube stands out. It’s a 53-minute bottle film — taking place almost entirely in a single room. A man awakes in a white cube, unsure of where he is or how he got there. There are no windows, no door. Just walls of white panels.

It doesn’t take long for someone to open a section of the wall and bring in a stool for our nameless man in the cube. But when he closes the “door” behind him, our protagonist can’t open it back up. And thus begins the parade of people, dozens of them, taking turns going in and out of various invisible doors in the titular cube.

The interactions start off strangely enough — why is there strawberry jam on the stool? Who is this woman who claims to be the protagonist’s wife even though he doesn’t recognize her? But they quickly escalate, calling into question the nature of reality, our protagonist’s sanity, and raising questions about what the cube is exactly. Jim Henson himself even makes an uncredited cameo as the voice of a gorilla in a tutu.

As people come and go, delivering supplies to the man, harassing him, or even attempting to seduce him, the room changes around him inexplicably. Beds, couches, fully stocked liquor cabinets, and other furniture mysteriously appear. A full band slips in and sings a song with the line “you’ll never get out ‘til you’re dead,” before it’s revealed to be a recording as the record skips repeatedly on the word “dead.”

The Cube offers many questions but no answers. Is the man living in a simulation? Is he on TV? Are the people around him actors? Is any of it real at all? Does matter exist?

Even in a post-Twilight Zone world, The Cube feels uniquely bizarre, more akin to the modern dystopian anthology series Black Mirror than anything else. While it’s not true lost media, it remains relatively obscure. It only aired twice, there’s a sold-out DVD listing on Amazon, and it only occasionally makes an appearance on streaming services in any official capacity.

Your best bets right now are a pair of YouTube uploads, both embedded above. One is a much higher-quality transfer of a black-and-white kinescope film with remastered audio. Unfortunately, it also cuts out most of the song due to copyright. The other upload is full color and retains the song, but is a generally lower quality rip with muddier image and audio. Regardless of which one you choose, it’s a wild and thoroughly enjoyable ride that shows just how twisted the mind of Jim Henson could be.

Jonathan Rinderknecht was facing arson charges for setting a fire on New Year’s Day in 2025, which became one of the deadliest wildfires in LA history. To make their case, prosecutors turned to location data from his iPhone, security camera footage, and witness testimony. But they also turned to his ChatGPT logs.

Prosecutors said that Rinderknecht had ChatGPT generate images of fire, asked the chatbot, “Why am I so angry all the time?”, and ranted to it about how the wealthy were destroying the world. They also pointed to a screen recording in which Rinderknecht asked ChatGPT whether someone could be blamed for a fire if it was lit by their cigarette.

But the jurors were unconvinced. The trial ended in a deadlock when the jury voted 10-2 in favor of the defense. That led the judge to declare a hung jury and a mistrial.

One juror told CBS LA that she didn’t believe the ChatGPT logs were proof of anything, saying, “I talk to ChatGPT all the time.” She said it actually made her “angry” that they were suggesting his use of the chatbot indicated some sort of character flaw.

This is The Stepback, a weekly newsletter breaking down one essential story from the tech world. For more news about the streaming industry, follow Emma Roth. The Stepback arrives in our subscribers’ inboxes at 8AM ET. Opt in for The Stepback here.

Streaming was once a reprieve from cable. Not only could you watch whatever you wanted at any given time, but you didn’t have to sit through five-minute-long commercial breaks. And the best part was the price: Netflix, for example, cost just $7.99 / month when it launched its standalone streaming service in 2010. Amazon’s Prime Video was the same, offering ad-free streaming as a perk with its Prime membership. Hulu was an outlier at the time, but that’s because it allowed you to stream ad-supported TV shows and movies for free. It later added a $7.99 / month Hulu Plus subscription, with limited advertising.

But the commercial-free streaming model was so appealing that when companies launched rival streaming services, most made them ad-free by default. Disney Plus cost $6.99 / month when it launched in 2019, followed by the $4.99 / month subscription to Apple TV, and the launch of HBO Max’s $14.99 / month streaming service. (There were some exceptions, of course, like Paramount Plus, which launched with ads for $4.99 / month, and NBCUniversal’s Peacock, which offered some titles for free alongside a $4.99 / month ad-supported tier with access to all of its shows and movies.)

As the streaming industry matured and viewers settled into the services of their choice, many executives soon realized that they couldn’t make money just by adding new subscribers in an already saturated market. Moreover, Netflix lost subscribers for the first time in more than a decade in 2022, while the streaming businesses run by its competitors remained unprofitable. Streamers needed a solution — especially so they could keep spending the billions required to fill their libraries with movies and TV shows.

In between steady waves of price hikes, major streaming services have adopted advertising one by one in a bid to boost profitability. HBO Max first rolled out a cheaper, ad-supported tier in 2021. Even Netflix, whose former CEO Reed Hastings said the service would never have ads, walked back on its promise by introducing its ad-supported plan in 2022. Disney Plus followed suit, and Amazon Prime Video automatically placed its subscribers in an ad-supported tier, forcing them to pay extra to get access to commercial-free viewing.

Now that nearly all of the most popular streamers have ads, the pricing gap between ad-supported and ad-free tiers has grown wider. Streaming services often target ad-free tiers with the biggest price increases, with Netflix most recently raising the price of its standard and premium plans by $2, compared to $1 for its ad-supported tier. Now, the price to go ad-free on Netflix is $19.99 / month (or $26.99 if you want 4K HDR) — more than double the original $7.99 / month subscription.

HBO Max similarly issued a $2 hike for ad-free subscriptions last year. It now costs $18.49 / month for its standard plan or $22.99 / month for premium, a sizable leap from its original $14.99 / month ad-free subscription. Prime Video also doubled the price to go ad-free earlier this year and locked 4K streams behind the pricier, $4.99 / month subscription that gets tacked onto your $14.99 / month Prime membership. And Disney Plus, which once offered an ad-free subscription for $6.99 / month, now costs $18.99 / month commercial free.

There’s a strategy behind this: Executives across the streaming industry have reported earning more average revenue per user (ARPU) on ad-supported tiers, as they rake in cash from viewer subscriptions and the ads they sell to them. In January, Netflix said its advertising business earned $1.5 billion in 2025, making up just a fraction of the company’s $45.2 billion total revenue. But Netflix’s ads tier is growing — now reaching more than 250 million viewers every month — and the company expects its advertising revenue to double to $3 billion this year.

More people are signing up for these ad-supported tiers as commercial-free plans grow more expensive. Among streaming services that offer cheaper, ad-supported plans, nearly half of their current subscribers in the US have signed up to the tier with ads, according to research from the analytics firm Antenna.

Apple TV remains the only major streaming service that doesn’t offer an ad-supported subscription — though questions remain about how long it will hold out. Apple currently shows ads during live broadcasts like Major League Soccer games. While some have speculated about Apple bringing ads to the rest of its service, the company pushed back on rumors, with its head of services Eddy Cue saying, “I don’t want to say no forever,” but “there are no plans.” Apple TV has steadily raised prices alongside other streamers and now costs $12.99 / month.

Meanwhile, Netflix is finding more ways to show ads to viewers, such as allowing brands to use AI to “blend” their ads with one of Netflix’s shows or movies. Other streamers are experimenting with different kinds of ads as well, with Peacock showing them on the profile selection menu and others invading paused screens. Unfortunately, viewers on a budget really have no other choice but to deal with ads — many of which are very repetitive or appear at the wrong times.

That’s why some viewers are looking toward alternatives, like the ad-filled but free-to-watch YouTube, which remains Netflix’s biggest rival, and free ad-supported streaming TV (FAST) services such as Tubi, Pluto, and The Roku Channel. Some services are trying to balance affordability with ad-free streaming, with the Roku-owned Howdy offering over 10,000 hours of TV shows and movies without commercials for $2.99 / month.

While some of these options may serve as a temporary haven from high prices, they just don’t have the premium titles streaming has become known for. It’s become more expensive than ever to access shows and movies without interruptions, taking away what drew most people to streaming in the first place.

• Fox’s plans to acquire Roku could benefit both companies: Roku by potentially being able to tap into Fox’s library, and Fox by getting a hold of Roku’s trove of user data.
• Surging subscription costs have led some viewers to pirate cable using rogue streaming boxes.
• As Netflix snaps up YouTube’s top talent, YouTube is hitting back by launching exclusive shows of its own, starring big names like Trevor Noah and Alex Cooper.

• Janko Roettgers, who writes the Lowpass newsletter, dives into what makes Apple TV the new HBO — and it’s not just because it doesn’t have ads yet.
• How much would you pay to get rid of ads? This survey from Consumer Reports found that 3 in 10 people subscribed to an ad-supported tier would pay less than $5 more to go ad-free. It also has some other fun tidbits about how annoying people find ads while streaming, and when they think is the best time to see commercials.
• Fortune highlights how streaming fatigue is driving more people — especially Gen Z — to subscribe to and then cancel streaming services after they watch specific shows.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Emma RothClose

  Emma Roth

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Emma Roth

• ColumnClose

  Column

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Column

• StreamingClose

  Streaming

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Streaming

• The StepbackClose

  The Stepback

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All The Stepback

The founding story of Nest is pretty much a perfect tech myth. A legendary product maker (in this case, Tony Fadell) helps create one of the most successful products ever (the iPhone) and then rides off into the sunset to enjoy the rest of his life, only to have an experience that drags him back for one last job. For Fadell, that job was to try and reinvent the thermostat. And maybe change the way our homes work forever.

On this episode of Version History, we tell the story of the early days of Nest. The Verge’s David Pierce, Nilay Patel, and Jennifer Pattison Tuohy describe Fadell’s frustration with outdated, expensive temperature controls, and his swift entry back into the product scene.

Nest was remarkably successful, remarkably quickly, and a lot of that is due to Fadell himself. The thermostat itself, though, was a bit of a mixed bag. It got a lot of things right — and Nest’s big idea about the future was eventually proven remarkably correct — and it also missed the mark in some important ways. It shipped with one heck of a screwdriver, though.

This is the third episode of the fourth season of Version History. If you haven’t yet checked out the story of the Harmony remote or the Roomba vacuum, make sure you catch up! It’s smart home season this season, and we’re having a great time. Here’s how to get every episode, and all our other fun stuff, as soon as it drops:

If you want to know more about the early days of Nest, here are some links to get you started:

I’ve seen lots of so-called “smart” bike locks over the years, but none so far could justify the added cost. A newcomer that got its start securing ATMs for banks is trying to change that. There’s nothing wholly unique about the TMD Chain Lock, but the combination of materials, performance, and insurance-friendly ART-2 certification makes it worth considering.

TMD’s first bicycle lock combines a Bluetooth proximity sensor and motion alarm with a slender core of hardened steel chain wrapped in a soft and lightweight sleeve of high performance Dyneema and Kevlar fibers. That makes this lock tough, yet flexible enough to conveniently wrap around your seat post when cycling. TMD also designed the keyless lock to be shareable with friends and family, and capable of withstanding shock, rain, dust, and extreme temperatures. That makes it highly suited to city bikes parked outside and ridden all year long.

I like the TMD Chain Lock I’ve been testing, but spending €249 (about $283) on something that typically costs far less is a big ask.

$280

The Good

• Strong yet flexible lock won’t scratch bike
• Digital key can be shared with friends and family
• Convenient if you lock your bike several times a day
• ART-2 certified for insurance
• Loud enough alarm

The Bad

• Very expensive
• Lacks USB-C charging
• Annoying operating sounds

The chasm separating ATM security from bike security isn’t as wide as it might seem. TMD developed a centralized, keyless system for banks to authorize access to its ATMs for service, replacing the physical master keys that created a massive security risk whenever they were lost or stolen. It’s this same keyless knowhow that TMD now brings to the Bluetooth bike lock.

There are many advantages to turning phones into digital keys, especially in bike-centric societies like the Netherlands, which TMD calls home. For example, I could ride to the train station and have an authorized friend pick the bike up later. And the keyless convenience quickly adds up for those of us that use our bikes to commute to work, pick up the kids from daycare and football, run errands, and shop for groceries. In my home city of Amsterdam, it’s not unusual to grab my keys more than a dozen times a day to lock and unlock my bike.

I tested the TMD Bike Lock in the 110cm (about 3.5 feet) length with an iPhone 15 Pro running the latest version of iOS. Generally — I’d say 19 out of every 20 attempts — the TMD Chain Lock immediately recognized my approach, allowing me to seamlessly unlock it with a push on the lock’s only button. When I did have to wait, it took no more than a second for the button to flash blue, indicating the lock had detected my nearby phone in a pocket or bag. Pretty good, but that convenience comes with some tradeoffs.

The bike sat unlocked across the street since it was still within Bluetooth range of my phone at this cafe.

My bike, parked outside this window, would automatically unlock whenever I cooked.

Charging the lock off my laptop. It’ll also charge off a phone’s USB-C port.

Don’t lose this proprietary charging cable!

For example, my bike remained within range of my phone when parked outside a cafe, so anyone could have walked up and unlocked it. I had the same issue when cooking in my kitchen with the bike parked just outside the window, as I normally do. In both cases, I either had to move the bike or turn off my phone’s Bluetooth radio. And since my phone is now my bike key, if it’s stolen, my much more expensive e-bike is vulnerable as well.

TMD says its Chain Lock is “immune to conventional drilling and picking,” but all bike locks can be defeated by a determined thief. The Chain Lock has earned an ART-2 certification issued by an independent Dutch organization that tests and certifies bike locks on a scale from one to five. Most Dutch insurance companies require ART-2 at a minimum to validate policies on expensive e-bikes and cargo bikes. It’s akin to something like a Sold Secure Silver or Gold rating in the UK, a 2 Roues in France, and VdS approval in Germany. (The US and Canada do not have a single, centralized, independent testing body that dictates insurance requirements for bicycles.) Should your bike be stolen, you can produce a digital log file for the insurance company from within the TMD app as official proof that your bike was locked.

To deter theft, the lock also features an integrated motion alarm that can be armed and disarmed in the app or with a double press of the lock’s button. At 100dB it’s less shrill than I expected, but it does the job when it senses the lock being jostled. It shuts off automatically after about 10 seconds of inactivity.

TMD tells me that a future software update will let you disable these sounds.

The lock is rather noisy in day-to-day usage. Unlocking it is accompanied by a loud tone that hits 75dB when measured at a distance of one meter. I hate that it draws attention to me and my expensive e-bike. Just flash green and unlock — that’s more than enough. TMD tells me that a future software update will let you disable these sounds.

TMD claims the lock’s battery will last up to nine months on a single charge. However, my test unit has already dropped 16 percent after a week of testing, suggesting a lifespan closer to six or seven weeks. But that’s likely because I store my bikes inside my house when not in use, so the lock’s Bluetooth radio is constantly being lit up by my phone’s proximity.

To prevent water and dust ingress and leave more room for the battery and speaker, TMD says it opted for a shallow magnetic pin connector – not USB-C – to charge the Chain Lock. That’s a shame because plenty of phones have USB-C ports and offer even better IP68/69 protection against rain and debris. That means you’ll need to have the lock’s proprietary cable on you should you ignore the low-battery warnings and let it die. It can be quickly recovered, however: I saw a completely dead (and locked) Chain Lock spring back to life after just two or three minutes of charging off the USB-C port of an iPhone.

You can still unlock the TMD Chain Lock should your phone die. You just have to tap in your four digit “emergency code” on the lock’s button. Choosing a code like 1324 would require 1 push, then 3 pushes, then 2, and 4. The ring around the button changes colors to visually prompt you for each new number. The TMD app also lets you share the lock with friends and family, who can also come to your rescue with their own phones.

1/6100cm is my preferred length.

For added peace of mind, it’s always a good idea to double-lock your bike, so most urban cyclists combine a chain lock with a ring lock that’s permanently mounted to the frame and prevents the rear wheel from spinning. TMD has already announced an “anti angle grinder” U-Lock and GPS-equipped Ring Lock coming later this year. Having my phone unlock both of my bike’s locks simultaneously would really up the convenience factor and better justify TMD’s pricing.

Right now, paying €249 when equivalent analog bike locks costs a quarter of that would be foolish for most people. The ART-2 certified ABUS 8900, for example, in the same flexible 110cm length, costs around €60 (about $68). If you can afford the TMD Chain Lock and really hate keys, or need a keyless lock that can be shared digitally with others, then have at it. Otherwise, it won’t justify its exorbitant price.

• Security rating: ART-2 certified
• Operation: Keyless Bluetooth
• Deterrence: Motion activated 100dB alarm
• TMD Locks App: Manages sharing, location tracking and more (iOS and Android compatible)
• Sizes: 70cm (1.3kg), 110cm (1.7kg) and 160cm (2.1kg)
• Colorways: Navy, Beige and Black
• Power: 9 month battery life. Charge to 100% in 2 Hours
• Durability: IP57 water and dust resistant
• Warranty: 2 years
• Pricing: from €229

All photography by Thomas Ricker / The Verge

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Thomas RickerClose

  Thomas Ricker

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Thomas Ricker

• Electric BikesClose

  Electric Bikes

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Electric Bikes

• ReviewsClose

  Reviews

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Reviews

• RideablesClose

  Rideables

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Rideables

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

• TransportationClose

  Transportation

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Transportation

Teenage Engineering has already issued multiple substantial updates for its surprisingly capable $329 EP-133 KO II sampler. Its latest is one of the biggest yet. OS 2.5 adds audio over USB, selectable sample rates for lo-fi fun, sample reverse, an arpeggiator, equal-length autochopping, and it extends the maximum length of a sample from 20 seconds to 40 seconds by capturing mono, instead of stereo, audio.

Sample reverse is such a simple feature that it’s shocking it wasn’t implemented earlier. An arpeggiator doesn’t always make a ton of sense on a sampler, but the KO II sounds so incredible repitching samples (like the SK-1 successor I’ve always wanted) that here it’s a huge addition. It’s easier than ever now to crank out a synthpop bassline.

The combination of longer sampling time and equal-length autochopping also makes the KO II much better suited to chopping up and rearranging samples and loops. While the transient-based autochopper was great for isolating drum hits in a break, it was less useful for melodic content.

My favorite new feature, though, is the new selectable sample rate. You can stick with the standard 46k kHz if you want, but the new 32 kHz mode adds some nice character, while 26 kHz goes full lo-fi, turning crunchy and digital.

Those are just the headlining features. There’s also improved time stretching, new scales, per-pad time shifting, and tons of bug fixes.

The OS 2.5 update is also available for the KO II’s reggae-themed sibling, the Riddim. While the bizarre EP-1320 Medieval sampler is finally getting an OS update, it’s only getting USB audio, and nothing else. We’ve asked Teenage Engineering why the EP-1320 is regularly left out of the update fun, but have yet to hear back.

You can update your OP-133 KO II, EP-40 Riddom, or EP-1320 Medieval here.

What’s the Password? has a simple concept: To solve each of the game’s more than 100 puzzles, you have to type in the right four-digit password on a number pad. That might sound like a limited constraint. But the simplicity gives solo developer Dan DiIorio, better known as TrampolineTales, lots of room to play with clever ideas. Over the course of a few hours, the game never stopped surprising me.

Puzzle clues come in several different formats. Some are written text; the very first puzzle is a sentence that tells you what numbers to punch in. But you’ll also have to decipher four-digit codes from things like the blinking digits of a clock, nonogram puzzles, solving arithmetic, and even counting the number of teeth on a key. Clues can get a little meta, too, like ones referencing a password on the game’s cover art or information in menus.

The puzzles can be tough. I’d get a few right away, but for most, I’d have to try a few solutions to see what worked. Fortunately, there’s no penalty for wrong guesses, so you can try as many combinations as you want. Some stopped me cold for minutes at a time. If you’re stuck on a puzzle for a while, the game will give you an option to skip it, and occasionally you can get a hint.

The game is backed by a calming, jazzy soundtrack that loops a single song over and over, and it pairs well with the game’s entirely black-and-white visuals. The whole vibe made me feel like some kind of noir gumshoe. But I also found that the slow drone could help when I was stuck on a particularly tricky puzzle; my mind would occasionally wander to the meandering music, and every once in a while, that idle thinking would lead me to a puzzle’s solution.

In the end, I couldn’t solve everything. It took me about three hours to work my way through What’s the Password?, and I completed about 77 percent of the puzzles. But the game reminded me of the value of patience and a good night’s sleep; coming back in the morning with a fresh mind and coffee in hand, I usually had an easier time figuring things out. A lot of the fun of What’s the Password? was seeing what unexpected format would be thrown at me next, and that’s what gives the simple premise such depth.

What’s the Password? is out now on PC via Steam, iOS, and Android.

Maraget Atwood, the storied author of The Handmaid’s Tale and The Blind Assassin, was interviewed as part of the Babell Literary and Cultural Festival in Porto, Portugal. As it usually does at these things, the issue of AI came up, and Atwood didn’t mince words.

According to Deadline’s recap, Atwood said she’d used an AI chatbot exactly once, Anthropic’s Claude, and came away unimpressed. She was looking for information about the British detective series Father Brown and, well:

”Claude gave me the wrong answer, or it lied. Of course, it didn’t know it was lying because it’s not a human being; it’s a large language model… It had skimmed and sampled a lot of television reviews, but they never give away the ending in online criticism, so it was misled by the things it had read about the show.”

She didn’t have particularly kind words for the people who rely on AI either, calling them “opportunists” looking for the easy way out. But of course, as she pointed out, all LLMs are only as good as the data they’re fed, and putting your faith in a machine trained on scraped, previously published, and possibly out-of-date information isn’t the best idea.

“Human beings are not robots, but they are opportunists, so if there’s an easy way to cheat and it’s hard to detect, people will do it… But the thing about AI is that it’s garbage in, garbage out. Even people who use it for business reasons have to check it because it makes mistakes.”

Apple is looking to alleviate some of the pressure on its supply chain by seeking an exception from the Trump administration to buy RAM chips from CXMT, a company blacklisted by the Pentagon over ties to the People’s Liberation Army, according to the Financial Times. The skyrocketing prices of RAM and storage have driven Apple to raise prices on almost all of its products this week, so it makes sense that it would seek alternative sources.

Legally, Apple isn’t barred from buying chips from CXMT, but doing business with a company tied to the Chinese military would carry serious reputational risks. It’s possible that CXMT could still find itself the target of export controls for undermining US security. The company was on a list of proposed additions to the so-called “Entity List” by the Commerce Department, but held off because the White House was in the middle of trade negotiations with China.

It’s unclear if the administration would give its blessing to Apple. Tim Cook has spent significant time trying to build bridges with the Trump administration, presenting the president with gaudy statues and attending a screening of the Melania movie, directed by accused rapist Brett Ratner. But if the White House granted Apple permission, such a decision would likely face significant blowback. John Moolenaar, Republican chair of the House China committee, told the Financial Times that:

“Apple choosing to partner with a Chinese military company would be a grave mistake… Helping the [Chinese Communist Party] succeed in its plans to dominate critical supply chains will make our country’s tech industry and economy more dependent on China at a time when we must build secure tech supply chains with our allies,”

Four years ago, overlooking a canal in Amsterdam, the smart home industry collectively launched Matter, the one interoperability standard to rule them all. Heralded as the solution to the industry’s struggles, Matter was built on open standards and existing technologies and is the result of years of collaboration between traditional rivals, including Apple, Google, Amazon, and Samsung.

Matter promised an end to walled gardens and ecosystem lock-in. It promised to make a smart home device, like a lock, lightbulb, or sensor, easy to buy and set up. It promised you could choose any brand, use any platform, no expertise required — it would just work.

“Matter long-term won’t be successful until everybody can use it at parity. That’s the goal. And all the companies know that.”

— Tobin Richardson, CSA

It didn’t. Today, Matter is still not there. Adding devices to your smart home is still cumbersome and finicky; sharing them across ecosystems can be unreliable, and some features users expect aren’t part of Matter and still require the manufacturer’s app. But, based on what I saw at the Unify conference in Austin, Texas, last week, Matter is much closer to its promise of a simple, reliable, interoperable smart home.

Following a rocky launch and years of broken promises, the Connectivity Standards Alliance (CSA) designated 2025 as the year to fix Matter — and I have seen significant improvements. This year at Unify, the CSA’s first public-facing event since the Amsterdam launch of Matter, version 1.6 of the spec was announced.

The most significant new feature, Joint Fabric, will let you create a single smart home network that can be controlled by any Matter platform — which is the Matter we thought we were getting all along. The question is, will the platforms adopt it? And that’s a question that has been at the heart of Matter’s struggles to date. It turns out that corralling an industry is not easy.

A panel at Unify hosted by Jennifer Pattison Tuohy featuring representatives from Google, Samsung, Homey/LG, and CableLabs. Here, Daniel Moneta of Samsung answers a question about Samsung appliances not being part of Matter, saying they are focused on offering a great first-party experience through SmartThings. Image: CSA

The mood at the three-day conference was optimistic but realistic. There was widespread recognition that Matter hasn’t done what it set out to do, but that the work is being done to get it there.

I asked Tobin Richardson, the CEO of the CSA, whether he believed Matter had achieved that initial promise that you could buy a device with a Matter logo and it would be easy to set up and use on any platform. “We’re on the path,” he hedged. “In some ways, yes. But in some places, universally and globally, we still have a few more milestones to get there.”

The biggest concern among many of the attendees I spoke to — which included representatives from silicon vendors, manufacturers, and ecosystem companies such as Google, Amazon, and Samsung — is the widening gap between the spec and its implementation by ecosystems. This disconnect has led to more consumer confusion, rather than less.

Developed by the Connectivity Standards Alliance (CSA), whose members include Apple, Amazon, Google, and Samsung (and others), Matter is an open-source, IP-based connectivity software layer for smart home devices. It works over Wi-Fi, ethernet, and the low-power mesh networking protocol Thread. Matter currently supports most of the main device types in the home. These include security cameras, lighting, thermostats, locks, robot vacuums, refrigerators, dishwashers, dryers, ovens, smoke alarms, air quality monitors, EV chargers, and more.

A smart home gadget with the Matter logo can be set up and used with any Matter-compatible ecosystem that supports that device type via a Matter controller and controlled from multiple smart home platforms at the same time with a feature called multi-admin.

Amazon Alexa, Google Home, Samsung SmartThings, and Apple Home are some major smart home platforms supporting Matter, along with hundreds of device manufacturers.

Matter 1.6 launched this month, but Apple, Google, and Amazon are barely at 1.3, which came out two years ago. SmartThings has committed to adopting new specs within six months of release, but Samsung still hasn’t added Matter support to its appliances.

Without platform parity, Matter’s promise simply can’t be fulfilled. George Yianni, head of technology at smart lighting company Philips Hue, summed up the general feeling: “It would be great if, within a year of new features being published in the spec, it was broadly supported by the ecosystems.”

Matter may have mostly solved the initial challenge of getting everything to work together, but it’s still working on getting everyone to work together. “Matter long-term won’t be successful until everybody can use it at parity. That’s the goal. And all the companies know that,” said Richardson. While he concedes there’s a natural tension with competitors like Google, Apple, and Amazon working together, he says they know that the result is a bigger market. “These companies want to make money, but that’s a good thing for Matter,” said Richardson.

Technologist Kevin Ashton, who coined the term Internet of Things and developed WeMo at Belkin, put it best, saying on stage that companies need to “be brave enough to compete on features and benefits rather than trying to build a prison for your customer.”

Ultimately, it is the CSA’s role to make competitors work together. “We create a safe space for them to collaborate,” Jon Harros, head of testing and certification at the CSA, told me. “Collaboration is still painful, but it is collaboration, and it is the way to move the needle forward.” He points to the Interoperability Lab they established and that the platforms are now accepting Matter certification for their Works With badges as evidence.

Six years after Apple, Amazon, and Google issued the first-ever press release with all three company names in the headline, Richardson said the collaboration remains strong. “Yes, there are cracks all the time,” he conceded. “But not in the foundation. In some ways, it is less confrontational now, more congenial over how we’re going to solve this. Early on, there was a real wresting of control.”

Over three days at Unify, I heard from representatives of all the major platforms and manufacturers across the industry, as well as the heads of the groups that manage the major wireless protocols: Thread, Wi-Fi, Zigbee, and Bluetooth. The universal message was that, by building on open standards and collaborating to find solutions, the smart home industry is moving toward a stronger future. And despite the slower-than-expected adoption from Google, Apple, and Amazon, I was repeatedly assured that the sheer resources these companies continue to pour into Matter is evidence that they’re in it for the long term.

Dan Wroclawski of Consumer Reports moderates a panel with Tobin Richardson of the CSA; Vividh Siddha, president of Thread Group and Director of software engineering at Apple; Neville Meijers, CEO of Bluetooth SIG; and Kevin Robinson, CEO of Wi-Fi Alliance. Photo by Jennifer Pattison Tuohy / The Verge

One concrete example I was given: Earlier this year, Ikea, an early member of Matter, rolled out an entire line of new Matter-over-Thread smart home products. It was the first large-scale, real-world deployment of Matter-over-Thread to a broader consumer audience, and it quickly ran into serious problems. Customers either couldn’t connect to the devices or they were dropping offline, and while not all buyers were affected, it was a big enough problem that the industry took notice.

“Collaboration is still painful, but it is collaboration, and it is the way to move the needle forward.”

— Jon Harros, CSA

At the conference, engineers from several of the major ecosystems told me that they all gathered with Ikea for a weeklong troubleshooting session, working from 7AM to 10:30PM each day — even going to the local Best Buy to buy five of every Thread border router and tracking down ISP router manufacturers on LinkedIn to alert them to bugs they’d found that were impacting the devices’ ability to connect to users’ networks. They tell me the efforts unearthed numerous potential problems and led Ikea to roll out several fixes.

This type of collaboration feels unprecedented for any industry and is a strong indicator of the investment in the standard. It’s clear that the engineers and product teams at these companies are dedicated to Matter’s success. But I still have concerns about the continued investment of their management.

Apple’s slow-and-steady approach seems the most solid. Google has certainly shown renewed interest in the smart home, propelled by Gemini. But I see Amazon as the biggest concern for Matter. The company’s stance seems more on the side of “we support everything” than “we believe Matter is the best solution.” It has also begun moving further toward its own proprietary Sidewalk protocol since Ring founder Jamie Siminoff’s return to the company.

A lot hinges on Matter’s perceived success in the market. Several companies appear to be circling the standard, not fully committing until it becomes the default, ready to retreat if they’re not forced by market pressures to adopt it. But by the numbers, Matter is looking strong.

As a new standard, it has achieved a lot in a relatively short time. It now covers most smart home device types, including, finally, security cameras, and has certified over 1,200 unique certified products. The CSA has 940 member companies, with 300 actively working on the standard and more arriving. ADT just joined the board of directors, bringing another large, legacy company deeper into the fold.

Thread has seen a significant increase in certified products in the last two years. Image: Thread Group

Similarly, Thread, one of Matter’s two wireless connectivity protocols, has also seen a huge uptick in interest, including a 27 percent increase in membership over the last two years and over 1,000 certified devices, with 151 arriving just this year. There are now 71 Thread-certified components, meaning silicon providers are building solutions for manufacturers, creating a strong foundation for new companies looking to use Matter in their products.

My biggest takeaway from Unify is that despite the problems the standard has faced, it’s not in danger of failing. Despite some outliers and cautious approaches, the industry coalesced around Matter four years ago, and that coalition remains strong.

That doesn’t mean Matter where it needs to be, however. Along with faster platform adoption and a need for more devices in the newly supported device types (cameras in particular), it still needs to deliver on its initial promise of simplicity. As I said at the panel I hosted at the conference, for me, success will be when I can feel confident that when I hand my 80-year-old father a Matter lightbulb, he can get it set up and working with no help from me.

Buying devices that support Matter is still confusing for customers. Photo by Jennifer Pattison Tuohy / The Verge

For manufacturers, the roadblocks to that simplicity are still too many. Yianni at Philips Hue told me that while he believes Matter has landed as the “definitive smart home interoperability mechanism,” it’s still confusing and complicated for his users.

In particular, this comes back to the slow adoption of the spec by ecosystems. The initial experience Matter delivered for Hue users was worse than what they could offer without it. “Matter launched too early, but a lot of those teething problems have been solved, with a lot of hard work from many companies,” he said.

“Most consumers just assume things will work.”

— George Yianni, Philips Hue

The challenge now is bringing those solutions to people’s homes. Yianni said he’s most excited for the just-announced Joint Fabric. “Now we actually have the possibility to have a single Thread network and a Matter fabric for the home, like we have a Wi-Fi network. That’s how everyone assumed it would work.” However, as everyone has learned, just because it’s in the spec doesn’t mean it will be implemented.

Additionally, Yianni feels that communication to consumers about what Matter is has been disappointing. Too much has been placed on device manufacturers, and the ecosystems still push their users to look for the Works With badges. “The badges are a crutch,” he said, arguing that most regular customers aren’t looking for them. “Most consumers just assume things will work.”

He and others I spoke to believe it’s worth the investment by all the ecosystems and the CSA to make this much simpler for people looking to buy smart devices, and to reinforce that the Matter logo is all you need to look for if you want to know whether a device will work with your smart home. “Because today it’s just confusing,” he said.

It should be so ubiquitous that the idea of a smart home device without Matter is as absurd as a wireless router without Wi-Fi.

John Bartucci, of Fortune Brands, the owner of Yale and Moen, pointed out during a panel that retailers also rely on the badges. “Retailers want products that won’t get returned; correct badging is how consumers trust it’ll work at home.” Austin Stewart of Schlage Locks agrees that talking with customers and retailers is still “a Works With story.” For their customers, the Apple Home badge means this lock works with the Apple Home app, “and that’s the experience you get,” he said. As Matter grows and becomes more ubiquitous, that story may change. He compared it to Wi-Fi, a term so common that his grandmother knows it, but today, Matter is not there yet. “It’s the experience that we’re delivering, not necessarily the technology,” he said.

Ultimately, the Matter logo needs to become synonymous with smart home connectivity, in the same way Bluetooth is synonymous with wireless audio and Wi-Fi with wireless internet. “Today, if you go buy some new headphones, you don’t look for the Bluetooth logo; it should be the same for Matter with smart home,” said Yianni. It should be so ubiquitous that the idea of a smart home device without Matter is as absurd as a wireless router without Wi-Fi.

That brings us back to the core concept of Matter: trust. The standard promised that I could just hand my father a smart lightbulb and trust that I would never get the call asking why he’s sitting in the dark. We’re not there yet. But, given the momentum at Unify and the evidence I saw of continued investment by many manufacturers and ecosystems, Matter feels closer to fulfilling its promise than it’s ever been.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Jennifer Pattison TuohyClose

  Jennifer Pattison Tuohy

  Senior Reviewer, Smart Home

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Jennifer Pattison Tuohy

• AnalysisClose

  Analysis

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Analysis

• MatterClose

  Matter

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Matter

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

• Smart HomeClose

  Smart Home

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Smart Home

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

Kai Wright is the co-host of Stateside with Kai and Carter over at the Guardian. But Wright has been bringing his unique insights to listeners for years. He’s also hosted Notes From America, The United States of Anxiety, and Indivisible. He’s a Peabody Award-winning journalist who has profiled powerful men, explored what it means to be American, and chronicled the AIDS epidemic.

When he’s not diving deep on sex, race, and politics, he’s gardening, listening to John Coltrane, and steadfastly refusing to buy a new phone. Wright was kind enough to give us a glimpse into his daily routine, how he unwinds, and offer some simple but powerful advice.

What is your most indispensable tool?

Which is the most underappreciated?

Wheelbarrow. Seriously, though. Haul around enough dirt and rocks, and you will come to appreciate the brilliant utility of this basic design.

What is one thing you wish you could change about your phone?

I wish I could have bought one smartphone and never ever had to buy another one again, forever. Why must it constantly go obsolete? (Of course, I haven’t actually bought a new phone in I dunno how many years; I just take my partner’s hand-me-downs. So I’m complaining on cranky principle.)

How many tabs do you have open right now?

In the window I’m using right this moment, I have 10 tabs open. Yes, this is quite a reasonable number, thank you. I do have 11 more windows open, however, and cannot bring myself to account for what’s in them.

Which was the most disappointing gadget you’ve ever owned?

I won’t name brands, but I promise you can make a coffee machine too complicated. I’m trying to get caffeinated so I can become a functioning human. I don’t want an engineering challenge.

What is one thing you wish you had created?

I wish I had thought of The Memory Palace – which is still, to me, podcasting at its best. Short, sweet, deeply human histories of our world. One person, telling you a well-written story.

What creation are you most proud of?

Can I say my whole, entire garden? It’s an exercise in constantly learning and applying the knowledge and trying again, and I hear from veteran gardeners that you never actually figure it out. Perfect.

What’s the best piece of advice you’ve ever received?

Live an honest life. Meaning, try your best to tell yourself the truth, and make choices accordingly.

What do you do when you’re feeling stuck?

Listen to John Coltrane records.

What’s the last piece of physical media you bought?

What would the tagline for your biopic be?

“What had happened was…”

What’s the last GIF or meme you used?

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Terrence O’BrienClose

  Terrence O’Brien

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Terrence O’Brien

• InterviewClose

  Interview

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Interview

• PolicyClose

  Policy

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Policy

• PoliticsClose

  Politics

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Politics

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

Nostalgia remains a powerful force. So much so that, in exploring the echoes of a late-’90s childhood spent skimming the water of Corneria and sneering “cocky little freaks!” in time with a monkey encased in a Gundam suit, I’m simultaneously describing playing Star Fox 64 (Lylat Wars if you’re nasty) in 1997 and streaming it through Nintendo Switch Online today.

The franchise has been revived through a splashy remake on the Switch 2, but it’s also a series that has not seen an all-new entry since Star Fox Zero on the Wii U. Yet Nintendo’s neglect of the series has been gently offset by indie creators. Ex-Zodiac and Whisker Squadron: Survivor recently offered echoes of Star Fox, and now two upcoming games, Huskrafts’ Rogue Eclipse and Wild Blue Skies from Chuhai Labs — founded by former Star Fox programmer Giles Goddard — offer hope that this desolate genre might be resurrected.

Though perhaps it’s unfair to describe Star Fox’s absence in terms of abandonment. “It’s just that times have moved on,” Goddard tells The Verge. “Each iteration were great games for their time,” but now, he suggests, their enduring popularity may be because you don’t see them much. “I think people are starved of originality more than any particular style or genre of game.”

For indie developers who hope to forestall the obscurity of games reminiscent of Star Fox, that translates to dealing with risk-averse publishers. “When I was pitching Rogue Eclipse, the response I generally received from most labels was that the genre is dead,” says Huskrafts’ Husban “Mcdoogleh” Siddiqi. It echoes the experiences of others, including Flippfly creative director Aaron San Filippo, who says many publishers “told us they just couldn’t see a big enough market to justify our budget for Whisker Squadron: Survivor.”

Unable to persuade publishers to take the leap, developers like Flippfly turn to crowdfunding as a means to resuscitate Star Fox. This follows crowdfunding success stories like Hollow Knight, Pillars of Eternity, Shovel Knight, and Undertale, which similarly revived underserved genres.

Siddiqi isn’t convinced by arguments that the audience for Star Fox-alikes, or more broadly arcade flight shooters, is limited. He points to the success of 2019’s Ace Combat 7: Skies Unknown as proof of that (Bandai Namco announced its sequel at The Game Awards 2025). As is often the case, however, an existing audience is not a guarantee that if you build it, they will come. In February 2025, Flippfly let its development team go for lack of funding, shelving a more ambitious follow-up to Whisker Squadron: Survivor in the process.

Challenging as funding is, bringing Star Fox back from the brink requires capturing what Siddiqi describes as the “push-forward, fast, frenetic, and kinetic approach to combat” that marks its genre in ways that makes sense for a modern audience. That’s a test for indie developers, but also for Nintendo with its Star Fox 64 remake on the Switch 2.

“Not everyone wants a low-resolution game,” Ben Hickling, developer of Ex-Zodiac, says, adding that others might not enjoy a 240fps reimagining of childhood favorites either. In revisiting underserved genres, there is a balance to be struck between capturing the realities of past games and the rose-tinted forgeries, replete with erroneous improvements, that we conjure in our memories.

“I think people are quite happy with me making Ex-Zodiac more responsive and snappier, as I think that’s how people remembered the original being, despite that not being the case,” Hickling continues. “To be honest, the whole game is kinda my version of how I imagine Star Fox in my head.”

That’s one approach. Another is to pull from more than one source. Rogue Eclipse draws influence from Star Fox, but also Armored Core, Mobile Suit Gundam, Returnal, and Battlestar Galactica (to name only a few Siddiqi lists).

This is echoed further in Chuhai Labs’ approach. Wild Blue Skies, according to director Francis Pétrin, is less about being a Star Fox clone than capturing a broader memory of Saturday mornings — games, cartoons, and all — in the 1990s. That Wild Blue Skies happens to bear a striking resemblance to Star Fox? “I guess we only really know how to make one kind of on-rails shooter!” Goddard laughs.

Yet evoking the past is only useful insofar as it supports engaging gameplay. “My blind affection for the past could mean we’re inadvertently bringing in baggage or inheriting systems which could be unnecessary for the experience,” he says. “I viewed this as an opportunity to see if we could solve some of the issues that individuals may have come up against in the past with flight games in general.”

One way developers are making that happen is by building on top of the basics that games like Star Fox did so well: making flight feel natural. “What immediately comes to mind is just how effortless and responsive the flight controls were,” Siddiqi says of 2001’s Star Wars: Rogue Leader. (Mark Haigh-Hutchinson famously spent a year perfecting its prequel’s controls.) “I can still recall, to this day, playing the Death Star Run mission for the first time. I imagine it’s one of the reasons why Rogue Eclipse has manifested.”

More pointed is the desire of developers to capture the broad shades of their own experiences and share those in a visceral, and interactive, way. Readers of a certain age will hear their own childhoods echoed back at them when developers talk about their love for the genre, and of days when games were more personally scarce and therefore more impactful.

“When I was a kid living in rural Wisconsin, we didn’t have any gaming consoles, but you used to be able to rent them from the video store,” says San Filippo. “So we’d get them as a birthday treat, and we’d play Star Fox.”

That, Goddard suggests, may well be at the heart of this steady uptick in Star Fox-style games. Much as the 2010s produced a wave of 8- and 16-bit-inspired games, he says, “a generation of people who grew up in the 3D era when these games were more popular now are becoming game developers themselves.”

As much as craving Star Fox-inspired games may feel like staring back in time, however, it’s not all about the past. The metamorphosis of collapse has made the gaming industry more risk-averse, certainly, but it’s also pushed it into the greater conservatism of trend-chasing and homogenization. More and more, we’re having to look to the past, and periods of genuine innovation, to find shreds of creativity and originality. This is most obvious in underserved genres from a gaming world we can no longer truly grasp. In returning to Star Fox, we are in fact mining experiences that feel genuinely new and unique for their absence in a gaming industry that often feels anything but.

“Ultimately,” Siddiqi concludes, “arcade flight experiences still offer a more playful, exaggerated fantasy gameplay fulfillment that players simply may not be able to get elsewhere.”

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Geoffrey BuntingClose

  Geoffrey Bunting

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Geoffrey Bunting

• EntertainmentClose

  Entertainment

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Entertainment

• GamingClose

  Gaming

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Gaming

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

Tim Cook recently said price increases were “unavoidable” and described the company’s pricing as “unsustainable.” The 16-inch MacBook Pro saw its price go up by $300. The 11-inch iPad Air went from $599 to $749. Even the HomePod Mini got a $30 bump to $129. Cook squarely placed the blame at the feet of the AI industry, which is not surprising. RAMageddon has already come for your desktop PCs and gaming consoles. The Xbox has seen its price climb nearly 25 percent depending on the model, and Nothing even canceled an entire phone launch. Apple is just the most recent to jack up prices and point the finger at AI.

The price hikes are “basic economics,” says Tim Derdenger, associate professor of marketing and strategy at Carnegie Mellon University’s Tepper School of Business. As the tech industry has raced to win the AI war, “the price of RAM has skyrocketed because the memory manufacturers have reallocated their production lines to produce new HBM memory for AI data centers and away from consumer DDR5.” And when the cost of components goes up, companies tend to pass those costs on to consumers.

But this isn’t some fluke, or temporary supply chain problem. Companies are choosing data center clients over ordinary buyers because “the same chip earns far more inside an AI server than inside a consumer device,” according to Srikanth Jagabathula, professor of technology, operations, and statistics at the NYU Stern School of Business. Regardless of whether people are clamoring for more AI, and more AI data centers, or not.

Companies like OpenAI, Google, and Microsoft have thrown around unprecedented amounts of money, outbidding companies like Apple for RAM and storage, creating what even Sam Altman has admitted is a bubble. This imbalance has led to record earnings for companies like Micron, which manufactures memory chips. “This shortage is not temporary and might extend into the next few years … And because the increase is lasting rather than temporary, simply absorbing the cost is not a sustainable strategy,” Jagabathula says.

But Apple has posted record earnings for at least four quarters in a row, and its margins on hardware sales are much higher than the industry standard. Its markups are estimated to be between 30 and 40 percent, depending on the product. TechInsights and The Wall Street Journal estimate that it’s even higher on the iPhone 17 Pro, perhaps as much as 47 percent. According to TheStreet, margins on smartphones are typically between 15 and 25 percent. Data on laptop margins is harder to come by, but estimates put it between 10 percent and 25 percent for most of the industry.

Apple is actually among the last of the major tech companies to raise its prices. But why are customers being asked to foot the bill when Apple seems well-positioned to absorb these costs?

Ari Lightman, professor of digital media and marketing at Carnegie Mellon University’s Heinz College, described it as “spot on” to say it’s hard to square Apple’s public financial statements and Tim Cook’s description of its pricing as unsustainable. He said raising prices was “without a doubt” about appeasing shareholders who demand constant growth.

Lightman points to Apple’s lagging behind in the AI race, the uncertainty around installing a new CEO in John Ternus, and the lack of a hit new product category as putting pressure on the company.

“There’s a lot of things that investors can really beat them up on,” he said, and “if they’re going to be selling the stock and promoting the stock to large institutional investors … in terms of being one of the most valuable companies, then they have to tell a really good story.” And that story is one of huge margins and profits even in the face of rising costs and AI-driven supply constraints.

The AI boom is touching almost every facet of our lives, but this week, it came particularly hard for our wallets with the announcement of another round of price hikes for the Xbox, and even the Arduino got caught up in the memory crunch. I spent hours talking to marketing and business experts, exchanging emails and phone calls, and no one could give me a satisfying answer to why the price of popping up more data centers should be consumers’ costs to bear.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Terrence O’BrienClose

  Terrence O’Brien

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Terrence O’Brien

• AIClose

  AI

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All AI

• AppleClose

  Apple

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Apple

• BusinessClose

  Business

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Business

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

Hi, friends! Welcome to Installer No. 134, your guide to the best and Verge-iest stuff in the world. (If you’re new here, welcome, hope you’re okay in all this heat, and also you can read all the old editions at the Installer homepage.)

This week, I’ve been reading about Polymarket lies and Jalen Brunson and the Arts and Crafts Movement, watching Not Suitable for Work and way too much of the World Cup, trying to catch up on The Bear before the final season starts, playing a bunch of Hank Green’s 4×3 game, trying my absolute hardest not to preorder this amazing Stanley Kubrick collection even though I’m obviously going to, buying a bunch of new AirTags and charging cables during Prime Day, poring over oral histories of data centers and Steven Spielberg, and thinking rude thoughts about whoever decided to kill Notion Mail.

I also have for you a smart speaker worth a look, a great new remake of a great old game, a couple of excellent video essays, and a Netflix hit to binge on your vacation.

Oh, and two more things. One, we’re off next week for the holiday weekend. And two, I have a question: What’s your reading setup? I mean reading in the broadest sense: books, long articles, newsletters, weird zines your kids make, anything and everything. I want to know the apps you swear by, the decade-old Kindle you won’t let go of, your go-to used bookstore, your favorite brand of highlighter, anything and everything. I have some thoughts on the subject, too. We’ll talk.

(As always, the best part of Installer is your ideas and tips. What are you reading / watching / listening to / playing / gluing together this week? Tell me everything: installer@theverge.com. And if you know someone else who might enjoy Installer, forward it to them and tell them to subscribe here.)

• The Google Home Speaker. Nine months later, it’s finally on sale! I mostly really like it so far — for its size and $99 price it’s a very serviceable speaker, and since all I really need is “put this on my calendar” and “play this song on Spotify,” Gemini is a winner. Still a little worried about Google’s ongoing interest in the smart home, though…
• Star Fox. I was very much raised by the Nintendo 64, but I confess I never really got into Star Fox. This is good news for Adult David, because now I get to experience this game — which is by all accounts a pretty faithful and solid remake of the original — for the very first time. I have a lot of flying to do.
• The Atonemo NTS Radio Player. I don’t know why I’m such a sucker for adorable little music players, but there’s just nothing I can do about it. This one is beautiful, and I love that it both supports streaming services and offers a bunch of just-turn-it-on radio options. The NTS infinite mixtapes always seem to hit.
• “The Revolt Against Technology.” Really great video from Jared Henderson about what he calls “the new Luddites,” or more broadly the ongoing backlash against everything tech has become in our lives. I don’t agree with everything in here, but I nodded a lot at the stuff about the future we were promised that simply never arrived.
• Claude Tag. Work chat is so obviously a useful place for AI as a way to find information and accomplish simple stuff. There are going to be a lot of weird social questions to figure out here — can you @claude a request from your boss, right in-line in Slack? — but I suspect I’ll end up using this a lot.
• Meta Glasses. I’m so torn on these. Love that they’re cheaper, love that they’re lower-profile, even love a couple of the styles. But… Meta. I wish I could get a pair of these that were just a camera, a microphone, speakers, and precisely nothing else.
• Deltarune Chapter 5. A bunch of new content for an absolutely beloved game just dropped, and both the Steam charts and my social feeds suggest that people are into it. If you’ve been in this world for a while, it apparently ties a lot of things together; if you’re new, like me, it feels like the perfect time to dive in.
• Avatar: The Last Airbender season two. An Installerverse favorite is back! Heard a bunch of good things about this season, too; the general consensus seems to be that nothing could or will ever top the original, but that’s okay and this one’s a great time. I’m always in the market for shows that are just a great time.
• “Why Everyone Wants A ‘Live-Service’ Game.” Jason Schreier’s (still new-ish) YouTube channel is full of good reporting and good stories, but this one in particular did a really nice job of answering a really complicated question: Why does everyone keep making these huge, expensive, oversized games that keep failing miserably? I think I get it now.

Three or four times a year, I decide to try and make my iPad my main computer. This is almost always Christopher Lawley’s fault. He is always making these extremely cool, glamorous videos about how to use your iPad! Tablet life seems so peaceful and sane when Chris does it.

I’ve been a fan of Chris’ work forever, and got excited all over again when I saw that he’s been working on a new tasks / planner app called Parchment. (Not to brag, but I am in the TestFlight.) We’ll dig into the app another time, but I figured this was a good moment to ask Chris to share his iPad setup as it stands right now. Maybe I’ll go full iPad again.

Here it is, plus some info on the apps he uses and why:

The device: 13-inch, 2TB M5 iPad Pro. I like working from the iPad because it is a more focused environment. I have ADHD and I can easily distract myself. A lot of people complain about the iPad because it is constrained compared to the Mac; I find that to be its biggest benefit. When I need to focus on some script writing, podcast prep, photo or video editing, or even administrative tasks the iPad is the device I grab.

The wallpaper: The wallpaper I am currently using is a photo I took at Bass Lake. About 12 hours from this point, a massive snowstorm rolled in and trapped us for an extra day — it was a fantastic trip.

The apps: Files, Safari, Mail, Obsidian, Drafts, Parchment, Reminders, Calendar, Stocks, Notability, Codex, Siri, Reeder, Quick Reads, Final Cut Pro, Lightroom, Photoshop, Messages, Indigo, Apple Music, Overcast, PCalc, Shortcuts.

Oh boy, do I love apps. That was the whole reason I started my YouTube channel almost 10 years ago.

I have to start with my beloved Obsidian. At its core Obsidian is a lightweight note-taking/writing app. If you will allow me a self-indulgent plug for a second: If Obsidian seems daunting to you I made a starter’s guide for it. Obsidian also has a web clipper that I use for research, saving articles, and even YouTube videos. It’s really good at pulling a lot of metadata.

During Apple’s OS beta season I always switch back to Reminders and Calendar as my core productivity apps. I do this so I can properly cover them in detail in the fall. I have some issues with Reminders and how long it takes to input simple stuff, but it’s very powerful for a free task manager. I manage my whole business out of a task manager, and it’s just this free tool that comes with our devices.

I spent four years working a full-time job while trying to build my YouTube channel. I never automated my creative work, but even with a YouTube channel there is a ton of administrative work that goes behind the scenes. For years I used Shortcuts (technically it was Workflow—that was what the app was called before Apple bought it). Lately I have been using agentic AI, specifically Claude Cowork and Codex. I’ve been using this so much that I built a shortcut that jumps right into Codex. I even had Codex extract its own app icon so I could just place it in the dock.

A few creative apps I use on my iPad are Final Cut Pro, Lightroom, and Photoshop. I did try to go all in on Apple’s Creator Studio, but the lack of lens profiles in Pixelmator Pro specifically caused me to go back to Lightroom. For me, I can get my work done in all of these apps. They do have some features missing that their desktop counterparts have, but I make it work. It all goes back to the iPad being the platform I like to work on.

My pal Matt Birchler has been on a bit of a tear lately making a bunch of apps. One of his more recent ones that I have fallen in love with is Quick Reads. This is a read-it-later service, with support for listening to articles with really good, natural-sounding voices.

I have been desperate to replace Twitter ever since its downfall. I have ended up in two places, Mastodon and Bluesky. Indigo combines both feeds into a single timeline. You can even post to both platforms at once. I really like it as a way to keep up with friends and people I follow.

An app that isn’t on my home screen but I use a lot is the Nintendo Music app. I love this app for when I need to get work done. I am somebody that can’t focus when music with lyrics is playing, but a completely quiet room drives me up the wall. Nintendo has done a great job curating some of the best music in video game history into one place. There is even a killer feature for some supported tracks to extend the song up to an hour. This works with a lot of overworld music; lately, I have been playing the Breath of the Wild and Tears of the Kingdom tracks. The Animal Crossing tracks are also great to work to. There is even a feature to sync the Animal Crossing tracks with real-world time.

I also asked Chris to share a few things he’s into right now. Here’s what he sent back:

• I just finished 007: First Light. What a fantastic game. As somebody who has seen every 007 movie a few times (yes, even Die Another Day), I loved this version of Bond. What a great game.
• I’ve been really into 3D printing lately. I got a Bambu Lab A1 for Christmas. I had been printing Gridfinity almost nonstop, organizing my office/studio. I finally finished that. Now it’s time to do the rest of the house.
• Lastly, my girlfriend and I adopted a new puppy. Her name is Riley. She is an almost 4-month-old golden retriever. She has her own Instagram account that my girlfriend manages.

Here’s what the Installer community is into this week. I want to know what you’re into right now as well! Email installer@theverge.com or message me on Signal — @davidpierce.11 — with your recommendations for anything and everything, and we’ll feature some of our favorites here every week. For even more great recommendations, check out the replies to this post on Threads and this post on Bluesky.

“I have been looking into the Polastax project, which basically uses a cheap film holder to let you shoot Instax film on old Polaroid pack film cameras. It’s super kludgey: You can only take one shot at a time and then you have to put it into an Instax wide body to ‘develop’ the shot. But most other ways to do this cost hundreds of dollars or require an engineering degree, this is slightly more attainable!” — Richard

“Watching Star City (spin-off of For All Mankind) actually made me rewatch FAM season 1, and there are nice call-backs. These TV shows are top notch for me.” — Fabrice

“I got a pair of XReal One Pros and I’ve been using them to watch movies and play games on the Steam Deck. Really cool tech. The way it can 3D convert any video on the fly so seamlessly is amazing. It has a very crisp image as well.” — Yoinks

“Over-the-air HD digital broadcast is still a thing and, coupled with a cheap DVR like HDHomeRun makes catching the World Cup when time permits super easy (and free, too!).” — David

“Currently I’m following up on some Installer advice from a few weeks ago and listening to soundtracks while I’m working. It’s difficult to focus, though, while the score for Into the Spider-Verse is pulsing through my speakers.” — Ben

“Just started Shoe Dog: A Memoir by the Creator of Nike, by Phil Knight. Excited to hear the story of one of the world’s biggest brands, especially right now as it struggles to adapt to the modern era of running shoes.” — Ryan

“RE: PopSocket Low-Pro Grip. OhSnap did it first, and it’s great!” — Deb

“Lately I’ve gotten into fountain pens and inks. I love, love, love the online shop by Italian company Stilo & Stile. Inks are swatched and so are most of the fountain pens, plus you can whittle it down to every nook and cranny! JetPens has a terrific online store as well. Probably two of the best online stores I’ve run into.” — Elif

The game On-Together is kind of having a moment right now. Though I’m not even sure I’d call it a game. It’s just an online space: You join a server, walk around for a while, find a spot to sit, and sit. Sit a while, get tickets, buy new stuff for your avatar. That’s it! But there’s something about the idea of this sort of digital co-working that really seems to work for people — it’s not at all a new concept, but I think there’s something really cool about the way On-Together actually makes these spaces feel communal.

Right next to this game’s popularity is the concept of “admin nights,” which is basically when you get together with a bunch of strangers to accomplish all your boring tasks together. I love this idea! We should do an Installerverse admin night one of these days. I have… a lot of boring tasks to accomplish. And in general, I think there’s something powerful about finding even these very low-touch ways to be together. We should all do that more.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• David PierceClose

  David Pierce

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by David Pierce

• GadgetsClose

  Gadgets

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Gadgets

• InstallerClose

  Installer

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Installer

• StreamingClose

  Streaming

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Streaming

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

“Instead of starting from scratch each time, a skill guides Copilot through the steps, applying the right structure and formatting, and helping produce an output that is easier to review, reuse, and trust,” Brian Jones, vice president for Excel at Microsoft, said in a blog post.

Users can access a library of pre-built finance skills or create their own custom skills and save them as a SKILL.md in OneDrive, where the Copilot assistant can access them. Microsoft’s partners are also building their own skills, including finance software vendors such as LSEG, Ramp and Velixo — these are “coming soon,” Microsoft said. Custom skills are available today via the Insider channel and generally available next month.

A new “plan” feature is aimed at giving users greater oversight of the AI assistant’s proposed actions before it starts interacting with spreadsheet data. The Copilot assistant can now draft a list of planned interactions — such as changing a formula — and, before it gets to work, ask the user to “approve, edit, or answer clarifying questions,” said Jones.

We’ve covered so many deals during Prime Day that my head is spinning. But after four days of doing our damndest to try and help folks save money, the thing I’m most hyped for is a simple tool for fixing my most egregious mistakes. I bought a pair of Vampliers.

If you’re unfamiliar, Vampliers are Japanese-made pliers designed for one simple task: removing stripped screws. I don’t even remember when I first heard of them, but I threw one on my wishlist long ago and thought, “I should get one of these eventually.” That time is now: the standard 6.25-inch VT-001 Vampliers is on sale for $25.58 (about $6 off) at Amazon.

Over the years, I’ve abandoned far too many stripped screws in door jambs, junction boxes, and even under my desk. And now I finally have a means to correct my mistakes. Maybe it’s the full pot of coffee I’ve been drinking every day since Tuesday, but I’m done with Prime Day and I’m done with stripped screws.

After a rollercoaster negotiation process with the Trump administration that dragged on for two weeks, Anthropic’s Mythos 5 is finally back in action — at least, somewhat, for a select group of organizations, according to a letter from the government to Anthropic that was viewed by The Verge. Fable 5, however — the public-facing Mythos-class model — appears to still be in limbo, with no apparent timeline for a rollout agreement.

The letter, dated June 26th and sent by Commerce Secretary Howard Lutnick to Anthropic co-founder Tom Brown, who had been recently leading negotiations, states that there’s been a “revision to the license requirements” based on the fact that Anthropic has recently “worked with the U.S. government to address risks” associated with Mythos 5 and Fable 5.

Anthropic spokesperson Danielle Ghiglieri told The Verge in a statement that the company had “received notice from the US government that Mythos 5, our strongest cybersecurity model, can be redeployed to a small group of cyber defenders and infrastructure providers.” She added, “We are working to provision the approved set of providers and restore their access to Mythos 5 as quickly as possible. We are pleased to see this progress and continue to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again.”

The US government did not lift the export control directive it hit Anthropic with two weeks ago, which barred any foreign national from accessing either model (including Anthropic’s own employees). Instead, the government essentially made an exception for Mythos 5, approving a select group of organizations for access the same way it did for OpenAI’s GPT-5.6, announced earlier today. Under this exception, Anthropic employees who aren’t US nationals and members of the approved organizations who aren’t US nationals are all greenlit to access Mythos 5, according to the letter.

“These efforts have yielded significant progress,” Lutnick wrote. “In addition, Anthropic has committed to work with the U.S. government on protocols and standards and releases for [Mythos-class models]. In light of this progress … I have determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model.”

Pressure has been building on the Trump administration to make a change in the case-by-case type of regulatory environment it recently adopted, especially as Anthropic’s competitors’ cybersecurity-focused models kept getting better and better — and even pulling ahead of it on some cybersecurity-focused benchmarks. Pressure was also building within the US AI industry, particularly fears about the AI advancements China could make while top US AI labs were sidelined. Plus, there was the fact that top US government departments, like the National Security Agency, had lost access to Mythos 5.

Now, Anthropic has essentially the same deal OpenAI has: Release the model in a limited preview setting, for approved organizations only, such as trusted enterprises and the US government itself. Both AI labs hope that more general availability comes soon, both for enterprise deals and for public access (like with Anthropic’s Fable 5), but it’ll be ultimately up to the Trump administration. Though many have called for AI regulation, including AI labs themselves, a current consensus among some tech leaders is that the past two weeks haven’t been the way to do it.

In fact, OpenAI wrote in the company’s GPT-5.6 blog post, “We don’t believe this kind of government access process should become the long-term default. It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them. We are taking this short-term step because we believe it is the strongest path to broader availability in the coming weeks, while we work with the Administration to develop the cyber Executive Order framework and a repeatable process for future model releases.”

Lutnick wrote in his letter to Anthropic, “All other requirements of the June 12 letter remain in effect until further notice.” He added, “I reserve the right to reevaluate and adjust the scope of license requirements on [Mythos 5 and Fable 5], should circumstances change.”

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Hayden FieldClose

  Hayden Field

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Hayden Field

• AIClose

  AI

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All AI

• AnthropicClose

  Anthropic

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Anthropic

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

We’ve already rounded up the best Philips Hue deals of Prime Day, but if you’re looking for something a little more budget-friendly, Govee’s latest sale is worth checking out. The company has heavily discounted several of its color-changing smart lamps, including the Table Lamp 2 ($53.99, down from $79.99), Floor Lamp Basic ($59.99, regularly $99.99), the portable Table Lamp Classic ($55.99, down from $79.99), and Uplighter Floor Lamp ($125.99, down from $179.99). I own all but the Portable Table Lamp Classic and regularly recommend them to friends.

No matter which lamp you choose, you’re getting a similar experience. Each supports millions of colors and adjustable white light, comes with a variety of preset scenes, and can sync its lighting to music. You can also create custom scenes, schedules, and automations in Govee’s app or control the lamps with your voice using Amazon Alexa or Google Assistant. Plus, with the exception of the Floor Lamp Basic, they all support Matter, allowing them to work with other major smart home platforms too.

The biggest differences come down to size and where they’ll fit best. The small Table Lamp 2 is ideal for a nightstand or desk, while the portable Table Lamp Classic swaps the power cord for a rechargeable battery, making it easy to move from room to room or take outside onto a patio. Just keep in mind it isn’t weather-resistant, so you’ll only want to take it out when it’s not raining.

The Floor Lamp Basic and Uplighter Floor Lamp, meanwhile, are both designed to brighten larger spaces. The minimalist Floor Lamp Basic is the smaller and slimmer of the pair. If you want something more dramatic, the taller Uplighter offers three independently controllable lighting zones that can project different colors onto your walls and ceiling at the same time.

And Apple is the Hamburglar

Apple’s AI strategy is even more brilliant than Nvidia’s. 

It’s built on a three-tier routing system. When you ask Siri to do something, a built-in orchestrator in the operating system decides how complex the task is. According to third-party estimates, around 85% of requests are handled on your Apple device by Apple’s own small, efficient models. (It does things like summarizing text, prioritizing notifications, cleaning up photos, or suggesting replies.) Roughly 12% of all queries get sent to Private Cloud Compute, Apple’s own server infrastructure running Apple’s larger models on Apple silicon in Apple-owned data centers. Only the hardest 3% of queries get routed to an external partner model.

This design lets Apple avoid the ruinous cost of training a frontier model from scratch. Microsoft, Google, Meta, and Amazon each spend tens of billions of dollars per year on GPU clusters, energy, and research teams to build and run trillion-parameter models. Apple doesn’t. Its own models are deliberately small and run on chips Apple already sells you, so the inference cost is basically absorbed into the device. It only needs a frontier model for that tiny sliver of hard queries, which is where the partnership strategy kicks in.

We’ve arrived at the final day of Prime Day, which at this point should probably be called “Prime Week.” We’ve found discounts on all manner of gadgets, including TVs, smart home tech, chargers, headphones, and more. Some of the best deals have started selling out at some retailers, so if you’ve been craving a popular upgrade like the AirPods Max 2, time is running low.

The good news is that our team is still hard at work, and in addition to the deals that remain in stock, the retailers sometimes save up a few extras for the last day (like this Echo Spot that got a little cheaper). This roundup is our pride and joy; the culmination of over four days of deal hunting by our entire team. We’ve worked tirelessly for the last week and arrived at a list of over 120 discounted items (and growing) that we’re happy to share with you.

Of course, our Prime Day coverage spans every category The Verge staff touches, and is a great place to explore the full breadth of discounts we’re able to find on the stuff we’ve tested, regularly use, and love. We genuinely enjoy helping you save on cool tech and fun gadgets that are actually worth your hard-earned money, especially when everything is getting more expensive.

Smartwatch and wearable deals

Home theater and speaker deals

Update, June 26th: Struck some out of stock vendors and added new deals to gaming, headphones, and smart home.

Apple just raised the prices on Macs and iPads in response to the rising costs of memory chips, right in the middle of Amazon Prime Day. That means existing discounts (even small ones) on Apple laptops like the MacBook Neo, MacBook Air, and MacBook Pro just became significantly better deals.

As an example, the 13-inch MacBook Air M5 that used to start at $1,099 and is currently on sale for $949.99 (at Costco) will soon carry a full retail price of $1,299. That initial $150 discount just became a savings of $350 compared to Apple’s upcoming pricing.

There’s no guarantee how long existing sales will stick around, and no official word on when major laptop retailers will reset all their Mac pricing to reflect the increases. Best Buy and B&H Photo have already updated their pricing, but some older prices or even cheaper deals are lingering at Amazon. Costco lists its member-only deals as valid through June 26th or 27th.

If you were considering an Apple computer purchase, it might be best to make a move while you can. We’re outlining all the deals we can find below.

$590

Apple’s most affordable laptop, with a 13-inch display, A18 Pro chip, 8GB of RAM, and optional Touch ID fingerprint sensor. Read our review.

Read More

The base 256GB MacBook Neo is selling for $589.99 at Amazon and Costco (in the citrus color only). Normally that’d be a paltry $10 discount, but with Apple raising the Neo’s base price to $699 it’s now a savings of $110. The 512GB version with Touch ID is $689.99 (a $110 discount compared to the new price) at Amazon and Costco. Best Buy still has the base 256GB model and 512GB model for the old prices of $599 and $699, respectively. That at least saves you the $100 price increase for now.

$950

Apple’s 2026 MacBook Air has an M5 chip and starts with 512GB of storage that’s significantly faster than previous generations. It’s still as excellent as ever despite a higher base price. Read our review.

Read More

The base 13-inch MacBook Air with an M5 chip, 16GB of memory, and 512GB of storage is $949 at Amazon, Best Buy, and Costco. That’s such a better price than the $1,299 Apple is now charging, but you have to be a Costco member. The 13-inch Air with two more GPU cores and 1TB of storage is $1,149.99 ($449 off the new price) at Amazon The 15-inch Air with base 16GB / 512GB configuration is currently on sale for $1,149.99 (also $350 off its new price) at Amazon, Best Buy, and Costco.

As good as the Neo is, the Air is still Apple’s best all-around laptop. Its balance of performance, battery life, and portability is still unmatched. And for $949 or $1,149 you’re getting a laptop that should last you the better part of a decade for what’s sadly now such a good deal.

$1550

Apple’s 14-inch M5 MacBook Pro takes everything we loved about the M4 model — including all-day battery life and a 12-megapixel webcam — and improves upon it with an M5 chip. It also offers faster storage and a 10-core CPU / 10-core GPU that’s up to 3.5x faster at AI tasks. Read our review.

Read More

The base 14-inch MacBook Pro with an M5 chip is my personal favorite sweet spot of Apple’s lineup if you do frequent content creation but can’t fathom paying much higher M5 Pro / M5 Max prices. It’s on sale for $1,549.99 ($450 off Apple’s new MSRP) at Amazon, Best Buy, and Costco. The base model revised for 2026 comes with 16GB of RAM, 1TB of faster storage, and has lots of ports. It’s the ideal step up from a MacBook Air, but paying Apple’s new price of $1,999 makes it much less attainable than it was.

$2050

The M5 Pro and M5 Max versions of the latest MacBook Pro deliver exceptional power and speed, making them ideal for creative work while offering best-in-class battery life. Read our review.

Read More

The higher-end M5 Pro and M5 Max MacBook Pros come in 14-inch and 16-inch sizes. Here’s a rundown of what each one’s base configuration starts at with Apple’s new pricing, and what you can currently buy it for:

• 14-inch MacBook Pro with M5 Pro (15-core CPU / 16-core GPU, 24GB, 1TB): $2,049.99 ($449 off the new price of $2,499) at Amazon and Best Buy
• 16-inch MacBook Pro with M5 Pro (18-core CPU / 20-core GPU, 24GB, 1TB): $2,549.99 ($449 off the new price of $2,999) at Amazon
• 14-inch MacBook Pro with M5 Max (18-core CPU / 32-core GPU, 36GB, 2TB): $3,299.99 ($800 off the new price of $4,099) at Amazon or $3,319 at Best Buy
• 16-inch MacBook Pro with M5 Max (18-core CPU / 32-core GPU, 36GB, 2TB): $3,649.99 ($750 off the new price of $4,399) at Amazon

There are of course many different configurations of current MacBook Pros out there, with RAM and storage upgrades that climb to crazy amounts — too many to neatly list without it getting overwhelming. But even base configs for all of the above models are excellent.

$1150

The latest version of Apple’s all-in-one desktop computer features an M4 chip and support for Apple Intelligence. Unlike its predecessor, the base model now offers 16GB RAM by default, along with a new 12-megapixel webcam. Read our review.

Read More

The M4 iMac from 2024 is going from a starting price of $1,299 to $1,499. You can still get a base model with 8-core M4 chip, 16GB of memory, and 256GB SSD in various colors for $1,149.99 ($350 off the new price) at Amazon, Best Buy, and Costco. The 10-core M4 iMac with four Thunderbolt 4 ports instead of just two is $1,349.99 ($350 off the new price) at Best Buy and Costco.

We may not all agree with why the 24-inch iMac still exists or who needs it, but it’s a lovely all-in-one computer. And the Mac Mini and Studio are impossible to find thanks to AI bros and their OpenClaw agents. Womp.

Update, June 26th: Revised copy where applicable to reflect that some retailers have transitioned to Apple’s new pricing. Removed or struck deals / old pricing that’s no longer available. Changed headline.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Antonio G. Di Benedetto
• AppleClose

  Apple

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Apple

• DealsClose

  Deals

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Deals

• GadgetsClose

  Gadgets

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Gadgets

• LaptopsClose

  Laptops

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Laptops

• Prime DayClose

  Prime Day

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Prime Day

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

• Verge ShoppingClose

  Verge Shopping

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Verge Shopping

In this case, the hackers had gained access to login details and passwords for active users of the networks, including the IT professionals guarding it. ASIO had set up a specific team to deal with the issue of cyber sabotage.

Australia isn’t alone in facing threats from the same state actors, Burgess said. “We struggle to find a single country in our region that has not been compromised by this state’s cyber apparatus.”

This meant that Australia is facing a persistent threat in the future, one that could have consequences for the way that the critical infrastructure is deployed and managed. “The biggest challenge is the cumulative one: in a degraded security environment defined by concurrent, cascading, compounding threats, when resources are limited, how and what do you prioritize?” he said.

Chinese tech giant Tencent is set to launch an AI assistant inside WeCom, its Slack-like collaboration tool for enterprises. The new tool, Dayuan, is built on the latest large language models from Chinese AI developer DeepSeek.

Tencent announced the news in a post on Chinese messaging platform Weibo by Tencent’s public relations manager Zhang Jun. Dayuan will automatically understand user requests and will respond according to the demands of the user, he wrote, according to a translation by Bloomberg. “At any time within WeCom, simply swipe left to summon Dayuan. It can intelligently recognize the interface you’re on, understand what you’re asking, and help you resolve issues more effectively,” he wrote, according to the report.

In addressing the Chinese enterprise market, Tencent has an advantage over other companies in the AI space because it has a vast reservoir of customers who use WeCom. Earlier this month, it announced a range of AI productivity agents to address the demand for more AI tools across enterprises.

Tencent has been intensifying its efforts in the AI space in an attempt to beat US competition. In April, it launched an updated version of its Hunyuan model to catch up with more established AI companies such as ByteDance, Alibaba and DeepSeek.

The launch of Dayuan with its vast supply of user data will provide a step-up for Tencent and will reinforce Chinese efforts to establish serious AI competition to US products.

Altman reportedly told staffers that the administration was getting nervous about the capabilities of the latest AI tools. It didn’t go as far as forbidding access to foreign users but it’s clear that the White House is looking to act as the power of the new models becomes more apparent.

The administration’s actions will undoubtedly cause some anxiety among AI companies, particularly in light of OpenAI’s and Anthropic’s upcoming IPOs. There will be concerns that new software developments could be postponed or even halted. However, it should also be noted that the administration was already displeased with Anthropic over its moral stance on defense issues, so the action against Mythos should be placed in context.

Indeed, the government is trying to play down such fears. Bloomberg quoted a White House official as saying that the Trump administration continues to collaborate with frontier AI labs to develop shared approaches for addressing the challenges of scaling the technology.

Arduino’s Raspberry Pi-like microcomputer the Uno Q is joining many other devices in getting a price hike this week. The 2GB version of the Uno Q, which is powered by a Qualcomm Dragonwing QRB2210 processor, will increase in price from $44 to $59, while the 4GB model is getting a slightly steeper increase from $59 to $79. If you’ve been eyeing the Uno Q as a cheaper alternative to Raspberry Pi’s current lineup, you have a little over a week to grab either version at the current price as the increases won’t take effect until July 6th.

In a letter shared through the official Arduino blog, the company’s chief product officer Marcello Majonchi says Arduino’s “memory component costs have more than doubled” over the last six months alone, “fueled by the enduring demand for AI applications.” The company has so far been able to hold off on raising prices “thanks to support from Qualcomm” which purchased Arduino last October, but “with no near-term signs of relief in memory supply or pricing, continuing to absorb these costs is no longer possible.”

This consequential uncertainty is affecting tech stocks. Asian stock markets fell sharply on Friday, led by a sell-off in technology firms. Trading on South Korea’s Kospi was temporarily suspended as a result — for the third time this week. Apple shares are down again, and the sell-off is expected to continue.

So, who’s winning?

The beneficiaries of this memory squeeze are not hard to identify. While consumers face higher prices for phones, laptops and games consoles, the companies driving memory demand — the hyperscalers and AI firms building out server farms at extraordinary scale — are posting record revenues. The costs flow down; the profits flow up.

Some, including Naomi Klein, see it as a kind of strip mining of human intellect and ingenuity, a cultural wealth transfer in which human innovation is packaged and resold as competing chatbots, for a fee. 

I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.

1 The Trump administration has asked OpenAI to limit its next model release
It wants to vet the first GPT 5.6 users before a wider launch. (Bloomberg $)
+ OpenAI said each of the initial partners will be government-approved. (FT $)
+ It’s the first US firm to be told to restrict an AI model before release. (Axios)
+ Anthropic is also still feuding with Washington. (MIT Technology Review)

2 Apple and Xbox have hiked prices, blaming AI-driven chip costs
Some MacBooks, iPads, and Xboxes are going up in price by over 20%. (BBC)
+ Apple’s shares plummeted after the announcement. (NBC)
+ AI data center demand has pushed up memory and storage prices. (WSJ $)
+ The shortages have been dubbed “RAMaggedon.” (The Verge)
 
3 Colossal and the US are building an endangered species “biovault”
It aims to cryptopreserve over 2,300 plant and animal samples. (Wired $)
+ It comes amid growing threats to endangered species protections. (NYT $)
+ Colossal is also growing chickens in artificial eggshells. (MIT Technology Review)
 
4 The US has banned Polestar from selling its EVs due to anti-China rules
The Sweden-based company is majority-owned by China’s Geely. (CNN)
+ The ban is because its connected-vehicle tech is linked to China. (Reuters $)
+ What happened to China’s overseas EV factory boom? (Rest of World)
 
5 China is betting on humanoids to beat its demographic decline
It wants the robots to narrow the labour gap. (FT $)
+ Gig workers are training humanoids at home. (MIT Technology Review)
 
6 The “fingerprints” of a black hole’s event horizon have been detected
The discovery was made by studying ripples in space-time. (AFP)
 
7 OpenAI is now expected to delay its IPO until next year
It’s been spooked by choppy global markets and SpaceX’s slump. (NYT $)
 
8 Data centers have moved to the forefront of environmental lawsuits 
The litigation is linked to energy sources, water consumption, and air pollution. (Guardian)

9 A master gene that turns on human development has been uncovered
It results in cells forming a human body. (New Scientist $)

10 Grok’s most popular feature? Smut
It accounts for “well over half” of the chatbot’s traffic. (The Information $)

Quote of the day

“The most advanced AI is built by a handful of American companies, on American soil, under American law, and what the rest of us are permitted to do with it can change on a Friday afternoon.”

—Nathan Benaich, AI investor at London-based venture firm Air Street Capital, tells the Financial Times about the geopolitical reality of US AI dominance.

One More Thing

MAX-O-MATIC

How technology helped archaeologists dig deeper

In 1991, construction workers in Manhattan unearthed hundreds of coffins. Further investigation revealed that the remains were between 200 and 300 years old, and they were all African and African American.

This discovery came at an inflection point in scientific history. Breakthroughs in chemical and genetic analysis allowed researchers to figure out where many of these people were born, the physical challenges they faced, and even the routes they took from Africa to North America.

Today, archaeologists are using techniques they could only dream of then: lasers, 3D photography, lidar, satellite imagery, and more. These tools are revealing where people came from, how ancient cities were built, and the lives of those who built them.

The organization has chosen to work with state governments rather than the US federal government but even partnering at the state level may be too broad, he said, given that the industry makeup varies from state to state, with some leaning towards knowledge workers, while others are more manufacturing-focused.

AI-washing

Technology analyst Carmi Levy said Raise Us’s mission is laudable, but, “Unfortunately, for all its slick website and optimistic-sounding copy, Raise Us is possibly the highest-profile example of AI-washing thus far, a well-intentioned but ultimately futile attempt to illustrate that something is being done to cut through the uncertainty and empower the workforce of tomorrow. In reality, it’s a convenient vehicle for a diverse range of stakeholders, including state governments, educators, financial institutions, and companies like Amazon, Anthropic, Microsoft, and the OpenAI Foundation, to virtue-signal that they’re doing their utmost to solve the AI employment crisis and mitigate the impact of this generational technology deployment on everyday workers.”

Like Andersen, he noted that since many of those same companies have already had multiple rounds of tech-driven layoffs, “It’s more than a little rich for them to be joining an initiative whose mission revolves around protecting workers from the worst risks associated with AI.”

Apple is expected to shake up its usual Mac silicon release strategy, with Bloomberg’s Mark Gurman reporting that there won’t be Pro or Max versions of the upcoming M6 chip. Instead, Apple wants to “fast-track technologies that it originally planned to release later” with the M7 launch next year.

The Cupertino company will reportedly only release a base model M6 chip “as early as this year,” with the base M7 chip set to launch sometime in the first half of 2027. The M7 Pro and M7 Max are then expected to follow “as early as the end of 2027.” That could leave the most powerful Macs in a lurch since Apple’s base chips are typically aimed at entry-level MacBook, Mac minis, and iMacs. The upcoming M5 Ultra will likely be the only near-term upgrade opportunity for Apple’s higher-end devices.

This would mark the first time that Apple hasn’t released Pro and Max variants of its silicon processors, which, until now, had accompanied every new version since the M1. Apple is still wrapping up its M5 lineup — the M5 Pro/Max were released in March, and the flagship M5 Ultra chip is slated to launch “as early as this year,” according to Bloomberg’s Mark Gurman, as part of a new Mac Studio.

The M7 lineup, which is reportedly designed around “major advancements to on-device AI processing,” would be the next series to include Pro, Max, and Ultra variations that power Apple’s most powerful devices. The M7 Ultra is on track to release sometime in 2028, Bloomberg reports.

Trump Mobile’s T1 Phone is now available for anyone to buy directly. The phone has previously trickled out to a small number of preorder customers — though The Verge is still waiting for both of its handsets — but now you can claim your own without having to place a $100 deposit.

Android Authority spotted the change to the Trump Mobile website, which now lets you pay $499 directly for the phone, rather than paying $100 first. In true Trump Mobile fashion, they’ve half-assed the website update: One page on the site still says you can “join the waitlist” for the phone, while the homepage encourages you to “get in line now,” and still mentions the now-deprecated $100 deposit.

It’s not clear how fast you’ll actually get the phone. There’s limited evidence that Trump Mobile has shipped more than a handful of phones to customers so far, and it’s possible that anyone ordering today would still have to wait for the company to process existing orders first. Or not! This company is weird.

If for some reason you want to buy a T1 Phone today, it’ll set you back $499 plus tax. Trump Mobile still describes this as “promotional pricing,” though as of yet there’s no sign of the price rise that executives told me was coming back in February. We don’t know when the T1 Phone’s price will change, or how much it will eventually cost, though I was told it will be “less than $1,000.”

For those keeping track at home, I’m thrilled to confirm that as of this week The Verge has paid in full for both of its phones, though we’re still waiting for any confirmation that either has shipped. Any bets on when they actually turn up?

Got inside information on Trump Mobile or the Trump phone? Reach out securely from a personal device to tips@theverge.com, or see our How to Tip Us page.

In a study published in 2023, Emma Lawrence at the University of Oxford, who studies the effect of climate change on mental health, and her colleagues reviewed the evidence linking mental-health outcomes to ambient outdoor temperatures. They found that during heat waves, there was a 9.7% increase in the rate of hospital admissions for people with such conditions. 

“People who live with mental-health conditions are among the most susceptible to the physical impacts of heat,” says Lawrence. People with schizophrenia were found to have been three times more likely to die during the record-breaking heat wave that affected Canada in 2021, for example.

In order to protect people, we need a better understanding of the mechanisms underlying these effects. After all, a lot of things change when it’s very, very hot. Some people may end up stuck indoors, avoiding outdoor play and exercise, and it can be difficult to get a good night of sleep, for example. Sleep, socializing, and exercise are all really important for our mental health. 

But whether unusual heat does something specific to our brains is, as Wortzel puts it, “the million-dollar question.”

Research in lab animals suggests that excessive heat can alter the way chemical signals work in our brain. The levels of neurotransmitters like serotonin, for example, seem to increase when rats and mice are exposed to high temperatures, according to multiple studies. The heat may also interfere with the way networks in our brains communicate with each other. It might affect the way oxygen reaches our brain cells.

“There are so many biological reasons why brains may be negatively affected by heat,” says Wortzel.

It’s not exactly surprising that RAMaggeddon is making new tech hardware really expensive. But if you’ve been in the market for things like a new computer or tablet, this week has been filled with sticker shock. Given how many companies announced price hikes related to component shortages, it seems unlikely things will get cheaper any time soon.

Valve started the week off by finally revealing the price of the Steam Machine, its long-delayed, console-like PC that, in our tests, has performance on par with the PS5. It starts at $1,049, nearly double the price of the six-year-old PS5, and that’s for the base configuration with 512GB of storage and without a controller. A bundled controller adds an extra $79, and if you want the 2TB model, you’ll have to tack on another $300. (If you still wanted one soon at that price, well, I hope you entered the lottery.)

This high-end Steam Machine with a bundled controller costs $1,428. Photo: Kelsey McClellan / The Verge

Ahead of launch, Valve had warned that the component crisis had forced it to revise its pricing plans. When asked in a recent interview by PC Gamer the one thing he would change about the Steam Machine, Valve engineer Yazan Aldehayyat said: “Make it cheaper.” While the company hasn’t spelled out its initial planned price, it seems like if you shave off $250 or $300 from the current Steam Machine prices, you’ll land in the ballpark of where Valve was originally thinking.

Microsoft introduced some new, cheaper options for its Surface devices this week, though they’re still expensive. The company’s lineup now includes a 12-inch Surface Pro at $849 and a 13-inch Surface Laptop at $949, but they come with a major drawback: Microsoft cut the amount of included RAM in half, dropping it from 16GB to 8GB. Still, they are more affordable than the previous base model versions of those devices, which recently jumped up to $1,049 and $1,199 as part of a batch of Surface price hikes.

And just today, two major companies announced price increases. Apple introduced sweeping price hikes of its own affecting MacBooks, iPads, and even devices like the HomePod and Apple TV. The MacBook Neo, which was a steal at its original $599 price, now starts at $699, for example. (We’ve got a list of all of the new prices right here, and they’re not pretty.) Apple gave Bloomberg a statement on the price hikes, saying they were due to component cost increases at a scale the company has “never seen” because of the data center boom.

The MacBook Neo isn’t quite as good of a deal as it used to be. Photo by Amelia Holowaty Krales / The Verge

Xbox announced higher console prices of $100 or more, meaning the six-year-old Xbox Series S now starts at $499.99 with the price tags only going up from there. In a blog post, Microsoft blamed the component shortage, saying that “console storage and memory prices have increased by more than 2.5x and we expect another doubling by the fall of 2027.”

If you’re looking for a new computer or computing device, whether it’s for gaming, or for school, or whatever you might need it for, the memory and storage shortage means that hardware just costs more now. And with RAM and SSDs still in very short supply — a situation that will likely be the case for a very long time — these higher prices probably aren’t a temporary thing but are instead our new reality. If even Apple, known for its supply chain mastery, has been forced to raise prices, is anyone immune?

The big AI data center buildout is largely to blame. If an individual making a PC at home or a tech company looking for parts for their next product needs components like memory or storage, they’re competing with an AI company or a hyperscaler who might be willing to pay far more for what’s becoming increasingly limited inventory. Valve said as much in an interview published this week; if the company doesn’t take prices as offered, then the suppliers “never talk to us again.”

The competition for parts for data centers is all based on a bet that AI will be a foundational part of how people will use computers and interact with tech in the future. The hyperscalers are racing to build out infrastructure to train their AI models and offer AI-based services for people and businesses. But the lack of supply for components means that buying computers is increasingly more miserable for everyday consumers. Today, everything in tech is more expensive than it was yesterday.

There’s no end in sight: hyperscalers are spending eye-watering amounts of money to build compute capacity so they can shove AI into everything we do. That’s making computers more costly, and it could just be the beginning of what’s to come.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Jay PetersClose

  Jay Peters

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Jay Peters

• GadgetsClose

  Gadgets

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Gadgets

• GamingClose

  Gaming

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Gaming

• NewsClose

  News

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All News

• PC GamingClose

  PC Gaming

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All PC Gaming

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

As far as prices go, Apple is kind of a reverse canary in the coal mine.

With its famously generous margins and immense purchasing volume, it can afford to ride out price fluctuations in its supply chain in a way no other consumer tech company can. So when Apple raises prices across nearly all of its product lines, you know that shit is well and truly real.

That’s what happened earlier today: Apple increased pricing across Macs, iPads, HomePods, and even the Vision Pro. Prices jumped hundreds of dollars in many cases. The MacBook Neo’s key feature — a $599 starting price — is now $699. The iPhone appears to be safe for now, but I’d be shocked if we don’t see higher starting prices on the iPhone 18 series when it debuts in a few months.

It’s all so alarming because Apple doesn’t typically mess with pricing on its current models. It certainly doesn’t participate in anything as common as a sale. If you walk into an Apple store to buy a new MacBook, you can count on it being the same price no matter what day or week or month it is. Sure, you’ll find discounts from third-party sellers on certain products, or maybe a gift card with your purchase if Apple is feeling really generous. Otherwise, an iPad generally costs the same price, year-round, right up until the day a new model is introduced. If the price is going to go up, it’ll go up on the new model; the current one holds steady.

But even for a company like Apple, the memory crisis is re-writing the rules of consumer tech pricing. First it came for the game consoles: the PlayStations, Xboxes, Switches, and Steam Decks. They all received price hikes, blamed squarely on the memory shortage. It came for laptops. Phones have suffered, too. The Pixel 10A is a barely warmed-over version of the 9A, and its best feature is that it didn’t get any more expensive than last year’s model. Samsung’s S26 phones were victims, with less storage and higher prices than the previous models. Every corner of the industry has been touched by the crisis, and Apple’s price hikes today underscore what a crappy year it’s been for consumer tech.

The funny thing is, more than a few tech companies picked this year to debut unique, premium devices. It’s the unfortunate reality of the yearslong R&D cycle. Apple is poised to launch its most expensive iPhone ever if it debuts a folding iPhone as rumored. Valve released its much-anticipated Steam Machine at twice the price of PS5. Samsung released the Galaxy Z Trifold for a small fortune. An ambitious game console from a company with a great track record of improving its existing hardware? That might just weather the storm. But a big, expensive phone with a questionable value proposition? Well, we already know how that went.

If nothing else, RAMageddon is going to quickly sort out the winners and losers. And if there was any doubt remaining, we know now that every consumer tech company is being forced to reckon with the memory shortage — even Apple. The company may have stumbled its way to the other side of its AI debacle and right into a crisis of another kind.

This week, Instagram launched a series of new features for its smart TV app that are all designed to get people to spend more time on the platform through the biggest screens in their homes. In addition to vertical Reels, Instagram for TV — which is currently available for Amazon Fire TV, Google TV, and Samsung Smart TVs — users can now watch disappearing Stories and horizontal videos with aspect ratios similar to what you typically see on YouTube. And soon, Instagram will make a big push for longform, episodic content and TV-focused “live creator experiences.”

This foray into the TV space feels very different from the Meta-owned company’s previous attempts to capture more of our attention by adding functionalities borrowed from competitors like TikTok, Snapchat, and Periscope (RIP.) It’s all contingent on the idea that people want to sit on their couches to watch Instagram content that they would typically consume on their phones.

The fact that we take our phones with us basically everywhere means that Instagram is always just a few taps away. The inherent portability of phone-based Instagram is arguably the biggest reason why the platform has managed to hit 3 billion monthly users. Scrolling through Instagram’s discovery page is something to do while you’re bored and already swiping through your phone. It’s easy to fire the app up and send videos to friends while you’re commuting, waiting for an elevator, or using the bathroom. But Instagram’s latest pivot is geared toward a more stationary experience that’s meant to be shared with people in the same room. This is a big bet.

Instagram says that community feedback is what inspired it to roll out the TV app’s new features. Some of them — like being able to cast Reels from your phone to your TV — make a certain amount of sense. If someone wants to share a funny video with a group, it’d be easier to throw it up onto a larger screen instead of having everyone cluster around a phone. What feels a little more dubious are Instagram for TV’s new dedicated channels that are supposed to “make it easier to find videos everyone in the room can enjoy together” even though the feeds are tailored to individual users’ interests.

It’s also hard to imagine people wanting to watch a carousel of Stories — which are essentially vertically oriented multimedia slideshows — on their televisions. Stories look awkward on widescreen displays, and interacting with them using a remote is much less intuitive than swiping through them on a phone’s touchscreen. This is probably why Instagram is also testing a dedicated section for widescreen content. The company’s plan to develop longform content and episodic series with creators is a sign that Instagram knows its TV app can’t just be a place where people scroll through videos that only last a few seconds.

It’s obvious that Instagram is trying to capture some of the attention away from YouTube and newer upstarts like the many microdrama platforms that have begun popping up.

Microdramas are already a booming business. The sector’s revenue is estimated to hit $14 billion by the end of the year. On mobile, Instagram is in a very good position to get into the microdrama game given the app’s built-in, massive user base. Most microdrama apps require you to sign up and then either pay per episode or subscribe to monthly or yearly plans to access their content. If Instagram can convince creators to start producing more of these kinds of videos on their own, the company could very well turn itself into a new microdrama hub.

Instagram may be leading with the vertical content for the moment, but with the understanding that longform, horizontal videos need to become central to the TV app for it to succeed. Speaking with The Hollywood Reporter, Instagram’s VP of product Tessa Lyons said that the company sees shortform content as “a very accessible way” for creators “to get into telling longer and more episodic stories.” Lyons also explained that part of what prompted the Instagram for TV’s new emphasis on longer videos was seeing creators using Instagram to promote their projects that live on other platforms.

But even if more Instagram creators embrace longform, landscape video, the TV app is going to have a tough time becoming something akin to YouTube. Like Instagram, YouTube has copied some of its competitors’ features, but YouTube has always been a destination for a wide variety of content that lends itself to viewing on a larger screen. You might start watching a YouTuber’s skit or video essay, but it’s easy to wind up putting on a playlist of music videos or a feature-length studio film. The breadth of YouTube’s content library is why it continues to be the most-watched streamer in the US, and while there’s plenty to see on Instagram, the two platforms aren’t exactly comparable in terms of what they offer viewers.

There’s a level of commitment that comes with watching something on TV that Instagram content has never really had to worry about before. You don’t exactly have to lock all the way in to appreciate a short clip on your phone. But a video on the TV should be something that can sustain your attention for more than a few minutes.

This TV pivot might work if Instagram can actually get creators to start churning out polished series instead of flooding the space with more phone-native content. But without a sharp influx of compelling projects that are crafted specifically for TV viewing, this could end up being another instance of Instagram losing the plot while chasing its competitors.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Charles Pulliam-Moore
• CreatorsClose

  Creators

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Creators

• EntertainmentClose

  Entertainment

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Entertainment

• InstagramClose

  Instagram

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Instagram

• MetaClose

  Meta

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Meta

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

• TikTokClose

  TikTok

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All TikTok

Microsoft announced today that the price of all Xbox models will rise in August, the second time in less than a year as memory prices continue to wreak havoc on every industry from cars to computing. Thankfully, we’re already out looking for Prime Day deals, and we were able to find some discounts on both the Xbox Series S and Xbox Series X ahead of the price hikes.

The 1TB Xbox Series X is discounted to $573 and $573.99 at Walmart and Target respectively, a significant reduction from the usual $649.99 price point. This is the model with a built-in 4K Blu-ray player that will go up to $799.99 in just over a month — and it’s what I’d recommend if you’re shopping for a new console to play GTA VI on later this year.

If you’re just looking for the disc-free Xbox Series S, which is usually $399.99 but will rise to $499.99 in August, Dell has the 512GB model on sale for $349.99, which is the current best deal. Target isn’t far behind, with the same console listed for $379.99, which may be worth it if you want to pick it up locally or have other discounts at Target like the Circle Card.

Thanks to the component crisis, it’s a bad time to want a new computer. But if you are waiting on a preorder for the Framework Laptop 13 Pro — which Framework’s CEO has called the “MacBook Pro for Linux users” — the company shared good news on Thursday that might mean yours will cost less than you expected. The company announced that it has secured cheaper new PCIe Gen 5 SSDs from its supplier, Adata, and if you had put in an order for a configuration with a 500GB SSD, you’ll now get a 1TB SSD at a lower price. Framework also said you can also now order configurations with the new 1TB or 2TB SSDs onboard, though I’m not seeing them in the DIY configurator yet.

But Framework has some bad news, too. The company says that it has “gotten signals that CPU pricing will be increasing soon,” so it expects that it will have to “adjust the overall system price on Framework Laptop 13 Pro in the coming weeks for new orders.” Not a great sign for the component crisis, which has already led to major price increases across tech hardware, including some also announced on Thursday for Apple devices and Xbox consoles.

Framework initially announced the Laptop 13 Pro in April and originally planned to ship the first units in June. The company recently delayed those initial shipments to July due to issues with the laptop’s haptic touchpad and its custom display.

The Trump administration, apprehensive of potential security issues, has reportedly asked OpenAI to stagger the release of its next big-ticket model, GPT-5.6.

The Information reported that OpenAI CEO Sam Altman told employees Wednesday in a company Q&A that it would release GPT-5.6 in limited preview form — granting access only to a small group of enterprise customers — in compliance with a request from the federal government. During that preview period, the Trump administration itself would reportedly approve access for customers on a case-by-case basis.

It’s a more favorable deal than the Trump administration gave OpenAI rival Anthropic, which earlier this month received an ultimatum requiring it to suspend access to its Mythos 5 and Fable 5, models. The administration issued an export control directive that prohibited “foreign nationals” from accessing the technology (including any of Anthropic’s own employees who were not US citizens). The seemingly heavy-handed approach to AI regulation, after promises from the Trump administration that it would take a “speed wins” approach to the technology and encourage an American AI exports program, raised alarm bells across the tech industry.

Now, some of those concerns seem to be coming to pass — and in a decidedly uneven way, depending on the company.

Android 17 is getting a dedicated gaming mode for foldables that will put a virtual gamepad with touch controls on half of your screen to theoretically make it easier to play games.

With foldable gaming mode, which is set to launch in the coming months, the virtual controller emulates physical button presses at a system level and is designed to work “with any game that supports physical controllers,” says Google’s Mishaal Rahman on Reddit. For the actual inputs, the virtual controller will have a D-pad; left and right virtual sticks; A, B, X, and Y buttons; L1, L2, L3; R1, R2, and R3; and a start button. And you’ll be able to configure the gamepad in several ways, such as keeping the virtual joysticks inline or staggered from each other, scaling the size of the buttons, and toggling haptics on or off.

Turning on the mode “is as simple as unfolding your device, either before or after launching a compatible game,” Rahman says. You can also choose to hide the gamepad, and if you connect a physical controller, the virtual gamepad will turn off on its own.

“Android allows you to play a wide variety of games on the go,” says Rahman. “While touch controls work incredibly well for many titles, certain games are better enjoyed with physical gamepads. The problem is that carrying a Bluetooth controller or a snap-on gamepad with you everywhere isn’t always convenient. We want to bridge that gap, and we’re addressing it with a new feature in the Android 17 platform release that’s specifically tailored for foldable devices.”

“Vendors should provide verified accounts, smart rate limits, abuse detection, usage monitoring, contractual bans on distillation, incident disclosure, and audit rights,” Jain said. “Enterprises should ask how the vendor detects and blocks large-scale model extraction and can demand contracts that guarantee backup plans and financial refunds if the AI service gets attacked or suddenly shut down.”

Joshi said enterprise customers should also press vendors for greater transparency around model development and safeguards.

“Enterprise buyers should ask what training data was used, how it was trained, what guardrails exist, how they can audit it, and so on,” Joshi said. “Model publishers will have to come up with watermarking technology in models as well as model responses. So if the model ‘skills’ are stolen, they should be able to find the thief.”

YouTube is adding even more TikTok-like features to Shorts, including a new “clear screen” mode that removes the icons and text from the video you’re watching. In a blog post on Thursday, YouTube says it’s also replacing the “thumbs-up” button with a “heart” icon, and is introducing a new option that lets you watch Shorts at 2x speed.

As noted by YouTube, you can speed up a video by holding down the edge of your screen and lifting your finger when you want it to return to normal speed. You can lock the speed at 2x by pressing the player and swiping down. Another change allows you to mute watching by tapping the video and selecting the “mute” button.

These new additions make Shorts even more like TikTok, which has had the similarly-named “clear mode” for years — a feature that also removes captions, icons, and usernames from your screen. TikTok already offers a way to speed up videos as well, and lets users “like” a video by pressing a “heart” button.

Along with these updates, YouTube is removing the “dislike” button. It’s encouraging users to share their feedback with the platform by selecting the “Not interested” and “Don’t recommend this channel” options from the three-dots menu instead.

Although the digital services team doesn’t have full visibility into how Nuage has been received, Piédallu said feedback is positive, with the file storage and sync system largely invisible to users and operating well — aside from some bugs around synchronization at times. “They are very happy to use it. They don’t make a comparison to Google Drive or OneDrive…, it’s just working,” he said. 

The Collabora-based office application suite has been less well-received, in part because its interface is unfamiliar to many users. “When they want to edit documents, work on a [spreadsheet] or something like that, they want it to be exactly like they are used to — if they have Microsoft Office, they want [it] to work the same, to have the same options,” he said. 

Local administrations and school districts are not required to use Nuage; they can choose whether to deploy the platform or rely on proprietary software. Microsoft SharePoint and Office tools are still in use, for instance, though there are no figures available for how many people are using the software. The Ministry pays around 2.5 million euros a year for Windows licenses, for instance, across 50,000 devices for Ministry staff.

Concerning the MacBook Neo, Francisco Jeronimo, IDC vice president for client devices, said: “The $100 increase takes the entry model from $599 to $699, close to a 17% increase, one of the sharpest percentage hikes in the announcement, and it hits one of Apple’s best-selling, most price-sensitive laptops right as it is winning share. Raising the price of its fastest-selling entry product indicates that Apple believes demand will stay strong, given the value the product offers. It also signals Apple is willing to give up some unit growth at the bottom of the range to defend margin, and it widens the gap the Neo had just closed against Windows entry-level machines.”

The fact of the matter is that consumers are effectively paying an inflationary fee for the benefit of the ongoing global AI rollout. That rollout is already consuming vast quantities of investor cash, despite the shaky business plan of existing incumbents. 

“We have never seen a component price increase this much, this quickly,” Apple ​said in a statement. “We have shielded our customers from these increases so far, but we have now reached a point where we need to begin ⁠raising prices on a number of products, including today’s increases for iPad and Mac.”

In 2018, Microsoft said its researchers had detected evidence of their existence, an apparently major breakthrough it was forced to retract when the data was successfully challenged. Nature’s editors subsequently backed this up with the blunt note: “The results in this manuscript do not represent evidence for the presence of Majorana zero modes in the reported devices.”

Despite the setback, Microsoft persisted, in 2025 publishing a new paper in Nature that claimed the company had worked out how to exploit the principle in a new “transistor for the quantum age,” the Majorana 1 chip.

Earlier this month, the company launched its successor, Majorana 2, claiming that AI had helped to achieve a 1,000-fold improvement in reliability compared to the earlier chip. “With this progress, the team now expects to achieve a scalable quantum computer by 2029, cutting its original timeline in half,” the company announced.  

Artificial intelligence is rapidly reshaping retail, but not in the ways consumers might immediately notice. The biggest transformation may not be flashy virtual try-ons or chatbot shopping assistants, but in how decisions are made behind the scenes: how products surface in search results, how inventory moves through supply chains, how engineers ship code faster, and how retailers respond to customer behavior in real time. As legacy retailers navigate a fragmented and hyper-competitive landscape, AI is becoming an operating philosophy.

At Macy’s, that philosophy is more often defined by what senior director of engineering Murali Murugan describes as an “AI-first” approach. “AI first isn’t about adding intelligence on top,” Murugan says. “It’s about redesigning how decisions happen so the business moves faster and every experience feels more relevant by default.” Rather than layering AI onto existing workflows, Macy’s is embedding intelligence directly into systems that include personalization, search, operational planning, and software development itself.

The company’s strategy is reflective of a larger shift taking place across retail: moving from isolated AI pilots toward integrated systems designed to compress, as Murugan puts it, “the gap between the signal and the action.” Early efforts focused on narrow, high-impact use cases like search recommendations and customer engagement, where measurable gains in conversion and reduced friction quickly built internal momentum. “Once we established the quick wins, scaling was a business decision, not a technology debate anymore,” he says.

That momentum is now extending into conversational commerce through tools like Ask Macy’s, an AI-powered shopping assistant designed to act more like a personal stylist than a traditional search bar. Whether for a prom, a vacation, or a last-minute event, customers can describe what they need conversationally and receive curated recommendations informed by past purchases, preferences, and context.

Still, the company sees AI as more of an invisible layer augmenting human judgment than a replacement for it. The long-term vision is retail that feels increasingly seamless, adaptive, and personalized, powered by systems customers may never even notice are there.

“The real transformation in this all comes from continuous improvement,” Murugan says. “It’s about learning from the mistakes, quickly adapting to the newer technology standards that are coming into play, timing, and execution which compound into a meaningfully better customer experience.” 

This webcast is produced in partnership with Infosys.

This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff. It was researched, designed, and written by human writers, editors, analysts, and illustrators. This includes the writing of surveys and collection of data for surveys. AI tools that may have been used were limited to secondary production processes that passed thorough human review.

We’ve seen a lot of great Prime Day deals on wearables this year, from the Fitbit Charge 6 to the latest Apple Watch SE. But if you’re looking for advanced health features, the Apple Watch Series 11 is the one to check out, and it’s now down to its best price ever: $279 ($120 off) at Amazon and Walmart for the 42mm model. If you want a larger watch, the 46mm option is also down to a new low price of $309 ($120 off) at Amazon and Walmart.

While we think the Apple Watch SE — now just $199 ($50 off) — offers the best value for most people, the Series 11 is the model we’d recommend if health features are a priority. It offers capabilities you won’t find on the Apple Watch SE, including EKG support, atrial fibrillation detection, and FDA-cleared hypertension notifications. The larger battery is another advantage, lasting us nearly 48 hours in testing, so you can take full advantage of sleep tracking and sleep apnea detection. The bezels are smaller as well and the display is larger, making it easier to read at a glance, which is a welcome improvement for older adults or people with visual impairments.

Outside of its health features, the Series 11 retains everything we already like about Apple’s smartwatches. It supports handy gestures like double tap and wrist flick, an always-on display, and 5G connectivity, along with the usual smartwatch perks like support for Apple Pay. It’ll also be compatible with watchOS 27, which will introduce a variety of new Siri AI features and other software updates later this year.

This, the Foundation further added, creates a federated mechanism for agent discovery and verification without any reliance on any proprietary registry or centralized control.

Growing demand for an agent identity framework

ANS solves an emerging problem for enterprises, especially in scaling AI deployments, said Charlie Dai, principal analyst at Forrester, too. “The agent identity problem is already emerging in early production deployments, particularly where multiple agents interact across tools, APIs, and organizational boundaries without consistent authentication and accountability models,” he said.

“We have seen growing concerns around provenance, authorization scoping, and auditability in agent-to-agent interactions, especially in regulated industries and multi-vendor environments,” Dai said.

I wasn’t surprised when I got the call that my dad was dying, even though we’d been estranged for many years. He’d suffered addiction for decades and eventually ran out of time, which also meant he ran out of time to reconcile with me. About 15 years after we stopped talking, my aunt and uncle held the phone up to his ear — 1,400 miles away, between me in Connecticut and him in Nebraska — to help me say goodbye. I’ll never be sure if he understood my words, but, as I watched waves crash on the shore from the Long Island Sound, I cried and told him I loved him. I forgave him for things he probably never forgave himself for. I hung up, and, soon after, he died. Then I thought about EverQuest.

Dad bought me EQ for my birthday in 2000 when I had just entered high school. He brought it home with a grin on his face. EverQuest was one of the first massive multiplayer online roleplaying games, but at the time, the box just read “fantasy.” Years earlier, we’d explored Sierra Online adventures for MS-DOS. Compared to that, EverQuest was a revelation.

We got hooked on EverQuest immediately. My brother played a Wood-Elf Ranger, my dad played a High-Elf Wizard, and I played a Human Magician. Back in my dad’s small one-bedroom apartment, powered by CompuServe dial-up internet, we had to share EQ. If you know anything about MMOs, you’ll know that sharing one computer is a prickly situation. Taking turns in an MMO is unimaginable these days, the kind of thing you’d only be subjected to if Mr. Beast found a new way to torture people for money. We made it work somehow.

Our home was already primed to appreciate the spectacle of battle. Dad was obsessed with Lord of The Rings. It was the second movie I clearly remember him bringing us to years after the family went to see Titanic in theaters, because fantasy was a big enough deal to warrant the rare movie outing. He’d otherwise source plenty of lessons from cinematic battlegrounds, like this one from Braveheart: “First, learn to use this, then I’ll teach you to use this,” says William Wallace’s fictional movie-uncle Argyle, played by Brian Cox, tapping his noggin before he hoists a broadsword. It was one in a stable of apocryphal tales shared by a man who nearly only let himself cry in public while watching the Notre Dame football movie Rudy. Now I see that these things gave him a permission slip for being vulnerable. I wish he knew they grow permission slips on trees.

EverQuest ended up being one of the most important boxes I ever opened

Thanks to dad’s fateful decision to buy one of the first MMORPGs, EverQuest ended up being one of the most important boxes I ever opened. I went on to spend more than 10,000 hours in EQ and its sequel over the years. Most of those hours had nothing to do with playing the actual game. I formed friendships that could not have existed otherwise, meeting people from all kinds of backgrounds all over the world. You could be anyone you wanted, but EverQuest felt like a rare place online where what people wanted was just to be themselves.

Because of these friendships, I couldn’t put the game down, to the point where my grades at school suffered. This was a difficult thing to explain to parents at the time who justifiably couldn’t appreciate the idea of me spending a month hunting a digital winged horse in South Karana all to collect a Pegasus Feather Cloak for an epic quest that would result in… oh shit, I’m doing it again. Yeah, you can barely even explain this now without sounding like an internet-pilled maniac. Eventually I did capture the legendary Quillmane, but only because I spent weeks befriending a higher-level Ranger who could help me track it down. We kept in touch for a while.

Grades be damned. I ended up learning more from EQ than in a lot of my classrooms, including how to type 120+ words-per-minute by necessity, because you simply had to be able to tell your party as quickly as possible that a mob was on the way to kill everyone. It was an excellent way to avoid taking a typing class.

So it was exciting to learn that the latest owners of EverQuest, which is still going 27 years later, are doing a kind of reboot of the original. Sure, I’ve had problems with the franchise after it was sold off by Sony Online Entertainment and ended up in the hands of Daybreak Game Company. But there’s something rare happening here. Daybreak’s new EverQuest Legends is supported by the developer of an original EQ fan project, Project 1999, which revived the early experience via emulation and custom code for the people who were there from the beginning. It’s not the kind of deal on the scale of Rockstar buying FiveM, the team behind some of the biggest GTA roleplay servers, but it’s just as meaningful.

I’d already played Project 1999 and came away thinking it was one of the greatest digital game preservation projects ever. It’s a largely faithful implementation of the original game with some custom improvements — but forget all of that. Almost as soon as you sign into P99 you’ll be greeted by people who want to shower you with free gear and talk to a person with a similar background. Because, of course, who couldn’t be like-minded if they chose to be in this particular place in history at this particular time in the present. I’ve quietly joked to myself for years that if I make it to a retirement home, I’d be locked into EverQuest reruns, and that actually seems possible now.

Thanks to a closed beta code from the developers, I logged in to EverQuest Legends last week. I have a laundry-list of complaints about Legends, mostly because I don’t think it aims high enough. The “quality of life” improvements are perceptible only to the people who have been most hooked on the game since 1998 — anyone else would open this thing, think it’s an ancient computer virus, and immediately uninstall it. I find myself scratching my head at some decisions and wishcasting for others. But 24 hours of gameplay later, I remember why I kept coming back for years.

Back in 2000, I devoted myself to the adventure and ended up reaching much farther than the other turn-havers in the household. Dad, who was once very smart and ambitious — a former Marine Corps officer — set lowly sights in EQ, and I’ll never know why. At a time when the max level was 60, he resigned himself to forever live a level-12 life camping Orcs on the unofficially named “Orc hill,” just outside the tree city of Kelethin in the Greater Faydark zone. While my brother and I dared explore nearby Crushbone Castle, the Orc stronghold, dad was content to post up outside as if the beachhead had never been won.

For the 90 minutes we allowed him to enjoy the game without hounding him for our turn, he set out to do his best on this hill; dual-wielding a beer and a cigarette in one hand, and the keyboard in the other, harnessing his demons to rid another world of theirs. Occasionally he’d be interrupted by a friendly passerby who forced him to work on his keyboard skills, replying in the best possible haste with single-finger typing, well before the age of voice chat.

I have to imagine Sisyphus was happy on Orc hill

At some point, dad reached the limit of gaining experience from slaying lower-level creatures, which effectively froze him in time within the game. Thanks to getting a philosophy degree instead of joining the armed forces (sorry, dad), I know this is just The Myth of Sisyphus. Dad climbed Orc hill over and over again. He must have known at some point that he wasn’t making any progress. (Addiction, in its many forms, has a similar texture.) I have to imagine Sisyphus was happy on Orc hill. But, then, I think maybe he wasn’t. I never witnessed him doing the one thing really worth doing in EverQuest: making friends. I can barely remember the recipes for quests these days but I will forever remember the late nights I spent talking to strangers who were almost always willing to open up and offer their hearts to the unknown. I wonder if he ever snuck on late at night while I was asleep to do the same.

EverQuest Legends is a lot like Project 1999 in the sense that it’s entirely built around a pretty well-established group of nerds who will keep coming back to this place over and over again until the end of time. Commercially that makes it a quixotic quest for recruiting new video game buyers, but also an endearing one. At a time when the biggest video game companies in the world struggle to make live service games work, EQL is a blessing to those who grew up with the concept at the right time. This place was real for you, and it still can be.

The biggest irony is that EverQuest Legends is actually designed to be more lonely than its parent. It enables extremely powerful solo-playing options that make it possible to plow through much of the game by yourself. Aside from the global chat channel, I’ve noticed that people seem less chatty in general, which should be expected considering you don’t need other people as much as you used to. Smart move for a dwindling player base, but also maybe proof that lightning in a bottle diminishes the more you open it.

I’ve long thought if I got a blank check I’d buy the rights to EQ and reboot it myself — that’s how much this thing means to me. But as I get older, something like EverQuest Legends seems less like an opportunity to make an old video game great and more like a great opportunity to keep an old internet community together. Till death do us part, but at least now without the part where you have to spend hours retrieving your own corpse. Hey! There’s a quality of life improvement.

Even though I rarely thought about dad during the later years, I’d always had in the back of my mind, if we had ever reconnected, that I would have loved to share with him the advancements that came in video games. The mind-blowing graphics, set pieces, and depth — bordering now on what we expect from great movies. Just to see if there was a spark in his eye like I’d had in mine. Maybe to see that grin once more.

Now I just wish I could have gotten to share his favorite place with him again. A small little part of an old game where he would have felt right at home at level 12, guarding the tree city against the forces of Orc hill. It made him happy. And now it’s one of the last places I can visit his memory.

The author, revisiting Orc hill at night.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• TC SottekClose

  TC Sottek

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by TC Sottek

• AnalysisClose

  Analysis

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Analysis

• EntertainmentClose

  Entertainment

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Entertainment

• GamingClose

  Gaming

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Gaming

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

3 The EU has joined a US pact to stop relying on Chinese AI
Much of the rest of the world seems to still be a battleground for control. (FT $)
+ China is expanding its AI push in the Global South to counter the US. (The Wire China)
+ Chinese AI experts are freaking out about the AI arms race. (Wired $)

4 OpenAI and Broadcom have unveiled their first jointly designed AI chip
Jalapeño is built to power large-scale AI systems like ChatGPT. (NYT $)
+ It’s part of OpenAI’s push to “build the full stack.” (CNBC)

5 A new report shows ICE has built a vast hi-tech surveillance system
It includes facial recognition, drones, and data scraping.(Guardian)
+ Is the Pentagon allowed to surveil citizens with AI? (MIT Technology Review)

6 Electronics can now be printed onto living tissue
Which could enable smart implants and ingestible diagnostics. (The Economist $)

7 The data center boom is sparking a third wave of inflation 
Demand for memory chips is pushing prices higher.(WSJ $)

8 Companies are scrambling to curb spending on AI token “chewing”
Accenture data shows non-technical staff are draining budgets. (404 Media)

9 Claude Design is creating a bland wave of website uniformity
The AI tool is homogenizing the internet’s aesthetic. (The New Yorker $)

10 Elon Musk has lost his trillionaire status
Thanks to SpaceX stock coming back to Earth. (Business Insider)

To celebrate its new status as No. 1 in JD Power’s initial quality ranking among mainstream automakers, Ford is opening up about the challenges it has faced in recent years, especially around its reliance on automated systems in production and design. It turns out that those automated systems were not as robust as previously assumed, requiring Ford to hire experienced technicians — sometimes bringing back former employees — to correct errors made by the company’s robots.

In Ford’s view, AI is both powerful and prone to pitfalls. Its effectiveness depends entirely on the quality of the data used to train the AI models. In addition, the automaker underestimated the value of the institutional knowledge accumulated by its more veteran engineers who had worked through multiple vehicle-development cycles. And this combination of phenomena led to a drop in quality in Ford’s vehicles.

“Mistakenly, we thought that by just introducing artificial intelligence and adjusting the design requirements that we had, that that would produce a high-quality product,” said Charles Poon, VP of vehicle hardware engineering, in a briefing this week with reporters.

“Mistakenly, we thought that by just introducing artificial intelligence and adjusting the design requirements that we had, that that would produce a high-quality product.”

— Charles Poon, Ford’s VP of vehicle hardware engineering

According to Poon, some of the company’s most experienced personnel left before all of their accumulated knowledge could be fully transferred into Ford’s automated systems. That necessitated bringing back some of those employees to retrain those systems, or in some cases, mentor younger engineers who were currently struggling to maintain Ford’s vehicle quality. Poon said that Ford hired, promoted, or brought back over 350 experienced engineers to rebuild that layer of expertise. In addition to guiding younger engineers, they’ve also been tasked with improving the data collection and AI training that underpin Ford’s automated systems.

“That’s where some of our most experienced engineers have had experience solving and identifying those problems before they creep into the system,” Poon said.

Ford currently leads the industry in the number of recalls, and its quality ratings have slipped over the past several years. Those challenges became more pronounced recently, with the difficulties associated with the launches of the Explorer and Aviator, supply-chain disruptions during the covid pandemic, and the noticeable growth in the number of its vehicle recalls.

According to Ford’s COO Kumar Galhotra, the automaker eventually concluded that its approach to quality had become too fragmented. Different departments operated in silos, and the company relied heavily on a “find and fix” philosophy that focused on identifying defects after they appeared and correcting them as quickly as possible. While that approach could address immediate problems, it did not prevent those problems from occurring in the first place.

“We’re moving from that find-and-fix mentality to preventing issues before they occur,” Galhotra said. “We’re focused on enablers and early indicators versus outputs. Stop admiring the problem and start solving it.”

The transformation extends beyond vehicle hardware. Software and digital teams now work much more closely with vehicle engineering, manufacturing, and supply-chain teams, executives said. And Ford is now attempting to combine the speed and flexibility associated with software development with the rigor and validation requirements of automotive-grade engineering.

Historically, this wasn’t always the case. Ford was only discovering software bugs late in the process because it wasn’t fully leveraging the rapid iteration cycles available, Poon said. That said, the automaker couldn’t push out software updates as fast as consumer electronics companies with the mentality that it could “move fast and fix later,” Poon said. Vehicles, unlike smartphones, operate in a safety-critical environment where customers depend on software functioning correctly from the moment the vehicle is delivered. To fix this, Ford created a dedicated 40-person software quality assurance team with the sole responsibility of preventing problems before they occur.

But don’t think that Ford isn’t dedicated to integrating AI into more of its processes. The automaker says it has dramatically expanded its automated testing capabilities, adding more than 100,000 new AI-powered tests designed to identify edge cases and stress software systems under a wide range of conditions. Because the testing framework is highly automated, software changes can be rapidly revalidated even late in development, ensuring that modifications do not introduce new defects.

“Because these tests are highly automated, even if we have a late change in the software, we can rapidly run back through the entire validation process to guarantee it works perfectly well before it reaches the customer,” Poon said. “We’ve established software reliability as its own rigorous disciplines with strict metrics.”

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Andrew J. HawkinsClose

  Andrew J. Hawkins

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Andrew J. Hawkins

• AIClose

  AI

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All AI

• CarsClose

  Cars

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Cars

• FordClose

  Ford

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Ford

• NewsClose

  News

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All News

• TransportationClose

  Transportation

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Transportation

Climate change is squeezing the grid from all sides, affecting both supply and demand. Heat can affect power availability, from generation to transmission infrastructure, as I covered in my latest story. But climate change is also helping push electricity use higher—and countries in Europe and around the world will need to adapt. 

In the US, nearly 90% of homes have air-conditioning. That means many grids see their highest demand in the summer months, and the risk of brownouts and blackouts is at its worst. 

People are often quick to cast air-conditioning as a villain, and it’s true that the technology will account for a major chunk of the globe’s rising energy demand in the future. But the reality is that heat waves can be incredibly dangerous, and as climate change pushes temperatures higher, that risk is becoming more real in parts of the world that haven’t historically had to worry quite so much about heat. 

In Europe, air-conditioning is historically much less common, with about 20% of homes across the continent using it. Some countries, including those getting hit by this heat wave, have even lower rates—the UK comes in at about 5%, and Germany is around 3%. 

But those numbers are starting to tick up as people adapt to increasingly brutal summers. As they do, we should expect higher electricity demand, and stress for the grid—just as in the US. And utilities often have to look across borders to buy more power, driving prices up for everyone. 

“It’s not just an incremental step,” Jay Gambetta, the director of IBM Research, said during a press conference on Tuesday. “It’s a meaningful leap forward.” Within a decade, Gambetta expects chips with nanostacking to be widely used in data centers, where their improved efficiency could help the facilities better manage their energy consumption.

“Absolutely, it’s transformational,” says Dan Hutcheson, vice chair of TechInsights, a technology analysis company. “This puts another ten, fifteen years on the roadmap.” 

Compared to IBM’s previous state-of-the-art architecture, the company reports that chips built with this new approach can do as much as 50% more work in the same amount of time and be up to 70% more energy efficient. 

The architecture offers a general way of laying out transistors, and IBM will partner with semiconductor manufacturers to make the actual chips. It anticipates chip designers will deploy the design in many different types of chips, including GPUs and CPUs. “I expect to have many conversations with designers about how they can use this technology,” Huiming Bu, IBM’s vice president of global semiconductor R&D, said in the press conference announcing the new design. 

A layer cake

Engineers created IBM’s new chip layer by layer, like a cake. They start by fabricating transistors on one layer of silicon. Then, they place a silicon layer on top of these devices, and they fabricate another layer of transistors directly on top of that. Finally, they create the electrical connections between the two layers of transistors. This kind of vertical stacking, which combines two types of transistors, is known as a complementary field-effect transistor, or CFET, explains Qing Cao, a professor of materials science and engineering at the University of Illinois at Urbana-Champaign, who was not involved with the work. 

The company isn’t the only one pursuing this general approach. The biggest chip manufacturers—Intel, Samsung, and TSMC—along with competing research lab Imec in Belgium have been investigating CFETs. IBM says its design is distinguished by the fact that the second layer of transistors do not sit directly on top of the first layer’s transistors; rather, they are staggered, which the company says simplifies wiring, among other advantages. 

CFETs like those in IBM’s nanostack architecture contrast with another common approach to making two-tiered chips, such as AMD’s 3D V-Cache and Huawei’s forthcoming LogicFolding technology, Cao says. In those approaches, engineers fabricate the transistors on each layer of the chip independently before bonding the two together. IBM’s new method allows for more precise alignment of the layers, which is important for performance because transistors are so tiny, says Cao. 

Meta is bringing back the Facebook Creator Studio page manager, now “reimagined” as a standalone AI companion app. The new app aims to make it easier for creators to connect with their audiences, and show them “exactly how to grow on Facebook,” according to Meta’s announcement.

Meta’s AI Creator Assistant is a central focus of the newly reimagined app — users can ask the chatbot to provide performance tracking insights and tailored recommendations for improving engagement. It can also be used to find “the most important comments” left by the user’s audience, and “instantly draft replies in your voice.”

The Creator Studio app isn’t widely available yet, with TechCrunch reporting that Meta is currently testing it with “select creators.” There’s no mention of when it’ll roll out to everyone, but Facebook creators can join a waitlist to get early access. The original Creator Studio experience was shut down in 2023, in favor of pushing users to Meta’s more comprehensive Business Suite platform to manage their pages and schedule content.

Now suppose you have another computer, the Contrarian, and you put the Oracle inside it. When you give the Contrarian a program, it passes it to the Oracle and then does the opposite of whatever the Oracle says the program will do. So if the Oracle assesses the Contrarian’s program and thinks it will halt, the Contrarian will run forever. If the Oracle thinks the program will run forever, the Contrarian will halt. Either way, the Oracle’s assessment is wrong, so the classification problem is undecidable.

The proofs that Super Mario is undecidable rely on a more complex version of this idea. The team’s argument breaks down the video game using a technique called a reduction, in which mathematicians convert a problem they’re trying to solve into a problem they already know something about. “The classic example I remember in a math class is: How do you make a pot of boiling water?” Demaine recalls. “Well, I fill up the pot with water from the sink, and then I put it on the stove, and then it eventually boils. Okay, now I’ll give you a pot of water that’s already filled. How do you make a pot of boiling water? Well, I empty out the pot first and reduce to the previous problem.”

In their particular world of platforms and porcupines, the team broke down their Super Mario level into localized parts of Mario’s path called gadgets, which they could use to prove that the level was undecidable.

“A gadget in our sense is anything in your environment that decides whether or not you can go through one pattern [within a level],” explains Jayson Lynch ’12, MEng ’15, PhD ’20, a CSAIL research scientist and head of algorithms at MIT FutureTech. For example, in one gadget Mario might need to jump on a platform to avoid a monster as he makes his way across the screen. As a PhD student mentored by Demaine, Lynch spearheaded the formalization of gadget theory and worked on some of the earlier Super Mario papers but did not study the game’s undecidability.

One of Lynch’s favorite Super Mario gadgets is the door gadget, which works like a door that Mario can open, traverse, and close. The door in question is always either open (when the Spiny is on the right) or closed (when the Spiny is on the left). So if a Spiny is pacing back and forth on the left of the door, Mario has to navigate beneath the moving Spiny and jump up to hit a brick block just as the Spiny reaches it. This bumps the Spiny to the right side, which opens the door and allows Mario to travel across the traverse path and get to the spot where he can close the door. Once there, he must time another jump beneath the pacing Spiny to send it back to the left side of the gadget, closing the door behind him. 

Mario opens the door by bumping the Spiny from the left to the right. With the Spiny out of the way, Mario can go through the open door and follow the traverse path to the other side. Once there, he’ll be able to bump the Spiny back to the left and close the door.

Since a door is always open or closed, its state can be used to simulate a true or false statement, with open being true and closed being false. Earlier Super Mario papers had strung together multiple door gadgets to simulate a true-or-false problem that complexity researchers already knew to be hard. But to show undecidability, the team used Super Mario level editors to put together another device, called a counter gadget, that tallies the game’s monsters and obstacles. 

If you can build a machine with even just a few of those counters, Demaine says, you can simulate an arbitrary computer—one that could essentially do anything a non-quantum computer could do, given enough time and memory. And with no limit on the number of monsters, such a machine could have infinitely expandable memory, even though the level size stays the same, which he calls “pretty wild.” In other words, any theoretical computer can be built in a Super Mario level. “You could use it to solve anything you can use a computer to do,” says Demaine. “You could have it do your taxes, or compile your code, or run an LLM, or optimize your class schedule.” You might even build Super Mario levels that could excel at sudoku, construct optimal chess strategies, or prove any provable mathematical theorem.

After 15 years of answering etymological queries, when the bee was expanded in 2018 Sietsema began serving as a pronouncer for some of the earlier rounds as well.

AP PHOTO/CAROLYN KASTER

At the Greek Orthodox church in Springfield, Sietsema’s facility for languages proved useful when he served as a volunteer chanter, helping the priest lead services in Greek. “I do a good job with the liturgical Greek because I have the phonological knowledge to know how to make my mouth do the things that it needs to do to sound like authentic Greek speech as opposed to an American just rattling off Greek letters,” he says. 

He began taking evening classes in Byzantine chant, and before long the bishop was encouraging him to attend seminary. Merriam-Webster allowed him to work four 10-hour days so he could commute to Brookline to study at the Holy Cross Greek Orthodox School of Theology. And after four years, he earned a master of divinity degree.

Sietsema fully intended to go back to being a lexicographer, perhaps eventually getting ordained so he could serve as a substitute priest on weekends. But he’d made what he jokingly calls “a terrible mistake” at the seminary: He’d embraced his studies so enthusiastically that he became the valedictorian and had to give the commencement speech. The archbishop of America—the head of the Greek Orthodox church in the US—came up from New York to attend the ceremony, and he happened to be in need of a deacon who could also serve as a speechwriter. “A few weeks later, I got a call from the archdiocese saying ‘We want you to be ordained, and we want you to come to New York, and we want you to write for the archbishop,’” Sietsema recalls. 

In short order, he and his wife moved to the Upper East Side of Manhattan so he could begin his new post as Father Mark (he used his middle name because Orthodox priests must be ordained with a saint’s name, and there are no Orthodox Saint Brians). As deacon to the archbishop and then to his successor, he wrote their speeches and encyclicals on top of many other duties—including chauffeuring them through New York City traffic—and traveled with them around the country and to Greece, meeting President Clinton, ambassadors, members of Congress, Elie Wiesel, and South Africa’s Anglican Archbishop Desmond Tutu along the way. But after two years, as the father of a newborn, he was eager to move on from a job that required putting in as many as 14 hours six or seven days a week. So in 2000, he returned to Michigan to become pastor of the Holy Trinity Greek Orthodox Church in Lansing.

“The World Series can be a four-game sweep and the Super Bowl can be a blowout, but the National Spelling Bee always comes down to one last word.”

Not long after settling into parish life, Sietsema got an unexpected call from the Scripps Spelling Bee. His wife had served on the event’s word panel from 1997 to 2000, and he had traveled with her to one of the members’ off-site gatherings in 1998. He’d tagged along to dinner one night, and they were pleased to meet the person who was responsible for pronunciations in the bee’s official dictionary. But now, just a few weeks before the 2003 event, there was a crisis: The longtime pronouncer had suddenly died. The veteran associate pronouncer would step into his role and take on the job of giving spellers their words, but a new associate pronouncer would be needed to answer spellers’ questions about word roots, monitor pronunciations, and be prepared to serve as the pronouncer if needed. Could he do it? Honored to be asked, Sietsema got the okay from his bishop and said yes. 

Little did he know it would become a permanent gig. After 15 years of answering root-word queries, when the bee expanded in 2018 he began serving as a pronouncer for some of the earlier rounds as well—though never for the finals. Now he’s the head of a team of associate pronouncers. “It’s just wonderful to see these young people blossom right in front of you, asking their questions and analyzing the word on the spot and figuring out how it all goes together,” he says. He dismisses the idea that the kids have photographic memories, saying they’re “really just good little word detectives.” 

As a member of the bee’s word panel, Sietsema attends multiple daylong meetings to create and fine-tune each year’s list by mining the 500,000 or so words in Merriam-Webster’s unabridged dictionary. “For an introductory round, you want something that’s an interesting word, a useful word, but something that’s gettable,” he says. “For the later rounds, you really want to find something that’s going to challenge the speller. And it’s nice to have a word that’s analyzable.” “Rooty” words—those with obvious roots—are ideal. 

The advent of unabridged online dictionaries has streamlined how students prepare for the bee, which once required wading through the dictionary manually to compile word lists. Today, it’s easy to generate lists of words derived from a particular language to study their roots, for example. Meanwhile, the competition has become increasingly fierce, and once-verboten terms like geographical names are considered fair game. For some of the words in the hardest rounds, “it looks like you’re just taking a spoonful of alphabet soup,” he says. “And that’s for the spellers who really, really are committed to learning just about every word they can in the dictionary.”

As our graduates know better than anyone, preparation in calculus is effectively an admissions requirement at a place like MIT—which means that students in schools with no calculus classes are in practice locked out of an essential route to STEM careers.

Recognizing this glaring need, we set out to find a solution. With support and inspiration from the Siegel Family Foundation, in the fall of 2025 the Institute launched the MIT4America Calculus Project. Developed by the MIT Scheller Teacher Education Program (STEP) Lab, the Calculus Project recruits and trains MIT undergraduates and alumni to provide weekly long-distance calculus tutoring for students in underresourced high schools across the country.

Reflecting the Institute’s longstanding commitment to national service, the MIT4America Calculus Project supplies an innovative answer to a hard practical problem, and it taps the uncommon skill of MIT’s people to create opportunity for others and spread the educational impact of the Institute beyond our walls. 

The project is in its early phases, so far engaging 30 MIT undergraduates and seven alumni tutors. From its initial work with 14 school districts across the country, it’s on track to collaborate with about 20 this summer. 

The demand is clear—and the response from the students we’re reaching makes it all worthwhile. This spring, the first Calculus Project students were prepared for their AP exams, thanks to their own persistence, diligence, and curiosity—and to the generosity, care, and patience of a dedicated group of people from MIT.

While the court still hasn’t signed off on the massive settlement resolving Epic’s antitrust lawsuit against Google for having a monopoly over Android’s app store with Google Play, the tech giant says it will start rolling out changes to the way it handles billing for developers worldwide. As announced in March, the flat 30 percent billing fee is being replaced by “lower, decoupled fees” that partially decouple the billing and the app store.

How much of a cut Google will take from transactions now depends on whether it’s for a user whose first install came before or after the new structure, how much a developer has earned, and whether or not the developer uses Google Play’s billing system with its 5 percent additional fee, instead of an alternative system or linking to their own website.

For apps that make over a million dollars annually, that will be 20 percent for new in-app purchases and 10 percent for subscriptions. However Google has also announced Games Level Up and Apps Experience programs for “exceptional” and “premium ” experiences that meet its guidelines by working across platforms (like tablets, smart TVs, or Android Auto), meeting benchmarks for memory usage and crash rates, and supporting features it recommends (like cloud saves or phishing-resistant sign-ins) to qualify for a lower rate on both new and existing installs.

Other program changes will go into effect in some areas at the end of September, at the end of the year, before rolling out to the rest of the world after September 30th, 2027.

Rockstar Games has finally given Grand Theft Auto VI a price ahead of the game’s November 19th launch. But while announcing that the game would cost $79.99, Rockstar also confirmed that the physical versions of GTA VI won’t include a disc, and instead will only contain a download code inside the box. Not only is that a disappointing decision for people who like to own physical games, but given the scale of the next GTA, it also sets a bad precedent for the rest of the industry.

It’s no secret that digital games have overtaken physical ones. The sales data proves it: Capcom, for example, recently announced that 93 percent of its games were sold digitally in its last fiscal year, and it expects that number to go up next fiscal year. Devices like the PlayStation Pro, Xbox Series S, and Steam Machine don’t have an included disc drive, further exacerbating the issue.

There are a lot of advantages to buying digital. You can start a download from your couch. You can store multiple games on one hard drive so you don’t have to get up to play something else. Storefronts like Steam or the PlayStation Store don’t run out of inventory of the newest game you’re interested in, and you can often get games at a cheaper price thanks to frequent sales.

But it’s becoming increasingly clear that digital ownership has significant disadvantages, too. If a game you don’t own digitally is removed from a storefront, whether that’s for things like licensing, artificially limited availability, or even the store eventually closing down, your only option is to hope you can find a physical version. If your account on a platform is banned, even if that ban isn’t warranted, you might be locked out of your digital library with no way to play those games unless you buy them again or hope your account gets restored. You can’t sell or trade digital games you’ve purchased, and while there are ways to share digital games, they require some work and are usually intended just for families.

It’s also much harder to preserve digital games because they only “exist” on the hard drive of a console, PC, or device they were downloaded to. This is an issue across many industries, not just console games; there are multiple examples of things like mobile games and streaming shows becoming lost for good when they don’t have a physical version. Without physical versions, you also can’t find a used version of a game at a garage sale or a local game shop.

Yes, digital games can be much more convenient. But what you gain in convenience, you lose in permanence. And in many ways, digital games just aren’t as easy as handing a disc to a friend.

Rockstar didn’t immediately reply to a request for comment about why it’s going digital-only for GTA VI and if it ever plans to launch a physical version. We can speculate some reasons why, though. I’d suspect the decision was made in part to prevent leaks; by only being available digitally, Rockstar can ensure that GTA VI unlocks at the same exact time for everyone. (The game has already had significant issues with leaks.) The digital-only choice might also indicate that the game has a massive file size that’s too big for PlayStation and Xbox game discs.

Going digital-only isn’t an entirely new concept. Other publishers have taken the code-in-a-box approach in the past, like Bethesda did for Fallout 4 Anniversary Edition and Skyrim Anniversary on the Nintendo Switch 2. Nintendo also gives publishers the option to offer game-key cards for the Switch 2, which are another way to sell a box at a store that contains a license for a game that you have to download from the internet to play. But game-key cards aren’t locked to a Nintendo account, meaning you can share a game with a friend even though they’ll have to download it themselves.

In the case of GTA VI, everyone will have to buy their own copy of the game, whether that’s from a digital storefront or at a physical store to get a one-time download code in a box. (Though at least one retailer isn’t stocking the game.) And the entire industry will be watching how GTA VI does — it’s a game so big it’s altering the entire 2026 release calendar — so Rockstar’s code-in-a-box decision could mark a turning point that might embolden even more publishers to do the same. If more games follow suit, it would make a medium that’s already hard enough to share and preserve even more brittle than before.

Few things are as delightfully divisive as Android’s dark mode.

Some phones now ship with Android’s darker-style interface activated by default. Most reasonably recent devices offer it as a swift ‘n’ simple toggle. And most people, in my experience, have amusingly strong preferences about which approach they prefer — the standard Android “light” mode, in which screens tend to be bright and with shades of white as a foundation, and the dark mode (a.k.a. “dark theme”), where black and dark gray dominate and everything is much more muted and muddy.

It really is a night and day difference, so to speak — but no matter where you fall on the light vs. dark preference spectrum, it’s well worth your while to noddle over two pertinent points:

1. Android’s dark mode doesn’t have to be an all-or-nothing decision. With the right setup, you can use it as a dynamically activated sometimes switch that enables itself automatically based on different variables and gives you a darker, less glary motif when conditions call for it while leaving you with the lighter, brighter look the rest of the time.
2. Regardless of how often you’re using dark mode, a few easy adjustments will make it meaningfully more complete and effective as an end-to-end interface style for whatever you’re doing on your device.

It occurred to me recently that we’ve gone over several smart Android dark mode enhancements over the past weeks and months — and that, put together into a single power-pack bundle, these small-seeming items can add up to create a pretty dramatic difference in your Android-using experience, whether you’re a full-fledged dark mode convert or a more light-preferring vampire skeptic.

Here, specifically, are five easy ways to make Android’s dark mode meaningfully better for you.

[Keep the nerdy knowledge coming with my free Android Intelligence newsletter — something new and useful in your inbox every Friday!]

Android dark mode power-up #1: The app expansion

Up first is a feature that arose as part of our Android 17 discussion and sparked the entire idea for this collection — and that’s the one-tap switch in the latest Android version that forces every app on your phone to follow your dark mode preference, whether the program technically supports such a setting or not.

In Android 17, finding and flipping that switch will make every app turn dark whenever the system-wide dark mode is active. It eliminates the irksome exceptions that’ve traditionally stayed stubbornly light (due to developer laziness) even when your dark theme is on.

If you’ve got Android 17 on your phone already, it couldn’t be much easier to make it happen. Just look in the Display section of your system settings, tap the words “Dark theme,” then change the setting that shows up next from “Standard” to “Expanded.”

Android’s new “Expanded” option lets you force apps into dark mode compliance, even if they aren’t designed to do it on their own.

JR Raphael, Foundry

No Android 17? No problem: On devices with reasonably recent pre-Android-17 system software, you can actually find a switch buried deep in some developer settings that’ll let you enable the exact same option without any waiting.

Follow these instructions and enjoy your new universally consistent darkened dynamic.

Android dark mode power-up #2: A darkened web

That first trick fixes the issue of certain apps not following your dark mode preference — but what about the web? Most of us spend a fair amount of time in our browsers these days, and most websites won’t follow a dark mode setting and adjust their interfaces accordingly.

They absotively can, though. With the flip of a single switch buried within your browser’s bowels, you can force every website into a darkened motif whenever your system-level dark mode is up and running.

See?

Viewing a website in Chrome normally, at left, and with dark-mode-associated web darkening, at right.

JR Raphael, Foundry

Here’s the secret:

• Open up Chrome on whatever Android device you’re using. (And note that this will also work with any Chromium-based Chrome alternative, like Brave, Edge, or Vivaldi.)
• Type chrome:flags into the address bar.
• Type darken into the search box at the top of the screen that comes up next.
• See the line labeled “Darken websites checkbox in themes setting”? Tap the “Default” box beneath it, and change its setting to “Enabled.”
• Tap the blue “Relaunch” button at the bottom of the screen.

Now, when Chrome comes back a second or so later…

• Tap the three-dot menu icon in its upper-right corner.
• Tap “Settings” in the next menu.
• Scroll down until you see “Appearance.” Tap it, then tap “Theme.”
• Make sure the newly added box for “Apply dark theme to sites, when possible” — which we just magically made appear via our last little modification — is checked and active.

Chrome’s web-darkening option appears only after you’ve enabled an out-of-the-way flag.

JR Raphael, Foundry

And that’s it: From that moment onward, whenever your Android device is switched into its dark mode, any website you’re viewing within Chrome will automatically follow suit. Nothing more to it, and no further thought or action ever required on your part.

Not bad, eh?!

Android dark mode power-up #3: A dark mode schedule

Even as someone who isn’t into dark mode as a 24/7 sort of thing, I can definitely appreciate the presence of a dimmer, less glary look on my device in certain specific scenarios.

It’s incredibly easy to overlook or forget, but Google’s actually got a way to handle that for you. In fact, it’s been built into Android itself since 2020’s Android 11 release.

Just look in the Display section of your system settings and tap the line for either “Dark theme” or “Dark mode settings.” If you see a toggle alongside that line, make sure you’re tapping the actual words next to it — not the toggle itself.

Then look for the option to create a schedule.

An Android dark mode schedule, as seen in Google’s standard Android interface (at left) and in Samsung’s Android style (right).

JR Raphael, Foundry

You can create a time-based rule for when your device’s dark mode turns on and back off again, or — more intelligent yet — you can set it to automatically activate at sunset, wherever you are at any given moment, and then turn itself back off and switch you back over to light mode at sunrise.

You can also integrate dark mode into Android’s rarely noticed Bedtime Mode so that the screen getting dimmer is part of your pre-sleep winddown routine, if you really wanna get wild.

Or — ahem…

Android dark mode power-up #4: Contextual dark mode

A dark mode schedule is pretty forkin’ sensible. But the reality is that even with a time-based setup or a sunset-driven activation approach, you’ll still be using your phone in bright rooms with dark mode active and vice-versa.

And if you want Android’s dark theme present only when you’re actually in a dark room — as makes the most sense in my mind — there’s an even more intelligent option.

It comes our way via a handy little free app called Adaptive Theme. That app does one thing and only thing only: It automatically adjusts your device’s dark mode setting based on the actual ambient light around you, using your phone’s sensors rather than an arbitrary time or a not-always-relevant sunset status as a guide. It makes so much sense, you’ll find yourself wondering why your phone didn’t just work that way from the get-go.

The app does unavoidably have a slightly complex one-time setup, which I outline step-by-step here. It’s perfectly safe to do, though, and it shouldn’t take you more than a couple minutes to pull off.

And once you’ve done that, your Android dark mode will just work for you — flipping on when the lighting around you is dim (to your exact specifications) and flipping back off when you’re in brighter surroundings.

Adaptive Theme lets you take total control of how and when Android’s dark mode activates based on your environment.

JR Raphael, Foundry

Yes, please — and thank you. Last but not least…

Android dark mode power-up #5: The wallpaper wizard

Superficial as it may seem, the one piece of the puzzle we haven’t yet addressed — that isn’t ordinarily affected by Android’s dark mode setting — is your home screen wallpaper.

By default, whatever wallpaper you set at the system level stays the same even as your interface moves between its dark and light states — and when you’re anglin’ for a dimmer, less glary look in dark environments, that can be pretty darn jarring.

An app called, rather aptly, Dark/Light Wallpaper Scheduler is the answer you never knew you needed. It’s pretty self-explanatory — you tell it which wallpaper you want when your phone is in dark mode and light mode, then it automatically switches ’em out for you based on that status — but I wrote about it in detail here, if you’re interested in reading more about how exactly it works and how you can make the most of it.

And with that, my fellow Android-appreciating animal, your dark mode power-pack is complete. Now, would someone please turn off the lights? I don’t know about you, but all this talk of darkness has me hankering for a nap.

Wake up to even more useful Android wisdom every Friday with my free Android Intelligence newsletter — one practical new trick to try each week, straight from me to ye.

Welcome to day two of Amazon’s four-day Prime Day event, which, if we’re being honest, looks a lot like day one. That’s actually good news, though, because many of the best deals are still around, and some new ones have joined them. If you’ve got a Prime subscription, whether through a free trial or a discounted student membership, you’ll find our favorite deals below.

Not every Prime Day deal deserves your attention, so we’ve filled this guide with products we know and genuinely like. Our team has spent years living with, testing, and comparing everything from robot vacuums and TVs to headphones and smart home gadgets, and the deals below are the ones we can confidently vouch for. We’ve also sprinkled in matching prices from retailers like Best Buy, Walmart, and Target whenever we find them, so you don’t necessarily need a Prime membership to save.

If you’re shopping for something specific, we’ve got dedicated roundups covering Apple gear, budget-friendly picks, smart home devices, TVs, and much more. We’ll also be updating this guide throughout the day as new deals pop up and old ones disappear, so check back occasionally.

Smartwatch and wearable deals

Home theater and speaker deals

Update, June 24th: Pricing updated and added deals for several Lego sets, Philips Hue products, Echo Show smart displays, the Apple MagSafe Charger, Roborock Qrevo Curv 2 Flow, Kobo Libra Colour, and more.

But it’s also not so easy to stay there. Sceye’s aircraft, Frandsen says, has to be light enough to stay aloft but also strong enough to carry the necessary systems. It must soak up and store enough solar energy during the day to provide around-the-clock power to an electric fan that can maneuver the HAPS back into place when winds knock it out of position—mettle it proved in a 2024 test flight.

Since then, Sceye has been preparing for its big Japan test. In the flight pictured here from this spring, for example, the craft stayed aloft for 12 days as it flew to the coast of Brazil and spent more than 88 hours “parked” in various locations. Eventually, the company expects its platform could help satellite operators better serve densely populated areas.

Someday, Frandsen says, spotting a HAPS may be as common as seeing ships at port or trains on the tracks.

Just days after a gunman killed conservative activist Charlie Kirk, it became clear that President Donald Trump would use the assassination to fuel a crackdown on free speech. To avenge Kirk’s death, the administration vowed to go after so-called “antifa” (otherwise known as antifascist) terrorists. Now that promise is bearing fruit. This week, eight Texas activists were sentenced to between 30 and 100 years in prison — one for attempted murder, but most for supposedly belonging to an insurrectionary “Antifa cell,” including one sentenced to 30 years in part for moving a box of zines.

These unusually harsh sentences are a major victory for the Trump administration, one that will likely serve as a blueprint for targeting activists across the US. The emboldened administration has quickly celebrated them. In a statement, acting attorney general Todd Blanche said that the sentences show that the law will come down hard on “Antifa terrorists who attack law enforcement and federal facilities.” But many of those sentenced did no such thing.

The Texas cases concern a July 4th, 2025 protest outside the Prairieland Detention Facility in Alvarado, Texas. Roughly a dozen protesters set off fireworks and shouted messages in Spanish through a bullhorn. Then it escalated: a few people slashed an ICE van’s tires, broke a security camera, and vandalized a guard shack. When guards came out of the building and told the group to leave, some complied, but others stayed. After a police officer arrived on the scene and drew his gun, one person shouted “Get to the rifles” and fired a rifle he’d brought with him, according to charging documents. The officer, who was shot in the neck, testified that he “knew” his life was in danger and spent three to four hours in the hospital after being shot.

Benjamin Song, the shooter, said he fired because he thought the officer was going to fire on a protester, and was convicted of attempted murder. But in addition to the shooting charge, prosecutors dubbed Song the “leader of the antifa cell.” Alongside rioting and discharging a firearm during a violent crime, a conviction for providing material support to terrorists was also included — and he was sentenced to 100 years in prison.

Other defendants were convicted of lesser crimes ranging from rioting to providing material support to terrorists, catch-all charges the government has applied to acts as disparate as distributing anarchist literature and “dressing in black bloc.” But their sentences, too, spanned decades.

Two people — Savanna Batten and Elizabeth Soto — weren’t involved in planning the protest, arrived separately from the others, and left when guards told them to, before the shooting. They were each sentenced to 50 years in prison. Among other accusations, the government declared they “were part of a group that created and distributed insurrectionary materials called ‘zines.’” Daniel Sanchez-Estrada, who didn’t attend the protest at all, was sentenced to thirty years in prison for moving a box of zines — an act prosecutors claimed was “corruptly concealing a document or record.” Ines Soto, Elizabeth Soto’s husband, was granted a continuance and will be sentenced on July 1st, according to the Department of Justice. (So will seven others who pleaded guilty to one count of providing material support to terrorists, some of whom testified as witnesses in the trial.) The other sentences handed down suggest that his will be similarly harsh.

The Department of Justice reportedly conceded that the zines weren’t even illegal — they were made for a book club named after anarchist organizer Emma Goldman, which read about topics including feminism and “the eradication of artificial intelligence from the face of the earth.” But it claimed that the Sotos had, by tabling at a zine fair, provided “material support to terrorists.”

“The defendants’ violence and terrorism is an assault on Democracy,” said Reed O’Connor, a Republican-friendly judge who among others handed down the sentences. O’Connor argued that the government needs to “deter this type of conduct.” In a statement to The Guardian, Song called it “collective punishment.”

In a statement, FBI director Kash Patel said the agency “remains committed to identifying, locating, and dismantling Antifa and its funding networks across the country,” and more cases are coming. Last week, prosecutors indicted 15 people in Minnesota on a raft of charges including conspiracy to impede or injure a federal officer, solicitation to commit a crime of violence, assault on a federal officer, and the destruction of federal property. Like the Prairieland defendants, the DOJ claims the 15 people charged in Minnesota are connected to “antifa.”

The DOJ is trying to punish people for trailing ICE officers — and if it succeeds, other activists could be next.

The indictment notes that the defendants are involved with the Black Cat Workers Collective and claims they “infiltrated and exploited lawful protests” in the Twin Cities, where thousands of people resisted Operation Metro Surge, a months-long DHS operation that resulted in the arrests of thousands of immigrants and the killings of Renee Good and Alex Pretti, which the federal government has suppressed efforts to investigate and prosecute. It accuses some — but not all — of the defendants of using homemade shields, using debris to obstruct traffic, and obstructing DHS vehicles with wood, leafblowers, and other items as they left the Whipple Federal Building to make arrests.

The indictment cites communications on Signal, which was used widely to coordinate mutual aid distributions and patrols of neighborhoods for ICE vehicles in Minnesota. Patel said during the protests that the FBI had opened an investigation into activists’ groupchats. Here, the indictment says defendants coordinated “anti-law enforcement action,” practiced operational security — or OPSEC — techniques, and engaged in “counter-surveillance tactics.” One defendant is accused of kicking a government vehicle and “causing dents.” Two are accused of traveling across state lines “with the intent to kill, injure, harass, intimidate, or place under surveillance another person.” In plain English, the DOJ is trying to punish people for trailing ICE officers — and if it succeeds, other activists could be next.

Though the indictment distinguishes between legal and allegedly illegal actions, its language is slippery. Erik Davis, a religious studies professor at Macalester College in St. Paul and one of the arrested activists, expressed disbelief at the charges at a court hearing last week. “I seem to be indicted for holding meetings,” Davis reportedly told the judge. Indeed, the 94-page indictment claims that Davis moderated an “Emergency Meeting on Resistance to ICE Operation” in January and sent messages about other meetings in Signal groupchats. The indictment notes that another defendant, Isaac Auman Sant, wrote an article for an “anarchist blog.” In his article, Sant allegedly mentioned watching someone break into an ICE vehicle. Notably, the indictment does not say that Sant vandalized the car, just that he was in the presence of someone who did.

The strategy here is guilt by association. Just as White House officials justified Alex Pretti’s death by smearing him as a domestic terrorist and “would-be assassin,” any of the thousands of regular people who resisted ICE’s siege of the Twin Cities could be called an antifa terrorist — and be condemned to a lifetime in prison as a result.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Gaby Del ValleClose

  Gaby Del Valle

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Gaby Del Valle

• AnalysisClose

  Analysis

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Analysis

• PolicyClose

  Policy

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Policy

• PoliticsClose

  Politics

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Politics

• ReportClose

  Report

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Report

He warned: “Microsoft missed the internet wave, the mobile wave and now it missed the AI wave.”

Show me the money

The most immediate likely impact on Microsoft after the three massive IPOs take place will be on the company’s stock price, which could well take another hit. Investors don’t want to miss out on the AI boom, and until now, there’s been a relatively small number of companies in which they could. When Anthropic and OpenAI join SpaceX as publicly traded companies, investors will have three more choices than they did only a few months ago. That could make it tougher for Microsoft to attract AI-focused investors. And that, in turn, could knock down its stock price. 

A lower stock price means the company will have a harder time raising money when it needs it. In addition, Microsoft won’t be able to as easily buy other companies in stock-only deals, because the value of its stock will be less.

A critique published in Nature Wednesday calls the basic technology behind Microsoft’s “breakthrough” quantum computing chip the Majorana 1 into question. Microsoft unveiled the chip in February 2025 and said it featured a brand-new technology known as a topological qubit. Topological qubits, they said, would be the “building blocks” for their future quantum computer. Microsoft announced the next generation chip Majorana 2 at Build earlier this month.

But in a peer-reviewed article, Henry Legg, a physicist at the University of St Andrews, reanalyzed Microsoft’s data on their device and argued that the company’s researchers did not conclusively demonstrate a working topological qubit in the first place.

Theory predicts that the electrons in this wire behave in a collective pattern known as a Majorana particle, for which the chip is named.

Proponents of quantum computing predict that the technology’s computational abilities will advance new medicine discovery, encryption, and machine learning. Companies like Google and IBM have already demonstrated more advanced machines than Majorana 1 or 2, although presently, no one has conclusively gotten any quantum computer to perform anything useful. But Microsoft claimed that Majorana 1, and subsequently Majorana 2, paved their path toward a practical quantum computer.

Microsoft’s design, unique among quantum computing companies, involves a tiny wire, thinner than a human hair, made of the semiconductor indium arsenide stuck to a superconductor. Theory predicts that the electrons in this wire behave in a collective pattern known as a Majorana particle, for which the chip is named. Microsoft wants to encode information in the properties of the Majorana particle. (A topological qubit is to a Majorana particle as a transistor is to silicon.)

Proponents of the Majorana particle think it is promising qubit material because theory predicts that when formed into topological qubits, the Majorana should compute with fewer errors than competing materials, such as superconducting circuits pursued by IBM. This suggests that ultimately, fewer topological qubits are needed to scale up to a useful quantum computer.

That is, if Microsoft has actually made a Majorana particle. “They haven’t convincingly shown that they have Majoranas,” Legg told The Verge. “You can’t make a qubit if you don’t have the Majoranas.”

In Legg’s critique, he writes that what Microsoft claims as a signature of the Majorana particle could actually be from the formation of quantum dots, which are electron-containing structures, in the device. Quantum dots would not be useful for building the quantum computer. He also writes that Microsoft cherry-picked their data.

“You can’t make a qubit if you don’t have the Majoranas.”

Microsoft’s team published a rebuttal in Nature disputing Legg’s interpretation of their data. Legg’s critique “does not constitute a substantial scientific challenge to our findings,” the Microsoft team wrote. Legg has not “proposed an alternative model that fits all of our data,” Chetan Nayak, a physicist leading Microsoft’s quantum team, told The Verge.

Legg first posted his critique on the online physics repository arXiv on February 26, 2025, within a week of Microsoft’s Majorana 1 announcement. It took a year for Nature to conduct a peer review and publish his article.

Meanwhile, on June 2, Microsoft announced a new chip, the Majorana 2, featuring what they claimed was the next generation of their topological qubits. The company says they can build a “scalable quantum computer” by 2029. “We 100% stand behind our results,” Nayak told The Verge. “We stand by our roadmap. We stand behind our long-standing commitment to scientific rigor and dialogue.”

Legg says the company’s characterization of Majorana 2, which Microsoft wrote in a non-peer reviewed manuscript, suffers from similar problems he pointed out a year ago. “Nothing in this [manuscript] resolves the fundamental issues that so many scientists have with this company’s previous claims,” Legg told The Verge.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

• Sophia ChenClose

  Sophia Chen

  Posts from this author will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All by Sophia Chen

• MicrosoftClose

  Microsoft

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Microsoft

• ScienceClose

  Science

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Science

• TechClose

  Tech

  Posts from this topic will be added to your daily email digest and your homepage feed.

  FollowFollow

  See All Tech

Rep. Anna Paulina Luna (R-FL) says her staff used AI for “spellcheck” in an amendment summary for a major defense bill, but denies it was used for the bill text itself and says “NO Legislation is ever drafted with AI.”

Luna issued the response after accounts on X began sharing screenshots of an amendment summary for the 2027 National Defense Authorization Act. It reads: “Identical to H.R. 100 (118th Congress).11:25 AM????Claude responded: Requires the Secretary of Defense to designate Department of Defense activities, support, and operations at the southwest land border as a named operation with…”

At first, Luna’s response seemingly indicated Claude had been used for amendment text. Her post said that “staff used AI to correct a draft text and didn’t edit.” Luna added: “Not a shocker. Most staff use it. I have told them to make sure they are double checking and more thorough.” After users on X began speculating that her staff is using the technology to write bills, Luna edited her response: “Yeah my staff used AI to spell/grammar check the amendment SUMMARY, not the actual amendment text itself,” her post now reads.

She followed it up with another: “FYI NO Legislation is ever drafted with AI. All bill text from the House comes from the House Legislative Council which is prohibited from using AI. The screenshot you’re referencing is an AI summary of the bill that’s also used for spellcheck, cmon man 🤣”.

“For engineering teams, Claude Tag will help reduce time spent on debugging through fragmented Slack discussions, summarizing long incident threads, pulling context across repos, tickets, and logs, and documenting decisions after the fact,” Jena said, while for business teams, “It could enable faster decision-making from thread summaries while reducing follow-ups in cross-functional work.”

Sohail Dev Majumdar, principal analyst at Gartner, though, sees greater benefits than mere productivity gains, particularly for CIOs and other technology leaders.

CIOs may need new governance and ROI metrics

The new offering reflects growing demand among enterprises for AI systems that can work across teams, retain organizational context, participate more actively in day-to-day workflows, and generate more measurable return on investment, he said.

The next frontier in AI may depend on a new web data infrastructure layer that can enable models to discover and map this ever-expanding digital realm. This layer must be able to navigate hundreds of millions of existing web domains and billions of new URLs created each week, delivering real-time information and overcoming technical barriers.

“The data suggests there’s far more data out there,” says Or Lenchner, CEO of Bright Data, a web data collection platform. “Think of the universe: It’s out there, but you don’t know what you don’t know.”

Enabling access to fresh, relevant, and trustworthy data

While early AI breakthroughs were driven by scaling training data and model size, organizations are now encountering a fundamental bottleneck: They need to keep pace with the dynamic, unstructured, and constantly evolving nature of web data in order to ground outputs in current and verifiable information. AI performance increasingly depends not just on model architecture but on a system’s compute, networking, retrieval, and data engineering capabilities—that is, the system’s ability to quickly and reliably retrieve data that is fresh, relevant, and trustworthy.

Traditional model training relies on snapshots of information collected at a particular point in time. Training AI on such static data is no longer sufficient. To track fluctuations such as competitor pricing, consumer sentiment, and market trends, companies need a constant feed of new information, pulling data in real time along with relevant context. Their infrastructure must therefore be able to handle millions of simultaneous interactions across websites that vary by geography, language, format, and access rules.

“If it can’t retrieve real-time information, it lacks context,” Lenchner says. “In a business setting, that’s not acceptable anymore. Stale answers lead to bad decisions and disappointed consumers.”

Speed is not merely a matter of convenience; it’s a matter of necessity. Today’s organizations operate in environments where prices, inventory, markets, security threats, and customer behavior change continuously. Delayed data retrieval can reduce the usefulness of an otherwise sophisticated model.

Using live, high-quality web data can also reduce AI hallucinations because the model has a more relevant knowledge base. This builds user trust. In fact, one survey found that 56% of AI practitioners said businesses need access to real-time web data to improve trust in AI outputs. To ensure the model runs efficiently and effectively, the information must also be pared down to the appropriate essentials. 

Despite the introduction of retrieval-augmented generation (RAG), where models pull in external data at the moment of a query, many AI systems still struggle to deliver outputs that are current, contextually relevant, and trustworthy in operational settings. According to Gartner, 60% of AI projects that are not supported by AI-ready data—accurate, structured, organized, and contextualized—will be abandoned by the end of the year. 

Apple’s iOS 27/macOS 27 cycle is revealing something new: AI is only as good as the operating system that supports it. The latest beta releases show that after two years in which the company has promised to become AI-native, testers finally believe it’s happening as Apple prioritizes improved system performance and Siri AI.

For example, the second developer beta (released this week) has clarified the vague “indexing” prompt that showed up two weeks ago, replacing it with a clearer message reading “Optimizing Search and Siri.”

What is indexing doing?

Developers digging into the code found the system is proactively building contextual maps of messages, notes, and photos, allowing the updated on-device architecture to swiftly pull up personal data without compromising privacy. It still takes time, but at least its purpose has been clarified.

The improved indexing seems to deliver smoother device performance overall, reflecting the deep architectural improvements supporting the entire release. Across communities like Reddit, early beta testers are reporting an unprecedented level of refinement for such an early build.

Siri gets snappier

Where external server support isn’t required, beta testers indicate Siri is responding faster. Many are also trying out the new Write with Siri feature that appeared in the second beta with a dedicated command for this feature situated above the keyboard in supported apps.

Write with Siri replaces the earlier Writing Tools panel with a natural language interface fully integrated with Siri, which means it can write responses informed by relevant information from messages, email, and other documents that Siri AI can access. When you tap the prompt, a text input field slides down from the Dynamic Island. The new tool will write contextually-relevant messages in Notes, Mail, and Messages on request, while the original Writing Tools remain.

Testers note that while the personalized context works remarkably well for tracking down past vacation details or messages, some functions such as Visual Intelligence and cross-app action tools are heavily restricted or throttled in early builds.

This reflects that Apple has only actively deployed limited server capacity at this point in the beta cycle, with more capacity scheduled to support full operation once the final versions ship. It is important to note that Apple continues to operate a wait list before providing beta testers with access to Siri AI.

The quiet stuff that matters

The update also checks off a massive list of quality-of-life bug fixes. A notorious glitch that broke screenshot cropping in the earlier beta has been resolved, and chronic Wi-Fi connection drops have been stabilized. Native utilities such as the standalone Passwords app received layout upgrades, adding a swift “+” button directly to the main dashboard to bypass multi-step menus.

The new betas also introduce a clutch of interesting cross-device tools; developers can now fully interact with their phone’s interface using a Mac keyboard and trackpad, which makes testing apps during development much easier. They also note that audio routes flawlessly through desktop hardware when working this way.

For most iPhone users, the big improvement is that Apple has made Handoff faster and more responsive. You’ll also find solid RCS upgrades in Messages and a new Insights feature in Wallet to help improve financial management.

An ecosystem getting ready

The new operating systems also appear to support future product development plans, with code identified in tvOS 27 reportedly including a variety of Apple Intelligence frameworks. That’s going to turn HomePods and Apple TV devices into useful, integrated AI devices — just as watchOS 27 will turn Apple Watch into the most widely-used wearable AI platform.

Apple plans to release the first iOS 27 public beta in July, with the final version arriving for everyone this fall. Sadly, the new Siri AI features that arguably underpin the release will not be made available in Europe or China due to regulatory problems, something that has upset customers.

Apple introduced a beta that feels remarkably close to a public release. But the bigger picture is that by improving the inherent architecture across its platforms, Apple is much better able to support the integrated, ecosystem-wide AI on which its future will be based.

Please join me on social media at BlueSky,  LinkedIn, or Mastodon, even better, please subscribe to The Core for your daily fix of human-curated Apple News.

Unit two at the Golfech nuclear power plant in southern France shut down at about 11:45 p.m. on June 22 when the river used to cool the plant got too hot. The move was a precautionary measure, according to Brid Nelligan, a spokesperson for EDF, the plant’s owner and operator.

The power plant takes in water from the Garonne River and then returns most of it to the river at slightly higher temperatures after using it to cool equipment. French regulations limit the temperature of that return stream, so the warm water (it was expected to reach 28 °C, or around 82 °F) forced the operator to shut down the plant.

EDF, which operates France’s entire nuclear fleet, is also limiting the output of other reactors across the country—one reactor at the Nogent-sur-Seine power plant was ramped down as of Tuesday, and more will follow later in the week, Nelligan says.

Extreme heat has affected France’s nuclear industry before. At least seven gigawatts’ worth of nuclear energy was forced to shut down across the country during a heat wave in July 2025, according to data from Ember Energy. That’s more than the entire grid of Ireland. 

This time, power plant outages and limitations aren’t expected to be drastic enough to affect the ability to meet demand in France, according to RTE, operator of the national electric grid. 

Nuclear power has made most of the headlines during this heat wave, but other forms of electricity generation face similar challenges. Hydropower plants frequently run into problems when dry conditions lower the amount of water available to generate energy and force them to decrease or shut off operations. In the first five months of 2025, high temperatures and low water conditions cut hydropower supplies in Europe by 13% compared with the year before.

The project takes inspiration from efforts to fight the covid-19 virus, where Veesler’s group was among those involved in the speedy development of vaccines, anti-viral drugs, and antibodies. 

According to Ransohoff, Intercept’s advisers will include Peter Marks, a former top FDA official, as well as Moncef Slaoui, the pharmaceutical executive who led the US coronavirus vaccine effort, Operation Warp Speed.

A key challenge for Intercept will be coming up with ways to counter many—even all—viruses at one time. That accounts for the group’s interest in air-cleaning technology, such as using strong ultraviolet light to inactivate viruses. The idea, the group says, is to remove viruses from the air in the same way municipalities remove impurities from the water supply before it’s piped to people’s homes.

The US funds about $6.5 billion a year in virus research through the National Institute of Allergy and Infectious Disease, or NIAID. But that agency’s budget hasn’t grown in recent years, leaving more room for private philanthropy.

And Stripe’s Collison brothers have become some of the most reliable philanthropists in viral research. After giving away “fast grants” to help labs during the covid-19 pandemic, they later joined other donors who committed $650 million to establish the Arc Institute, in Palo Alto, which has developed AI models for biological research.

“The diversity of viruses is just too large and seems daunting, so people don’t even try,” says Veesler. “I’m happy that someone is ready to help scientists, not accepting the status quo, and doing something different.”

Sometimes the challenges we face are giant, like tunneling beneath the seafloor. Some exist at the nanoscale, as with a new ASML machine powering the future of chipmaking. Others represent problems at a planetary scale and in truly unknown territory, like replicating a volcano’s mechanism to cool the Earth on purpose.

These incredible engineering stories show we can come together to get to work and, when the smoke clears, find we’ve made real progress. Subscribe now to read all of them—and more—in the full print issue.

Stripe, Anthropic, and OpenAI are backing an effort to stop respiratory infections

The common cold comes for us all—often more than once a year. And there is no way to prevent it. The best you can do is take vitamin C and stay away from people with the sniffles.

Now, the payment company Stripe is funding a new $500-million nonprofit aiming to prevent both the common cold and the flu. Its eventual goal is to get rid of respiratory viruses altogether.

Anthropic, OpenAI, and Bill Gates have also backed the venture, which will investigate whether modern technologies can counter the common cold and the flu. Dive into the nonprofit’s plans.

—Antonio Regalado

MIT Technology Review Narrated: inside the hunt for the most dangerous asteroid ever

As asteroid 2024 YR4 hurtled toward Earth, astronomers determined that this massive rock posed a higher risk of impact than any object of its size in recorded history. Then, just as quickly as history was made, experts declared that the danger had passed. 

This is the inside story of the network of global scientists who found, followed, planned for, and finally dismissed the most dangerous asteroid ever discovered —all under the tightest of timelines and with the highest of stakes.

—Robin George Andrews

That’s putting downward pressure on entry-level hiring as AI tools absorb more routine work, said Kye Mitchell, head of Experis US, a part of ManpowerGroup. “That doesn’t remove opportunity, but it changes the expectations. Employers now expect candidates to come in with hands-on experience, AI familiarity, and the ability to contribute faster,” Mitchell said.

Compensation remains strong for specialized, in-demand skills, while more commoditized roles such as customer service, helpdesk, and some entry-level positions  are flattening. “The shift overall is toward skills-based hiring, where demonstrable capability matters more than credentials alone,” Mitchell said.

Graduates who combine technical fundamentals with practical experience, AI fluency and strong communication skills stand out quickly. Job candidates can’t rely solely on academic credentials. 

Dyson’s PencilWash cordless hard floor cleaner is on sale for $199.99 ($150 off) at Amazon. This is the lowest price yet for the powered wet mop that’s a cousin to the silly-sounding-but-awesome Dyson PencilVac Fluffycones vacuum.

The PencilWash basically takes the sleek styling and design of the PencilVac and simplifies it into a single-purpose electric mop. It’s designed for all manner of hard home flooring, and its slim head and pivoting neck allow it to get low to reach underneath furniture. Dyson claims you can get 30 minutes of cleaning time before you have to pop it onto the included charging dock, and the 300ml water tank is estimated to cover up to 1,000 square feet before needing a refill.

While the pencil name is cute, and the mop certainly looks svelte, keep in mind that it weighs around five pounds (heavier than the PencilVac) before adding any water or cleaning solution to its tank. But its self-propelling brush should help push that heft around your space without most of the weight being felt in your hand — until you’re mopping stairs.

$200 isn’t cheap for a single-purpose cleaning tool (let alone its full price of $350), but the PencilWash costs much less than the $600 PencilVac Fluffycones. Our own smart home reviewer, Jennifer Pattison Tuohy, called the PencilVac “the perfect companion to your robot vacuum.” The PencilWash is like the companion for that companion.

Everywhere you looked, there was something to do. I was mostly working with a labor crew to jump on quick fixes to things like windows and doors blown out in the fighting. But there were, of course, bigger jobs too. Most notably, there were those massive fires to put out. The Iraqi army had set hundreds of oil wells ablaze, the majority of which were still spewing soot and oil smoke into the air. On bad days, the sky would remain dark all day, and the air would burn your eyes and hurt your throat.

It was so apocalyptic that none other than Carl Sagan warned of massive environmental consequences. If smoke from the oil fires reached the stratosphere, he predicted, the result could be akin to the 1815 explosion of the Tambora volcano in Indonesia, which triggered what’s known as “the year without a summer”; global temperatures dropped between 0.4 and 0.7 °C, and crops failed around the world. Fortunately, the plume from Kuwait never made it that high, and although temperatures did decline regionally, there was little effect on a planetary scale. As it turns out, predicting what will or won’t lower global temps is quite hard. (Reader, I am foreshadowing here.) 

Firefighters from companies with monikers like the Red Adair Company or Boots and Coots (as well as less colorfully named outfits, like Bechtel) rushed to Kuwait after the war’s end to figure out how to extinguish the gargantuan blazes and cap the wells. At a hotel in downtown Kuwait City, one of the few places with working phone lines, I would occasionally run into them covered head to toe in black oil and soot. 

Putting out the fires took all sorts of creative thinking. Engineers working in the burning oil fields figured out they could repurpose existing pipelines meant to pump oil out to sea to instead pump water in from the Persian Gulf. One company from Hungary rigged up a firefighting machine called Big Wind by outfitting an old Soviet T-34 tank chassis with two turbines from a MiG-21 fighter jet, each of which could blast 220 gallons of water per second. Sadly, I never got to see it in action (except in the movies). 

Other jobs were less cinematic but no less dire. The retreating Iraqi army had left booby traps all over. They snaked hand grenades into the plumbing (at a facility where I worked, among others). They planted mines everywhere, and those had to be found and removed. Many of them were small plastic “toe poppers” designed not to kill but to maim. Hunting them was a herculean effort. And although it was mostly successful, hundreds of thousands of them, by some estimates, still remain.

Which is to say, we can’t fix everything. But we can be ambitious. We can take on the challenge of making the world better through human ingenuity. 

That’s what the July/August issue of MIT Technology Review is all about. Sometimes the challenges we face are giant, if knowable, like tunneling beneath the seafloor. Some exist at the nanoscale and represent decades of investment and research, as is the case with ASML, a company with the singular ability to produce the machines that make the most advanced computer chips in the world. Others represent problems at a planetary scale and take us into truly unknown territory, like a future where we could engineer the veil of the Tambora volcano to cool the Earth on purpose.

By the end of my 90-day contract in Kuwait, instead of damage everywhere you looked you could see the fruits of a gargantuan international rebuilding effort. The air was not clean, exactly, but inhaling it no longer felt like smoking a pack a day. On the beach, which had been pocked with mines, people swam and splashed at the edge of the gulf. The lights were on. Water ran from the taps. The markets were open. It was a remarkably different place. 

Yes, forces both within and beyond our control will always break things. People will invariably make mistakes, or act out of their own self-interest to the detriment of others. But we can also come together to get to work and, when the smoke clears, find we’ve made real progress. 

You’ll notice that the Function Arguments pane for the IF function has fields for Logical_test, Value_if_true, and Value_if_false. In our “greater than or equal to 18” example, the logical test is whether the number in the selected cell is greater than or equal to 18, the value if true is “Yes,” and the value if false is “No.” So we’d enter the following items in the pane’s fields like so:

Logical_test: B2>=18

Value_if_true: “Yes”

Value_if_false: “No”

or just type the full formula into the target cell:

=IF(B2>=18,”Yes”,”No”)

This tells Excel that if the value of cell B2 is greater than or equal to 18, it should enter “Yes” in the target cell. If the value of cell B2 is less than 18, it should enter “No.”

The IF function in action.

Shimon Brathwaite / Foundry

Tip: When using functions like this, rather than entering the function repeatedly for each row, you can simply click and drag the tiny square on the bottom right of the cell that contains the function. Doing so will autofill each of the rows with the formula, and Excel will change your cell references to match. For example, when the formula we used in cell C2 that references cell B2 is autofilled into cell C3, it changes to reference cell B3 automatically.

Autofilling a formula to subsequent rows in the column.

Shimon Brathwaite / Foundry

Find more details at Microsoft’s IF function support page.

3. The SUMIF and COUNTIF functions

SUMIF is a more advanced SUM function that allows you to add up only the values in a range that meet the criteria you specify. To use this function, you must specify the range of cells to apply the criteria to, the criteria for inclusion, and, optionally, the sum range, which is the range of cells to total if that’s different from the initial range. The syntax is as follows:

=SUMIF(range,criteria,[sum_range])

Note that any criteria with text or mathematical or logical symbols must be enclosed in double quotes.

In the sales spreadsheet shown below, for example, suppose you want to total up only the sales that are more than $100. The criteria range is C2 to C9, and the criteria is “greater than 100.” Since you’re adding up the values in that same cell range (C2 to C9), you don’t need to supply a separate sum range. So your formula is:

=SUMIF(C2:C9,”>100″)

Using the SUMIF function.

Shimon Brathwaite / Foundry

What if instead you want to find the total for all sales in the East region only? To do that you’ll have to specify both the criteria range (cells B2 to B9) and the sum range (cells C2 to C9). This is the formula:

=SUMIF(B2:B9,B2,C2:C9)

Note that you don’t have to type out “East” for the criteria. You can simply type B2 or click the cell B2 to have Excel search for the text it contains.

Using SUMIF with both a criteria range and a sum range.

Shimon Brathwaite / Foundry

There is a similar function called COUNTIF that lets you create a count of values that meet specified criteria. The syntax is as follows:

=COUNTIF(range,criteria)

So to count the total number of sales in the West region, for instance, you supply the range of cells to apply the criteria to (B2 to B9), followed by the criteria (“West” or cell B3). The formula is:

=COUNTIF(B2:B9,B3)

The COUNTIF function can instantly count up items that meet your criteria.

Shimon Brathwaite / Foundry

What if you want to apply multiple criteria to your data, such as calculating total sales for books in the East region, or counting the number of sales over $100 in the West region? Excel can do that too, via functions called SUMIFS and COUNTIFS. These functions use more complex syntax than SUMIF and COUNTIF. For more details, use cases, and best practices for all four of these functions, see Microsoft’s SUMIF, SUMIFS, COUNTIF, and COUNTIFS support pages.

4. The CONCAT function

This function is useful for piecing together text from different cells into one complete string. For instance, maybe you have a worksheet with different columns for people’s first and last names, but you want to put first and last names together. Other common use cases are completing an address, reference number, file path, or URL. The syntax is as follows:

=CONCAT(text1,text2,text3,…)

In this example we will use CONCAT to combine a list of first names and last names into a full name with a space in between. To do so we simply place the cursor in cell C2, type =CON and select CONCAT from the list of options that appears. Next, select the cell that contains the first name (A2) and add a comma, a blank space surrounded by quotation marks, and another comma. Then add the last name by selecting the adjacent cell (B2) and hit Enter. Here’s the full formula:

=CONCAT(A2,” “,B2)

Next, click and drag the bottom right of cell C2 to autofill the formula in all the other rows.

The CONCAT function combined the values from column A with those from column B.

Shimon Brathwaite / Foundry

For more details and examples, see Microsoft’s CONCAT function support page.

5. The VLOOKUP function

This is one of the most commonly used functions in Excel and a valuable data analysis tool. VLOOKUP lets you look up a value in a table and return information from other columns related to that value. It’s very useful for combining data from different lists or comparing two lists to find matching items. To use this function, you need to provide three to four pieces of information:

1. The value you want to look for. This is known as the lookup value.
2. The range of cells to look in. This is known as the table array.
3. The column that contains the information you want to return, called the column index number.
4. Optionally, the type of lookup you want to perform: TRUE or FALSE. This is known as the range lookup. FALSE means you want an exact match for the lookup value, while picking TRUE returns the best approximate match. If you don’t specify a range lookup, VLOOKUP defaults to TRUE.

The syntax is as follows:

=VLOOKUP(lookup_value,table_array,column_index_number,[range_lookup])

The lookup value must be in the first column of cells you specify in the table array. The leftmost column in the table array has a column index number of 1, with subsequent columns numbered 2, 3, and so on.

In this example, we’ll look up what region our employees work in. To do so, we first need to specify the value that we are going to search for: the employee name Mike (cell A2). Next, we need to highlight the entire cell range (table array) that we want to look in: cells F2:G8.

Then we specify which column holds the information that we want. Rather than picking the column itself, we count from left to right within the table array. Since the column that contains the region is the second from the left, type in 2.

Lastly, we enter the TRUE (best approximate match) or FALSE (exact match) option. TRUE is typically only used with numbers or when you aren’t sure if the value you want is in the table. Since we know the value we want is in the table, we will pick FALSE. Generally, FALSE is the better option, as it returns more accurate results.

This is the full formula:

=VLOOKUP(A2,F2:G8,2,FALSE)

Use VLOOKUP to find values linked to other values in large data sets.

Shimon Brathwaite / Foundry

This is an oversimplified example using a small data set, but when you need to search through a spreadsheet with thousands (or tens of thousands) of cells, VLOOKUP is a huge time-saver and reduces the possibility of errors. For more details and examples, see Microsoft’s VLOOKUP function support page.

Using Copilot to create Excel formulas

Remembering all these different formulas and functions can be difficult, so another good approach for those who have access to Copilot in Excel (see details at the top of this story) is to use the AI tool to dynamically create formulas that meet your needs. Using Copilot, you can describe what you want the formula to do  and have the AI generate the syntax for you. In this section we’ll demonstrate with two sample data sets.

To access Copilot, click the Copilot icon on your ribbon toolbar or at the bottom right of your screen.

To invoke Copilot, click its icon, which you might find in the ribbon toolbar up top or floating at the bottom right of your Excel window.

Shimon Brathwaite / Foundry

The Copilot sidebar opens along the right side of your spreadsheet. Type in a prompt describing the formula you want created — in this example, a simple formula that adds up all the sales values in the data set. Be sure to instruct Copilot where to place the formula:

Create a formula in B7 that sums all of the sales

Describe the formula you want Copilot to create.

Shimon Brathwaite / Foundry

Click the submit arrow in the chat window, and Copilot will create and insert the formula to give you the result you want.

Copilot has inserted the correct Excel formula for the sum operation, with the result appearing in cell B7.

Shimon Brathwaite / Foundry

That was a simple one. Now try a more advanced example working from a more complex data set, as shown in the screenshot below.

A more complex data set for our second example.

Shimon Brathwaite / Foundry

Try the following prompt:

Create a formula in G2 that calculates total revenue per unit while accounting for the discount on each purchase. Then fill the formula down for all rows and create a summary table showing total revenue by Region.

Click the submit arrow, and it should create something like this:

Copilot creates the requested formula and fills in the results.

Shimon Brathwaite / Foundry

(There’s a lot more you can have Copilot can do in Excel besides creating formulas. See our story “11 cool things Copilot can do in Excel.”)

Tip: It’s not required, but it can be helpful to format your spreadsheet as a table when working with Copilot. This helps define more clearly the exact cells for Copilot to use when it creates formulas or takes other actions on your data.

You’re just getting started

In this story you’ve seen how powerful formulas and functions can be in Excel — and we’ve only the scratched the surface of what they can do. Once you get comfortable using them, you can explore some of the myriad prebuilt functions Excel offers and learn how to build more complex formulas (including nesting functions). That’s all beyond the scope of this article, but a great place to start is Microsoft’s “Overview of formulas in Excel” support page, which includes links to several helpful tutorials.

For those who have access to it in Excel, experimenting with Copilot is another great way to learn. You can review the steps it took and the formulas it created to get a better understanding of how formulas work and what they’re capable of.

This story was originally published in May 2023 and updated in June 2026.

More Excel tips and how-tos:

A new update for Google Home could make it less likely your smart home cameras mistake you for someone else, just because you’re facing away from the camera. Starting June 23rd, Google’s expanding its facial recognition feature so that people you’ve tagged in your Familiar Faces library can continue to be identified when their faces aren’t clearly visible, using “additional non-biometric signals (body size, clothing color, etc.).”

The Familiar Faces library will also begin automatically updating with the most recent images of everyone in your house, so you should get fewer inaccurate notifications from outdated examples.

Google also says its AI-generated video event descriptions “can now identify specific sounds — like dogs barking, alarms, or footsteps” and include them in the notes, even if the audio came from something off camera. All of these updates could help address some of the quirks The Verge’s Jennifer Pattison Tuohy noticed while trying out Google’s updated smart home system last year, like event logs with detailed descriptions of people who weren’t there, and things that didn’t happen.

In the new technique, these cells are injected directly into the body along with hydrogel microspheres that help them stay together and form connections with nearby blood vessels. The spheres have special properties that allow them to act like a liquid when they are closely packed together, so they can be injected through a syringe and then regain their solid structure once inside the body. “These microspheres provide the hepatocytes with a niche where they can stay localized and become connected to the host circulation much faster,” says MIT postdoc Vardhman Kumar, lead author of a paper on the work.

The researchers have shown that the cells can remain viable in the bodies of mice for at least two months, generating many of the enzymes and other proteins that the liver produces. In the study, they were injected into the fat tissue of the belly, but they could also be delivered to other sites.

“The way we see this technology is it can provide an alternative to surgery, but it can also serve as a bridge to transplantation where these grafts can provide support until a donor organ becomes available,” Kumar says. While patients would likely need to take immunosuppressive drugs, the researchers are exploring how they might get the hepatocytes to evade the immune system or use the microspheres to deliver immunosuppressants locally. 

OpenAI has introduced spend controls and enhanced usage analytics for ChatGPT Enterprise to enable organizations to monitor AI adoption, track consumption across teams, and set budgets for AI usage. But, analysts cautioned, it still can’t show how those costs lead to business benefits.

The new features provide administrators with centralized dashboards showing how ChatGPT is being used across an organization, enabling them to understand adoption patterns, and manage AI costs by setting budgets and tracking spending.

“The Global Admin Console brings ChatGPT and Codex credit usage into one view, so admins can see a more granular breakdown of credit consumption across users, products, and models — helping them understand where spend is coming from and how it maps to actual credit usage,” OpenAI said.

The Nex Playground is the family-centric, Kinect-like game console that made one Verge editor’s kids laugh, cry, and ask for more playtime, even when they were sick. The motion-based game play isn’t perfect, but it has won over plenty of parents — and it’s on sale for $239 at Amazon, down from its current MSRP of $299. That’s a healthy discount, though only slightly better than the Nex’s original $250 price.

Like so many other gadgets, the Nex Playground became a victim of RAMageddon when its price went up in April. It got a $60 discount shortly after for Amazon’s week of gaming deals — and again with this Prime Day deal — but it’s nothing like the $199 prices we saw before the RAM crisis hit. This is the new normal, I guess.

Still, any discount is welcome, especially when you’ll need a subscription if you want to play more than a few starter games. You can get a 3-month Play Pass for $49 to access a much deeper catalog of games, or pay $89 for a 12-month pass. The subscription economy and RAMageddon come for us all, but at least a sale can take a little bit of the sting out of it.

According to one Windows user forum, the issue is particularly frustrating because of this lack of error message. As one user put it, “‘Word won’t open from our workpaper system’ is functionally the same as ‘Word is broken.’ To an administrator, the difference determines whether the next hour is spent repairing Office, rolling back Windows, calling a line-of-business vendor, or opening a Microsoft support case.”

Microsoft has said that it is aware of the issue and is working towards a resolution.

Deleting mystery file

The company is also looking to fix a minor issue that has emerged from this latest round of updates: Users are finding that when items are deleted, the confirmation dialog box displays the internal Recycle Bin file name (for example, $Rxxxxx.ext) instead of the original file name. However, Microsoft stressed that in the Recycle Bin, the item still appears with its original name.

The Foundation has established a set of criteria to demonstrate conformity with what is expected. There are two layers: the Requirements and Guidance layers will help users determine what is actually required, while the Assessment Enablement layer will look at how those requirements are evaluated.

Appia stressed that what it is offering are not standards — which are set by recognized international bodies such as ISO/IEC — but a means of assessing what those standards mean and how they can be used by organizations. However, the Foundation said that some of the criteria that it is introducing may become standards themselves after a period of time.

The Appia Foundation is hosted by the Linux Foundation’s Joint Development Foundation, and its other members include Arm, Ericsson, Mastercard, Mitsubishi Electric, Omron, Schneider Electric, and Siemens. It is also looking to bring academics and government into the fold, so that it can establish an advisory board.

The data collected included full prompts and transcriptions, private conversations, people and performance data, Wired said, adding, “Meta executives have repeatedly defended the data-gathering project, saying it was necessary to train AI systems to operate computer software the way humans do, and that employees were the best examples for the artificial intelligence to learn from.”

Wired also quoted Stephane Kasriel, a Meta vice president overseeing AI research, who saying that the company discovered that unauthorized employees were found to have accessed MCI data on June 18, and that the hole was closed “within four hours.” But, he added, “ the initial fix didn’t stick, and access to the data had to be further locked down.”

In an email statement, Meta confirmed that the program was being halted for the time being. “We have carefully designed this program with privacy safeguards, and while we have no indication at this time that any data was improperly accessed by Meta employees, we’re pausing it while we investigate,” Meta said. 

Enterprises implementing agentic AI face a challenge: Which tools should they allow their agents to use, where can they be found, and how can they be used safely? A new protocol, Agentic Resource Discovery, or ARD, aims to let agents answer those questions for themselves. Behind it are Google, Microsoft, Cisco, Nvidia, Salesforce and others.

ARD aims to standardize the way that tools and services are shared across systems within a corporate domain. For example, when investigating a production problem, an agent may want to query engineering documentation and open support tickets, deployment history and observability systems, all of which could be managed by different registries and across different silos. There is no common layer that pulls them together. ARD has been designed to be that layer.

It operates across two levels. Catalogs and Registries. In the first, an organization publishes a catalog setting out its available capabilities. The Registries layer act as a form of search engine, crawling those published catalogs.

Version 2603 (Build 19822.20168)

Release date: April 9, 2026

This build fixes several bugs, including one in Outlook in which users could not close the Copilot chat pane using a keyboard. Users can now close the pane by navigating to the Close button using a keyboard or by using the assigned keyboard shortcut.

Get more info about Version 2603 (Build 19822.20168).

Version 2603 (Build 19822.20142)

Release date: March 31, 2026

This build includes “various fixes to functionality and performance,” according to Microsoft.

Get more info about Version 2603 (Build 19822.20142).

Version 2603 (Build 19822.20114)

Release date: March 24, 2026

This build fixes a single bug in which PowerPoint sometimes closed unexpectedly when opening a newly created empty file from the OneDrive folder.

Get more info about Version 2603 (Build 19822.20114).

Version 2602 (Build 19725.20190)

Release date: March 18, 2026

This build fixes an Outlook bug in which updating a single instance of a recurring meeting in a Microsoft 365 group calendar updated the entire series.

Get more info about Version 2602 (Build 19725.20190).

Version 2602 (Build 19725.20172)

Release date: March 10, 2026

This build introduces agent mode in Word, which adds a conversational chat experience that helps create, edit, and refine document content as you work. In addition, the build fixes a bug that impacted the rendering of extended characters in calendar items, causing certain characters to appear as question marks.

The build also plugs a number of security holes. For details, see Release notes for Microsoft Office security updates.

What IT needs to know: Because this is a security update, it should be applied relatively soon. Over the next few weeks, check for reports about problematic issues, and if all seems well, apply the update.

Get more info about Version 2602 (Build 19725.20172).

Version 2602 (Build 19725.20152)

Release date: March 3, 2026

This build fixes a bug in which closing a document sometimes remained in progress indefinitely after the Office app resumed from sleep or hibernation.

Get more info about Version 2602 (Build 19725.20152).

Version 2602 (Build 19725.20126)

Release date: February 24, 2025

This build fixes several bugs, including one that caused OneNote to close unexpectedly upon startup.

Get more info about Version 2602 (Build 19725.20126).

Version 2601 (Build 19628.20214)

Release date: February 17, 2025

This build includes, in Microsoft’s words, “various fixes to functionality and performance.”

Get more info about Version 2601 (Build 19628.20214).

Version 2601 (Build 19628.20204)

Release date: February 10, 2026

This build fixes a bug that sometimes prevented users from opening emails with the Encrypt Only label in Outlook.

It also plugs a number of security holes. For details, see Release notes for Microsoft Office security updates.

What IT needs to know: Because this is a security update, it should be applied relatively soon. Over the next few weeks, check for reports about problematic issues, and if all seems well, apply the update.

Get more info about Version 2601 (Build 19628.20204).

Version 2601 (Build 19628.20166)

Release date: February 3, 2026

This build includes, in Microsoft’s words, “various fixes to functionality and performance.”

Get more info about Version 2601 (Build 19628.20166).

Version 2601 (Build 19628.20150)

Release date: January 27, 2026