verge-rss

Image: Cath Virginia / The Verge, Getty Images

The Paris 2024 Olympics are nearly here. The opening ceremony starts on July 26th, and this year’s events run through August 11th. On the main NBC channel, you can see hours of live morning and afternoon coverage of events like swimming and gymnastics ahead of a three-hour primetime recap, to go along with livestreams of every sport and event and full event replays on Peacock.
One key aspect of NBC’s Peacock coverage will be an Olympics hub, which spotlights major events happening live, lets you browse by sport and by “star athlete,” offers an interactive schedule, and includes an up-to-date count of medal standings.

Image: NBC
Peacock’s Olympics hub.

NBC is also offering a way to catch up on the events of the previous day on Peacock — and, because it’s 2024, it of course involves AI. “Your Daily Olympic Recap,” a personalized daily recap, will be narrated by an AI-generated version of legendary sportscaster Al Michaels’ voice. Yes, really — I have some more information on that in another article.
Peacock will also have a new feature called Peacock Live Actions, which NBC describes as “a new interactive tool that lets fans choose their own viewing journey during live and primetime coverage.”
As an example, NBC says that while watching its “Gold Zone” whiparound show that serves as an Olympics-focused version of NFL RedZone and even features RedZone host Scott Hanson, you can choose to keep watching a specific live feed shown during the show instead of being shuffled to whatever Gold Zone covers next.

Image: NBC
Peacock’s Discovery Multiview.

If you want to watch a bunch of events on your TV, phone, tablet, or computer all at the same time, a special Peacock Discovery Multiview will include onscreen information like what’s at stake for a particular competitor or when the person onscreen is a first-time Olympian. The feature means you can keep up with the action even while your audio or captions are tuned in to a different feed. There will also be a “traditional” multiview that NBC says will be available for certain sports, including soccer, track and field, and wrestling. Both multiviews will let you watch up to four events at once.
Disclosure: Comcast, which owns NBCUniversal, is also an investor in Vox Media, The Verge’s parent company.

Image: Cath Virginia / The Verge, Getty Images

The Paris 2024 Olympics are nearly here. The opening ceremony starts on July 26th, and this year’s events run through August 11th. On the main NBC channel, you can see hours of live morning and afternoon coverage of events like swimming and gymnastics ahead of a three-hour primetime recap, to go along with livestreams of every sport and event and full event replays on Peacock.

One key aspect of NBC’s Peacock coverage will be an Olympics hub, which spotlights major events happening live, lets you browse by sport and by “star athlete,” offers an interactive schedule, and includes an up-to-date count of medal standings.

Image: NBC
Peacock’s Olympics hub.

NBC is also offering a way to catch up on the events of the previous day on Peacock — and, because it’s 2024, it of course involves AI. “Your Daily Olympic Recap,” a personalized daily recap, will be narrated by an AI-generated version of legendary sportscaster Al Michaels’ voice. Yes, really — I have some more information on that in another article.

Peacock will also have a new feature called Peacock Live Actions, which NBC describes as “a new interactive tool that lets fans choose their own viewing journey during live and primetime coverage.”

As an example, NBC says that while watching its “Gold Zone” whiparound show that serves as an Olympics-focused version of NFL RedZone and even features RedZone host Scott Hanson, you can choose to keep watching a specific live feed shown during the show instead of being shuffled to whatever Gold Zone covers next.

Image: NBC
Peacock’s Discovery Multiview.

If you want to watch a bunch of events on your TV, phone, tablet, or computer all at the same time, a special Peacock Discovery Multiview will include onscreen information like what’s at stake for a particular competitor or when the person onscreen is a first-time Olympian. The feature means you can keep up with the action even while your audio or captions are tuned in to a different feed. There will also be a “traditional” multiview that NBC says will be available for certain sports, including soccer, track and field, and wrestling. Both multiviews will let you watch up to four events at once.

Disclosure: Comcast, which owns NBCUniversal, is also an investor in Vox Media, The Verge’s parent company.

Read More 

Photo by Cooper Neill / Getty Images

Legendary sportscaster Al Michaels is going to give daily, personalized recaps of the Paris Olympics on Peacock — well, an AI-generated Al Michaels voice will. In practice, the effect is a lot like hearing a sports announcer’s voice in a video game like Madden, except it’s spitting out lines about real-life sports, which, in this case, means custom Olympics coverage.
Here’s how it works. To set up what NBC is calling “Your Daily Olympic Recap” in the Peacock app, you’ll provide your name (the AI voice can welcome the “majority” of people by their first name, NBC says in a press release) and pick up to three types of sports that are interesting to you and up to two types of highlights (for example, “Top Competition” or “Viral & Trending Moments”). Then, each morning, you’ll get your Michaels-led rundown.

Here’s an example of Peacock’s Olympics recap led by the AI-generated Al Michaels voice.

To help protect against potential AI-made weirdness, NBC says that “a team of NBCU editors will review all content, including audio and clips, for quality assurance and accuracy before recaps are made available to users.” But I still feel like there’s the chance somebody’s recap will include an AI-generated hallucination spoken out loud in Al Michaels’ voice, like highlighting the wrong athlete or bungling some unusual outcome in a sport.
The voice was trained using Michaels’ appearances on NBC, according to the press release, and the experience was built in-house, NBCUniversal’s John Jelley tells The Verge in a statement. “Our in-house Peacock team of engineers, product managers and data scientists developed a proprietary process to integrate, optimize and validate state-of-the-art large language model and voice synthesis technology to create this experience.”

Image: NBC

In the press demo where I heard the voice, it sounded convincing, but that’s what you’d expect from a demo. The real test will be when it’s generating millions of unique clips — NBC estimates there could be nearly 7 million personalized variants in the US during the games — crossing dozens of sports, each with its own unique terminology, and identifying an array of athletes from around the globe.
Peacock’s recaps led by the AI Al Michaels will be available starting July 27th in supported browsers and the iOS and Android Peacock apps. The first edition of the recap will feature highlights from the opening ceremony for everyone, and the personalized recaps will start on July 28th.
Disclosure: Comcast, which owns NBCUniversal, is also an investor in Vox Media, The Verge’s parent company.

Photo by Cooper Neill / Getty Images

Legendary sportscaster Al Michaels is going to give daily, personalized recaps of the Paris Olympics on Peacock — well, an AI-generated Al Michaels voice will. In practice, the effect is a lot like hearing a sports announcer’s voice in a video game like Madden, except it’s spitting out lines about real-life sports, which, in this case, means custom Olympics coverage.

Here’s how it works. To set up what NBC is calling “Your Daily Olympic Recap” in the Peacock app, you’ll provide your name (the AI voice can welcome the “majority” of people by their first name, NBC says in a press release) and pick up to three types of sports that are interesting to you and up to two types of highlights (for example, “Top Competition” or “Viral & Trending Moments”). Then, each morning, you’ll get your Michaels-led rundown.

To help protect against potential AI-made weirdness, NBC says that “a team of NBCU editors will review all content, including audio and clips, for quality assurance and accuracy before recaps are made available to users.” But I still feel like there’s the chance somebody’s recap will include an AI-generated hallucination spoken out loud in Al Michaels’ voice, like highlighting the wrong athlete or bungling some unusual outcome in a sport.

The voice was trained using Michaels’ appearances on NBC, according to the press release, and the experience was built in-house, NBCUniversal’s John Jelley tells The Verge in a statement. “Our in-house Peacock team of engineers, product managers and data scientists developed a proprietary process to integrate, optimize and validate state-of-the-art large language model and voice synthesis technology to create this experience.”

Image: NBC

In the press demo where I heard the voice, it sounded convincing, but that’s what you’d expect from a demo. The real test will be when it’s generating millions of unique clips — NBC estimates there could be nearly 7 million personalized variants in the US during the games — crossing dozens of sports, each with its own unique terminology, and identifying an array of athletes from around the globe.

Peacock’s recaps led by the AI Al Michaels will be available starting July 27th in supported browsers and the iOS and Android Peacock apps. The first edition of the recap will feature highlights from the opening ceremony for everyone, and the personalized recaps will start on July 28th.

Disclosure: Comcast, which owns NBCUniversal, is also an investor in Vox Media, The Verge’s parent company.

Read More 

Image: Figma

Figma is announcing a bunch of new features at its Config conference today, including a major UI redesign, new generative AI tools to help people more easily make projects, and built-in slideshow functionality.
Let’s start with the redesign, which is intended to “lay the foundation for the next decade,” according to a blog post. You’ll see things like a new toolbar, rounded corners, and 200 new icons. As part of the design refresh, the company wants to “focus the canvas less on our UI and more on your work” and make something that’s approachable to new users while still being useful to Figma experts.

Image: Figma
Figma’s “UI3.”

Figma says this is the company’s third “significant redesign” since Figma’s closed beta launch. The new look is rolling out as part of a limited beta, and users can join a waitlist if they want to try it out.
Beyond the redesign, the headline feature addition is new generative AI tools, which look like a useful way to quickly get started with a design. They’re basically a Figma-focused version of the “draft an email”-type AI tools we’ve seen many times.
In a briefing, Figma chief product officer Yuhki Yamashita showed me an example of how Figma could create an app design for a new restaurant. A few seconds after he typed the prompt into a textbox, Figma mocked up an app with menu listings, a tab bar, and even buttons for delivery partners like Uber Eats and DoorDash. It looked like a generic mobile app mock-up, but Yamashita was able to start tweaking it right away.
In another example, Yamashita asked Figma AI to spin up a design for a recipe page for chocolate chip cookies, and sure enough, it did — including an AI-generated image of a cookie. Over Zoom, it looked like a pretty accurate image, but I can’t imagine that a basic image of a chocolate chip cookie is hard for an AI generator to make.

Figma AI can also build a prototype app by connecting design pages and writing suggested content, among other things.

Figma is also introducing AI features that could help speed up small tasks in big ways, such as “AI-enhanced” asset search and auto-generated text in designs instead of generic Lorem ipsum placeholder text.
Ideally, all of the new Figma AI tools will allow people who are newer to Figma to test ideas more easily while letting those who are more well versed in the app iterate more quickly, according to Yamashita. “We’re using AI to lower the floor and raise the ceiling,” Yamashita says in an interview with The Verge — something CEO Dylan Field has said to The Verge as well.
Figma AI is launching in a limited beta beginning on Wednesday, and interested users can get on the waitlist. Figma says the beta period will run through the end of the year. While in beta, Figma’s AI tools will be free, but the company says it might have to introduce “usage limits.” Figma is also promising “clear guidance on pricing” when the AI features officially launch.
In a blog post, Figma also spelled out its approach to training its AI models. “All of the generative features we’re launching today are powered by third-party, out-of-the-box AI models and were not trained on private Figma files or customer data,” writes Kris Rasmussen, Figma’s CTO. “We fine-tuned visual and asset search with images of user interfaces from public, free Community files.”
Rasmussen adds that Figma trains its models so they learn patterns and “Figma-specific concepts and tools” but not from users’ content. Figma is also going to let Figma admins control whether Figma can train on “customer content,” which includes “file content created in or uploaded to Figma by a user, such as layer names and properties, text and images, comments, and annotations,” according to Rasmussen.
Figma won’t start training on this content until August 15th; however, you should know that Starter and Professional plans are by default opted in to share this data, while Organization and Enterprise plans are opted out.
The company is likely being specific about how it trains its AI models because of Adobe’s recent terms of service disaster, where the company had to clarify that it wouldn’t train AI on your work.
In addition to the redesign and the new AI features, Figma is adding a potentially very practical new tool: Figma Slides, a Google Slides-like feature built right into Figma. Yamashita says that users have already been hacking Figma to find a way to make slides, so now there’s an official method to build and share presentations right inside the app.

Image: Figma

There are a few Figma-specific features that designers will likely appreciate. You’ll be able to tweak designs you’ve included in the deck in real time using Figma’s tools. (Note that those changes will only appear in the deck — tweaks won’t currently sync back to the original design files, though Yamashita says that Figma wants to make that possible eventually.)
You can also present an app prototype right from the deck, meaning you don’t need to make a convoluted screen recording just to demonstrate how one piece connects to another. You can also add interactive features for audience members, like a poll or an alignment scale, where people can plot on a range if they agree or disagree with something.
Figma Slides will be available in open beta beginning on Wednesday. It will be free while in beta but will become a paid feature when it officially launches. The company is also adding new features for its developer mode in Figma, including a “ready for dev” task list.
This year’s Config is the first since Adobe abandoned its planned $20 billion acquisition of Figma following regulatory scrutiny. With the dissolution of the merger, Adobe was forced to pay Figma a $1 billion breakup fee.

Image: Figma

Figma is announcing a bunch of new features at its Config conference today, including a major UI redesign, new generative AI tools to help people more easily make projects, and built-in slideshow functionality.

Let’s start with the redesign, which is intended to “lay the foundation for the next decade,” according to a blog post. You’ll see things like a new toolbar, rounded corners, and 200 new icons. As part of the design refresh, the company wants to “focus the canvas less on our UI and more on your work” and make something that’s approachable to new users while still being useful to Figma experts.

Image: Figma
Figma’s “UI3.”

Figma says this is the company’s third “significant redesign” since Figma’s closed beta launch. The new look is rolling out as part of a limited beta, and users can join a waitlist if they want to try it out.

Beyond the redesign, the headline feature addition is new generative AI tools, which look like a useful way to quickly get started with a design. They’re basically a Figma-focused version of the “draft an email”-type AI tools we’ve seen many times.

In a briefing, Figma chief product officer Yuhki Yamashita showed me an example of how Figma could create an app design for a new restaurant. A few seconds after he typed the prompt into a textbox, Figma mocked up an app with menu listings, a tab bar, and even buttons for delivery partners like Uber Eats and DoorDash. It looked like a generic mobile app mock-up, but Yamashita was able to start tweaking it right away.

In another example, Yamashita asked Figma AI to spin up a design for a recipe page for chocolate chip cookies, and sure enough, it did — including an AI-generated image of a cookie. Over Zoom, it looked like a pretty accurate image, but I can’t imagine that a basic image of a chocolate chip cookie is hard for an AI generator to make.

Figma is also introducing AI features that could help speed up small tasks in big ways, such as “AI-enhanced” asset search and auto-generated text in designs instead of generic Lorem ipsum placeholder text.

Ideally, all of the new Figma AI tools will allow people who are newer to Figma to test ideas more easily while letting those who are more well versed in the app iterate more quickly, according to Yamashita. “We’re using AI to lower the floor and raise the ceiling,” Yamashita says in an interview with The Verge — something CEO Dylan Field has said to The Verge as well.

Figma AI is launching in a limited beta beginning on Wednesday, and interested users can get on the waitlist. Figma says the beta period will run through the end of the year. While in beta, Figma’s AI tools will be free, but the company says it might have to introduce “usage limits.” Figma is also promising “clear guidance on pricing” when the AI features officially launch.

In a blog post, Figma also spelled out its approach to training its AI models. “All of the generative features we’re launching today are powered by third-party, out-of-the-box AI models and were not trained on private Figma files or customer data,” writes Kris Rasmussen, Figma’s CTO. “We fine-tuned visual and asset search with images of user interfaces from public, free Community files.”

Rasmussen adds that Figma trains its models so they learn patterns and “Figma-specific concepts and tools” but not from users’ content. Figma is also going to let Figma admins control whether Figma can train on “customer content,” which includes “file content created in or uploaded to Figma by a user, such as layer names and properties, text and images, comments, and annotations,” according to Rasmussen.

Figma won’t start training on this content until August 15th; however, you should know that Starter and Professional plans are by default opted in to share this data, while Organization and Enterprise plans are opted out.

The company is likely being specific about how it trains its AI models because of Adobe’s recent terms of service disaster, where the company had to clarify that it wouldn’t train AI on your work.

In addition to the redesign and the new AI features, Figma is adding a potentially very practical new tool: Figma Slides, a Google Slides-like feature built right into Figma. Yamashita says that users have already been hacking Figma to find a way to make slides, so now there’s an official method to build and share presentations right inside the app.

Image: Figma

There are a few Figma-specific features that designers will likely appreciate. You’ll be able to tweak designs you’ve included in the deck in real time using Figma’s tools. (Note that those changes will only appear in the deck — tweaks won’t currently sync back to the original design files, though Yamashita says that Figma wants to make that possible eventually.)

You can also present an app prototype right from the deck, meaning you don’t need to make a convoluted screen recording just to demonstrate how one piece connects to another. You can also add interactive features for audience members, like a poll or an alignment scale, where people can plot on a range if they agree or disagree with something.

Figma Slides will be available in open beta beginning on Wednesday. It will be free while in beta but will become a paid feature when it officially launches. The company is also adding new features for its developer mode in Figma, including a “ready for dev” task list.

This year’s Config is the first since Adobe abandoned its planned $20 billion acquisition of Figma following regulatory scrutiny. With the dissolution of the merger, Adobe was forced to pay Figma a $1 billion breakup fee.

Read More 

Photo by Daniel Golson for The Verge

I’ll admit, I was too distracted by Rivian’s surprise announcement of a $5 billion deal with Volkswagen to develop software and electric vehicles to notice that the buzzy California company was also planning an investor day event on June 27th.
So when I woke up on Wednesday to see this shareholder letter — thanks to InsideEVs with an assist from BloombergNEF’s Corey Cantor on Threads — I was pretty surprised to see Rivian was essentially taking a page from Tesla by teasing us with some shroud-covered mystery vehicles.

Screenshot: Rivian

It’s already been a pretty jam-packed 2024 for Rivian: the introduction of the more affordable R2, starting at $45,000; the surprise reveal of the R3 and R3X, which briefly broke the internet with hype; the release of its second-generation R1T and R1S vehicles, with all new architecture, motors, and batteries; and of course, the company’s continuing financial struggles in the face of softening EV demand.
So to toss into the mix four — four! — new mystery vehicles is either a pretty bold strategy for the future or a desperate effort to stay relevant amid rising costs and cooling interest in electric vehicles. Either way, let’s take a look at these vehicles to see if we can make some guesses about what they are.
It’s already been a pretty jam-packed 2024 for Rivian
First of all, Rivian’s shroud-covered vehicles look very similar to the ones Tesla revealed during its company shareholder meeting earlier this month, which raises the possibility that both are just using the same tools in Photoshop to create these images.
Starting from the left, we have Rivian’s OG vehicles: the R1T truck, R1S SUV, and EDV (Electric Delivery Van) for Amazon. In the next column, we have the second-generation R1 vehicles that were announced earlier this month. On top of those, there’s our first mystery vehicle, which suggests the company could be working on an updated EDV built on the Gen 2 architecture.
Except that doesn’t really look like a van shape. It’s definitely more truck-shaped — although InsideEVs editor-in-chief (and former Verge contributor) Patrick George makes a convincing argument that we shouldn’t read too much into the various shapes. They could be stock images or just placeholders. If anything, Rivian could actively be trying to throw sleuths off the scent with these shapes. (Rivian spokesperson Liz Markman declined to comment on the mystery vehicles.)
If anything, Rivian could actively be trying to throw sleuths off the scent
In the next column, we can see the R2 and R3, both of which were unveiled in March. There’s another mystery vehicle on top of that column. And all three vehicles will be built on an electrical architecture called “MSP,” which stands for “midsized platform.”
And then there’s the fourth column with three shrouded vehicles built on a platform simply called “Affordable Mass-Market.” How intriguing! Earlier this year, the R2 and R3 were billed as Rivian’s big play for affordable EVs, with the former starting at $45,000. Rivian hasn’t put a price on the R3 or the performance trim R3X except to say it will be less than the R2.
So what are we to make of all of this? It seems likely that a few of these mystery vehicles will be trucks. Trucks are some of the bestselling vehicles in North America. Rivian’s first vehicle was the R1T. Trucks are in the company’s DNA. So does that mean we’re going to get an R2T or R3T or perhaps even X variants of both? It’s possible!
While the rest of the auto industry is prioritizing big, heavy electric trucks that sell for six figures, Rivian could shake things up with a cheaper, smaller truck. The company didn’t mention whether it was working on such a vehicle during the R2 announcement in March. And none of these silhouettes are instantly screaming “truck,” so it’s unclear whether that’s actually in the cards.
Trucks are in the company’s DNA
Another thing to note is that the affordable mass-market vehicles are all different shapes. I know, we can’t trust the shapes! But this could still mean that there are at least small, medium, and large vehicles in the works for that category. That seems important.
The fact that Rivian is working on a trio of vehicles that are even cheaper than the R2 and R3 is encouraging. Other automakers have stumbled in their attempts to produce truly affordable EVs. Tesla has gone back and forth on whether it will make a $25,000 “Model 2”-type vehicle. The Volvo EX30 is getting caught up in the trade war with China and likely won’t sell for its promised $35,000. Chevy is only now rolling out its Equinox EV but hasn’t yet offered the $35,000 version that was announced.
If Rivian joins the fray, that would be interesting. The company clearly doesn’t want to get pegged as an automaker that exclusively makes luxury vehicles on par with Land Rover or Mercedes-Benz, churning out expensive trucks and SUVs for the superrich. Still, it will need to break out of the “EV valley of death” in which its costs exceed its revenue and start showing signs of long-term survivability before it can tackle that project. The lifeline from Volkswagen — $1 billion right away, with an additional $4 billion to follow — will certainly help in that effort.
I like that despite its financial woes, Rivian is still playing the long game. Maybe soon enough we’ll get a peek at what’s under those shrouds.

Photo by Daniel Golson for The Verge

I’ll admit, I was too distracted by Rivian’s surprise announcement of a $5 billion deal with Volkswagen to develop software and electric vehicles to notice that the buzzy California company was also planning an investor day event on June 27th.

So when I woke up on Wednesday to see this shareholder letter — thanks to InsideEVs with an assist from BloombergNEF’s Corey Cantor on Threads — I was pretty surprised to see Rivian was essentially taking a page from Tesla by teasing us with some shroud-covered mystery vehicles.

Screenshot: Rivian

It’s already been a pretty jam-packed 2024 for Rivian: the introduction of the more affordable R2, starting at $45,000; the surprise reveal of the R3 and R3X, which briefly broke the internet with hype; the release of its second-generation R1T and R1S vehicles, with all new architecture, motors, and batteries; and of course, the company’s continuing financial struggles in the face of softening EV demand.

So to toss into the mix four — four! — new mystery vehicles is either a pretty bold strategy for the future or a desperate effort to stay relevant amid rising costs and cooling interest in electric vehicles. Either way, let’s take a look at these vehicles to see if we can make some guesses about what they are.

First of all, Rivian’s shroud-covered vehicles look very similar to the ones Tesla revealed during its company shareholder meeting earlier this month, which raises the possibility that both are just using the same tools in Photoshop to create these images.

Starting from the left, we have Rivian’s OG vehicles: the R1T truck, R1S SUV, and EDV (Electric Delivery Van) for Amazon. In the next column, we have the second-generation R1 vehicles that were announced earlier this month. On top of those, there’s our first mystery vehicle, which suggests the company could be working on an updated EDV built on the Gen 2 architecture.

Except that doesn’t really look like a van shape. It’s definitely more truck-shaped — although InsideEVs editor-in-chief (and former Verge contributor) Patrick George makes a convincing argument that we shouldn’t read too much into the various shapes. They could be stock images or just placeholders. If anything, Rivian could actively be trying to throw sleuths off the scent with these shapes. (Rivian spokesperson Liz Markman declined to comment on the mystery vehicles.)

In the next column, we can see the R2 and R3, both of which were unveiled in March. There’s another mystery vehicle on top of that column. And all three vehicles will be built on an electrical architecture called “MSP,” which stands for “midsized platform.”

And then there’s the fourth column with three shrouded vehicles built on a platform simply called “Affordable Mass-Market.” How intriguing! Earlier this year, the R2 and R3 were billed as Rivian’s big play for affordable EVs, with the former starting at $45,000. Rivian hasn’t put a price on the R3 or the performance trim R3X except to say it will be less than the R2.

So what are we to make of all of this? It seems likely that a few of these mystery vehicles will be trucks. Trucks are some of the bestselling vehicles in North America. Rivian’s first vehicle was the R1T. Trucks are in the company’s DNA. So does that mean we’re going to get an R2T or R3T or perhaps even X variants of both? It’s possible!

While the rest of the auto industry is prioritizing big, heavy electric trucks that sell for six figures, Rivian could shake things up with a cheaper, smaller truck. The company didn’t mention whether it was working on such a vehicle during the R2 announcement in March. And none of these silhouettes are instantly screaming “truck,” so it’s unclear whether that’s actually in the cards.

Another thing to note is that the affordable mass-market vehicles are all different shapes. I know, we can’t trust the shapes! But this could still mean that there are at least small, medium, and large vehicles in the works for that category. That seems important.

The fact that Rivian is working on a trio of vehicles that are even cheaper than the R2 and R3 is encouraging. Other automakers have stumbled in their attempts to produce truly affordable EVs. Tesla has gone back and forth on whether it will make a $25,000 “Model 2”-type vehicle. The Volvo EX30 is getting caught up in the trade war with China and likely won’t sell for its promised $35,000. Chevy is only now rolling out its Equinox EV but hasn’t yet offered the $35,000 version that was announced.

If Rivian joins the fray, that would be interesting. The company clearly doesn’t want to get pegged as an automaker that exclusively makes luxury vehicles on par with Land Rover or Mercedes-Benz, churning out expensive trucks and SUVs for the superrich. Still, it will need to break out of the “EV valley of death” in which its costs exceed its revenue and start showing signs of long-term survivability before it can tackle that project. The lifeline from Volkswagen — $1 billion right away, with an additional $4 billion to follow — will certainly help in that effort.

I like that despite its financial woes, Rivian is still playing the long game. Maybe soon enough we’ll get a peek at what’s under those shrouds.

Read More 

The excellent M2-powered Mac Mini is up to $110 off. | Photo by Amelia Holowaty Krales / The Verge

Apple may have released a bunch of new and improved laptops over the last few months, but the Mac Mini remains a solid desktop machine if you just need a basic computer for your home. And right now, Amazon is selling the latest M2-powered Mac Mini with the upgraded 512GB of storage for $689.99 ($110 off), which is just $10 shy of its all-time low price. You can also still buy the base model for $499 ($100 off) with a 256GB SSD and 8GB of RAM from Amazon, Best Buy, and B&H Photo.

If you’re looking for a desktop that’s more than capable of powering through everyday tasks, the Mac Mini is a good option. It’s a fast performer that’s even capable of some light video work, but it also offers a good port selection — including Thunderbolt 4 / USB-C and USB-A ports, an ethernet port, and an HDMI port. Add in futureproof specs like Wi-Fi 6E along with the old yet beloved 3.5mm headphone jack, and all in all, the desktop machine is a solid upgrade over its predecessor. Just bear in mind you’ll have to supply your own monitor, keyboard, and mouse, though you could easily buy those with the $110 you’re saving.

Read our M2-powered Mac Mini review.

A few more ways to save

Amazon Prime Day may not start until July 16th, but that isn’t stopping Amazon from releasing some early deals. We’re currently rounding up a guide to the best, so stay tuned — but in the meantime, one of my favorite deals is on the battery-powered Blink Video Doorbell with the Blink Outdoor 4 camera. Right now, you can buy both for $59.99 ($100 off) from Amazon, which is a new all-time low price. The smart security gadgets feature all the basics, like motion detection, night vision, and two-way audio, though they’re limited to 1080p video, and you have to pay extra for some features, like cloud-based storage.
The Meross Smart Wi-Fi Garage Door Opener is currently on sale for $39.99 ($20 off) at Amazon. The smart garage door device lets you control your garage door remotely via the Meross app and even sends notifications if you’ve accidentally left your door open. It works with a wide range of smart home platforms, including Apple HomeKit (and CarPlay), Amazon Alexa, and Google Assistant.
If you’re planning on traveling somewhere remote this summer, right now, you can save $300 when you buy the EcoFlow Glacier portable refrigerator for $799 directly from EcoFlow. The battery-powered EcoFlow Glacier combines a refrigerator, freezer, and ice maker into one. After some heavy usage, we found it easily lasts about a day and were impressed by how well the three-in-one appliance works, cooling down incredibly fast and producing 18 ice cubes in about 18 minutes. Read our review.
The OnePlus Nord N30 phone is on sale for $199.99 ($100 off) at Woot. For the price, the budget-friendly smartphone offers some standout features, including a 6.7-inch 120Hz display, all-day battery life, and 50W fast charging. It even comes equipped with a 3.5mm port for wired earbuds, and you can expand storage thanks to its microSD card slot. Just be aware you’ll have to compromise on camera quality, which is decent but can’t compete with midrange phone Google Pixel phones, and that the phone will only receive one update after Android 13.

The excellent M2-powered Mac Mini is up to $110 off. | Photo by Amelia Holowaty Krales / The Verge

Apple may have released a bunch of new and improved laptops over the last few months, but the Mac Mini remains a solid desktop machine if you just need a basic computer for your home. And right now, Amazon is selling the latest M2-powered Mac Mini with the upgraded 512GB of storage for $689.99 ($110 off), which is just $10 shy of its all-time low price. You can also still buy the base model for $499 ($100 off) with a 256GB SSD and 8GB of RAM from Amazon, Best Buy, and B&H Photo.

If you’re looking for a desktop that’s more than capable of powering through everyday tasks, the Mac Mini is a good option. It’s a fast performer that’s even capable of some light video work, but it also offers a good port selection — including Thunderbolt 4 / USB-C and USB-A ports, an ethernet port, and an HDMI port. Add in futureproof specs like Wi-Fi 6E along with the old yet beloved 3.5mm headphone jack, and all in all, the desktop machine is a solid upgrade over its predecessor. Just bear in mind you’ll have to supply your own monitor, keyboard, and mouse, though you could easily buy those with the $110 you’re saving.

Read our M2-powered Mac Mini review.

A few more ways to save

Amazon Prime Day may not start until July 16th, but that isn’t stopping Amazon from releasing some early deals. We’re currently rounding up a guide to the best, so stay tuned — but in the meantime, one of my favorite deals is on the battery-powered Blink Video Doorbell with the Blink Outdoor 4 camera. Right now, you can buy both for $59.99 ($100 off) from Amazon, which is a new all-time low price. The smart security gadgets feature all the basics, like motion detection, night vision, and two-way audio, though they’re limited to 1080p video, and you have to pay extra for some features, like cloud-based storage.
The Meross Smart Wi-Fi Garage Door Opener is currently on sale for $39.99 ($20 off) at Amazon. The smart garage door device lets you control your garage door remotely via the Meross app and even sends notifications if you’ve accidentally left your door open. It works with a wide range of smart home platforms, including Apple HomeKit (and CarPlay), Amazon Alexa, and Google Assistant.
If you’re planning on traveling somewhere remote this summer, right now, you can save $300 when you buy the EcoFlow Glacier portable refrigerator for $799 directly from EcoFlow. The battery-powered EcoFlow Glacier combines a refrigerator, freezer, and ice maker into one. After some heavy usage, we found it easily lasts about a day and were impressed by how well the three-in-one appliance works, cooling down incredibly fast and producing 18 ice cubes in about 18 minutes. Read our review.
The OnePlus Nord N30 phone is on sale for $199.99 ($100 off) at Woot. For the price, the budget-friendly smartphone offers some standout features, including a 6.7-inch 120Hz display, all-day battery life, and 50W fast charging. It even comes equipped with a 3.5mm port for wired earbuds, and you can expand storage thanks to its microSD card slot. Just be aware you’ll have to compromise on camera quality, which is decent but can’t compete with midrange phone Google Pixel phones, and that the phone will only receive one update after Android 13.

Read More 

Image: Verizon

Verizon has ditched its checkmark logo for a red and yellow “V” and, with the change, introduced a new commercial that revisits its “Can you hear me now?” tagline. It also announced updated home internet plans, now called “myHome,” that include a slew of service bundles like Disney Plus and Netflix as well as cloud storage options or a Walmart Plus subscription.
In the ad spot, that “now” turns into a lot more than just talking, with streaming video, filming TikToks, gaming, and FaceTiming shown to hammer home the point that this connection does a lot more than it did in 2002 when the campaign started. (That Paul Marcarelli, the original Verizon guy, defected to Sprint in 2016 goes unmentioned.)

Here’s the old logo and the new one in case you want to play a little game of “spot the differences.”

Image: Verizon
So long checkmark!

Meanwhile, in its MyHome offers, the services are $10 per month each so far, just like its myPlan offerings. Later this summer, the company says it’ll start offering Apple One bundles with iCloud Plus, Apple Arcade, Apple TV Plus, and Apple Music, too. (It didn’t announce pricing for that, however.) It also has its own cloud storage offering at the $10 monthly price point.
The internet plans, which will be 5G or LTE fixed wireless plans for most, start at $35.99 per month if you sign up for its phone plans and autopay. If you’re lucky, 5G plans could be as fast as 300Mbps (or up to a gigabit with its 5G Home Plus service). Verizon says its LTE plans get up to 50Mbps downloads. That pricing also applies to its Fios plans.

Image: Verizon
Verizon’s new myHome internet plans.

Image: Verizon

Verizon has ditched its checkmark logo for a red and yellow “V” and, with the change, introduced a new commercial that revisits its “Can you hear me now?” tagline. It also announced updated home internet plans, now called “myHome,” that include a slew of service bundles like Disney Plus and Netflix as well as cloud storage options or a Walmart Plus subscription.

In the ad spot, that “now” turns into a lot more than just talking, with streaming video, filming TikToks, gaming, and FaceTiming shown to hammer home the point that this connection does a lot more than it did in 2002 when the campaign started. (That Paul Marcarelli, the original Verizon guy, defected to Sprint in 2016 goes unmentioned.)

Here’s the old logo and the new one in case you want to play a little game of “spot the differences.”

Image: Verizon
So long checkmark!

Meanwhile, in its MyHome offers, the services are $10 per month each so far, just like its myPlan offerings. Later this summer, the company says it’ll start offering Apple One bundles with iCloud Plus, Apple Arcade, Apple TV Plus, and Apple Music, too. (It didn’t announce pricing for that, however.) It also has its own cloud storage offering at the $10 monthly price point.

The internet plans, which will be 5G or LTE fixed wireless plans for most, start at $35.99 per month if you sign up for its phone plans and autopay. If you’re lucky, 5G plans could be as fast as 300Mbps (or up to a gigabit with its 5G Home Plus service). Verizon says its LTE plans get up to 50Mbps downloads. That pricing also applies to its Fios plans.

Image: Verizon
Verizon’s new myHome internet plans.

Read More 

Google is addressing issues with its Find My Device network to help it actually find devices. | Image: Google

Google launched its upgraded Find My Device network in April, giving Android users the equivalent of Apple’s Find My network for locating trackers and offline devices. But some recent reviews have raised concerns over its ability to quickly and reliably find devices and trackers. Google says that it’s aware of the issue and plans to address it “in the coming weeks.”
In a statement provided to Mishaal Rahman, a senior contributor to sites like Android Authority and Android Police, which they shared on X, Google acknowledged the concerns over its Find My Device network. “We are actively working to roll out enhancements to how the Find My Device network operates that will improve the speed and ability of locating lost items over the coming weeks. Devices are continuing to join the new Find My Device network, and we expect the network to grow, which will also help improve lost device findability.”
The performance of crowdsourced tracking networks like Android’s Find My Device is dependent on the number of users actively using the service out in the wild. Initially only available in the US and Canada, the upgraded Find My Device network rollout has been slow and only recently expanded to the United Kingdom last month.
As Google states, the network’s performance will improve as it continues to grow, but in its statement to Rahman, it also suggested that users could improve it ahead of the planned enhancements by changing their Find My Device network setting to “With network in all areas.”
The Find My Device network setting defaults to “With network in high-traffic areas only,” which will wait until multiple Android devices have detected a lost item before sharing its location as calculated from several different location reports. As the name implies, that approach works best in busy locations like shopping malls or airports where multiple devices detect a lost item, with an extra layer of privacy for the Android devices sharing location info.
Changing that setting to “With network in all areas” can potentially help locate lost items faster in “lower-traffic areas” because instead of aggregating its location from several Android devices that have detected it, the location from just a single device is used. The reason “With network in all areas” isn’t the default setting is because it requires users to “opt in to sharing location info through the network,” according to Google, which some may be hesitant to do over privacy concerns.

Google is addressing issues with its Find My Device network to help it actually find devices. | Image: Google

Google launched its upgraded Find My Device network in April, giving Android users the equivalent of Apple’s Find My network for locating trackers and offline devices. But some recent reviews have raised concerns over its ability to quickly and reliably find devices and trackers. Google says that it’s aware of the issue and plans to address it “in the coming weeks.”

In a statement provided to Mishaal Rahman, a senior contributor to sites like Android Authority and Android Police, which they shared on X, Google acknowledged the concerns over its Find My Device network. “We are actively working to roll out enhancements to how the Find My Device network operates that will improve the speed and ability of locating lost items over the coming weeks. Devices are continuing to join the new Find My Device network, and we expect the network to grow, which will also help improve lost device findability.”

The performance of crowdsourced tracking networks like Android’s Find My Device is dependent on the number of users actively using the service out in the wild. Initially only available in the US and Canada, the upgraded Find My Device network rollout has been slow and only recently expanded to the United Kingdom last month.

As Google states, the network’s performance will improve as it continues to grow, but in its statement to Rahman, it also suggested that users could improve it ahead of the planned enhancements by changing their Find My Device network setting to “With network in all areas.”

The Find My Device network setting defaults to “With network in high-traffic areas only,” which will wait until multiple Android devices have detected a lost item before sharing its location as calculated from several different location reports. As the name implies, that approach works best in busy locations like shopping malls or airports where multiple devices detect a lost item, with an extra layer of privacy for the Android devices sharing location info.

Changing that setting to “With network in all areas” can potentially help locate lost items faster in “lower-traffic areas” because instead of aggregating its location from several Android devices that have detected it, the location from just a single device is used. The reason “With network in all areas” isn’t the default setting is because it requires users to “opt in to sharing location info through the network,” according to Google, which some may be hesitant to do over privacy concerns.

Read More 

Photo by Chris Welch / The Verge

Microsoft and Qualcomm have convinced software developers to lean into their latest Arm push. When I first used the Arm-powered Surface Pro X in 2019, I loved the hardware but disliked the software experience. Everything felt like it was lagging. Microsoft didn’t have native versions of Edge or its Office apps, and it was clear the Surface Pro X had been released too early. With little support from developers, Windows on Arm was unlikely to succeed.
Nearly five years later, the Windows on Arm experience has improved dramatically. Qualcomm’s new Snapdragon X Elite and X Plus processors deliver a Windows 11 experience that feels like any regular laptop. Microsoft and Qualcomm have also been pushing software developers to create more ARM64 native apps, and it has made a huge difference.
Apps like Photoshop, Dropbox, and Zoom are all native, as are entertainment apps like Spotify, Prime, and Hulu. Even Chrome, Opera, Firefox, Vivaldi, Edge, and Brave are all on Arm now. That’s a good start, but there are still many apps that will have to be emulated on these latest Copilot Plus PCs, which is where Microsoft’s Prism emulator comes in.
Microsoft claims Prism is as efficient as Apple’s Rosetta 2 translation layer and can emulate apps twice as fast as the previous generation of Windows on Arm devices. I’ve been testing the Surface Laptop over the past week and haven’t run into the erratic behavior I saw on Microsoft’s previous emulator, which also impacted battery life on the Surface Pro X. But I also haven’t seen the dramatic improvements in emulated app performance that Microsoft promised.
Microsoft’s claims are difficult to test without comparing previous Arm-based devices. YouTuber Gary Explains did exactly that, comparing the x86 or x64 versions of Firefox, Cinebench R23, and HandBrake on a Surface Pro X without Prism and then with the latest Windows 11 24H2 update that includes Prism.

Photo by Chris Welch / The Verge
Microsoft’s new Surface devices are powered by Arm-based Qualcomm chips.

Gary Explains found that Prism gave a 10 percent performance improvement in Speedometer 3 running on Firefox, an 8 percent jump in Cinebench R23 single core, and a 4.5 percent improvement in Cinebench R23 multicore compared to the previous emulator. HandBrake performance also improved by 8 percent thanks to Prism.
In my own testing, I’ve found that Prism handles compatibility for non-native apps well, but the performance varies depending on the complexity of the app. ShareX, a screenshot tool, works fine using the Prism emulator, but it’s a lightweight app. iA Writer and Notion aren’t native, but they run well on these latest Snapdragon chips, too. Discord also performs a lot better than I’ve seen on Arm in the past, but there’s still some occasional stuttering and a slight lag navigating between servers.
For more heavyweight apps, Prism doesn’t bring the experience up to what you’d find on an Intel- or AMD-powered laptop. Adobe’s Premiere Pro was practically unusable for editing a 4K video on the Surface Laptop, which is probably why Adobe is now blocking its installation on Snapdragon X Elite and Plus processors.
Blender is another example of an emulated app with underwhelming performance. Blender doesn’t detect Qualcomm’s Adreno GPU, so everything hits the CPU instead. The performance for rendering projects is terrible as a result, with one test I performed taking more than 15 minutes to complete, compared to just over two minutes on a 13-inch MacBook Air M3. Blender will soon have a native ARM64 version, but I tested the early alpha copy, and it only marginally improved the results because it’s still not picking up the GPU correctly.
Intel has dominated the laptop GPU market with its integrated solutions for decades, so I suspect Qualcomm still needs to engage with developers of software like Blender to ensure apps are optimized for its GPUs. Blender illustrates that Microsoft’s Prism emulator can’t solve everything.

Photo by Chris Welch / The Verge
Native ARM64 apps make the most of Microsoft’s new Surface devices.

Speaking of GPUs, games also don’t “just work” on the Snapdragon X Elite and X Plus, despite Qualcomm’s assurances. I didn’t make a big deal out of this for the Surface Laptop review because it’s not a gaming laptop, but gaming on Windows on Arm is disappointing right now. Shadow of the Tomb Raider kept crashing for me when I attempted to play, and most of the other games I tried just refused to launch. Fall Guys throws up an unsupported error, as does Halo Infinite. Destiny 2 didn’t even launch — no error, just a whole lot of nothing. Starfield did the same.
There aren’t many native Windows on Arm games, so Prism has its work cut out for it here. I managed to get Grand Theft Auto V working but with lots of frame stuttering. Cyberpunk 2077 also ran on the Surface Laptop 7th Edition but at around 26fps on average at low settings on 1080p resolution. The Witcher 3, Baldur’s Gate 3, Control, Rocket League, and Minecraft all worked out of the box, too.
The biggest issue here is that most anticheat services use kernel drivers that aren’t supported by emulation. BattlEye, a widely used anticheat service, is one of the rare exceptions that supports Windows on Arm, but it seems games like Destiny 2 that utilize this anticheat software will need to be updated to run properly here. Thankfully, there is a dedicated website that tracks which games are supported and run well. I’m not holding out much hope for Arm-powered gaming laptops anytime soon, though.

Screenshot by Tom Warren / The Verge
A lot of games use anticheat technologies that aren’t supported on Windows on Arm.

Another thing I’ve run into is apps just refusing to install. Google Drive is the big one here, as it throws up an error about the Windows architecture of Copilot Plus PCs not being supported. Google’s Drive app on Windows integrates into the shell much like Dropbox, which is something Microsoft didn’t originally support on Windows on Arm. There is, however, a native version of Dropbox that integrates into File Explorer, so hopefully Google is able to deliver a similar experience soon.
There are compatibility issues with external devices, too. I’ve seen reports about Brother printers and scanners not working well on Arm or simply that generic printer drivers don’t support all of the features you’d expect. There’s no easy quick fix for accessories that require driver support, and that’s only likely to come based on the sheer amount of people using these new Copilot Plus PCs. I’m less concerned about the driver issues here because I think most people will be able to plug in the type of accessories (webcams, printers, storage drives) you use on a laptop and have them up and running with the built-in drivers in Windows 11.
VPN apps are still an issue on Windows Arm, too. Bitdefender, NordVPN, and Private Internet Access don’t work. VPN developers use TAP and TUN virtual adapters and devices and need a signed driver from Microsoft to work correctly. Fortunately, Android Authority reports that those three VPN developers are all working on ARM64 versions.
That’s encouraging because the last time I used Windows on Arm regularly in 2019, I said, “Most of the apps I use on a daily basis haven’t been recompiled for ARM and probably never will be.” Now, it feels like app compatibility on Windows on Arm is changing on a daily basis, which is a scenario I wasn’t expecting to see five years ago.
While we’re in this transition point, you may need to use beta versions or download special builds of Windows apps that are ARM64 native — much like the macOS transition. That means the Windows Store versions of apps aren’t always ARM64, and you might be able to find the improved version on the web before the app store version is updated. That was the case initially with Slack earlier this month before the store version got updated.

Screenshot by Tom Warren / The Verge
Microsoft has some extra settings to control the Prism emulator.

For everything else, Microsoft does have some tools for power users that might improve app compatibility on Arm with existing unmodified x86 or x64 apps. There’s a program compatibility troubleshooter that can help enable or disable emulation settings, and you can also toggle these in the properties of an executable. You can control things like hybrid execution mode to force the use of x86-only binaries, disable floating point optimization that could impact performance, and much more. You can also modify how an emulated app uses multiple CPU cores, which might improve performance or compatibility in certain apps.
Ultimately, it’s down to app developers to focus on native ARM64 support for their apps. The sheer amount of native apps that are now available shows things are heading in the right direction. These new Qualcomm chips also provide the brute-force power to emulate apps a little better, alongside Microsoft’s Prism improvements. Day to day, I think most people won’t even run into app issues here because a lot of the key apps are already native or run well in emulation.
I’m confident a lot more ARM64 apps are still on the way. During my testing, benchmark tools and apps were updated to support ARM64, catching me by surprise. I’m willing to bet that we won’t be discussing Prism or emulated app performance as much in a year or two because native ARM64 apps will be as common as x64 apps are today after the transition from x86 began in the early 2000s. After 12 years of attempts to transition to Windows on Arm, it feels like Microsoft is finally about to succeed.

Photo by Chris Welch / The Verge

Microsoft and Qualcomm have convinced software developers to lean into their latest Arm push.

When I first used the Arm-powered Surface Pro X in 2019, I loved the hardware but disliked the software experience. Everything felt like it was lagging. Microsoft didn’t have native versions of Edge or its Office apps, and it was clear the Surface Pro X had been released too early. With little support from developers, Windows on Arm was unlikely to succeed.

Nearly five years later, the Windows on Arm experience has improved dramatically. Qualcomm’s new Snapdragon X Elite and X Plus processors deliver a Windows 11 experience that feels like any regular laptop. Microsoft and Qualcomm have also been pushing software developers to create more ARM64 native apps, and it has made a huge difference.

Apps like Photoshop, Dropbox, and Zoom are all native, as are entertainment apps like Spotify, Prime, and Hulu. Even Chrome, Opera, Firefox, Vivaldi, Edge, and Brave are all on Arm now. That’s a good start, but there are still many apps that will have to be emulated on these latest Copilot Plus PCs, which is where Microsoft’s Prism emulator comes in.

Microsoft claims Prism is as efficient as Apple’s Rosetta 2 translation layer and can emulate apps twice as fast as the previous generation of Windows on Arm devices. I’ve been testing the Surface Laptop over the past week and haven’t run into the erratic behavior I saw on Microsoft’s previous emulator, which also impacted battery life on the Surface Pro X. But I also haven’t seen the dramatic improvements in emulated app performance that Microsoft promised.

Microsoft’s claims are difficult to test without comparing previous Arm-based devices. YouTuber Gary Explains did exactly that, comparing the x86 or x64 versions of Firefox, Cinebench R23, and HandBrake on a Surface Pro X without Prism and then with the latest Windows 11 24H2 update that includes Prism.

Gary Explains found that Prism gave a 10 percent performance improvement in Speedometer 3 running on Firefox, an 8 percent jump in Cinebench R23 single core, and a 4.5 percent improvement in Cinebench R23 multicore compared to the previous emulator. HandBrake performance also improved by 8 percent thanks to Prism.

In my own testing, I’ve found that Prism handles compatibility for non-native apps well, but the performance varies depending on the complexity of the app. ShareX, a screenshot tool, works fine using the Prism emulator, but it’s a lightweight app. iA Writer and Notion aren’t native, but they run well on these latest Snapdragon chips, too. Discord also performs a lot better than I’ve seen on Arm in the past, but there’s still some occasional stuttering and a slight lag navigating between servers.

For more heavyweight apps, Prism doesn’t bring the experience up to what you’d find on an Intel- or AMD-powered laptop. Adobe’s Premiere Pro was practically unusable for editing a 4K video on the Surface Laptop, which is probably why Adobe is now blocking its installation on Snapdragon X Elite and Plus processors.

Blender is another example of an emulated app with underwhelming performance. Blender doesn’t detect Qualcomm’s Adreno GPU, so everything hits the CPU instead. The performance for rendering projects is terrible as a result, with one test I performed taking more than 15 minutes to complete, compared to just over two minutes on a 13-inch MacBook Air M3. Blender will soon have a native ARM64 version, but I tested the early alpha copy, and it only marginally improved the results because it’s still not picking up the GPU correctly.

Intel has dominated the laptop GPU market with its integrated solutions for decades, so I suspect Qualcomm still needs to engage with developers of software like Blender to ensure apps are optimized for its GPUs. Blender illustrates that Microsoft’s Prism emulator can’t solve everything.

Speaking of GPUs, games also don’t “just work” on the Snapdragon X Elite and X Plus, despite Qualcomm’s assurances. I didn’t make a big deal out of this for the Surface Laptop review because it’s not a gaming laptop, but gaming on Windows on Arm is disappointing right now. Shadow of the Tomb Raider kept crashing for me when I attempted to play, and most of the other games I tried just refused to launch. Fall Guys throws up an unsupported error, as does Halo Infinite. Destiny 2 didn’t even launch — no error, just a whole lot of nothing. Starfield did the same.

There aren’t many native Windows on Arm games, so Prism has its work cut out for it here. I managed to get Grand Theft Auto V working but with lots of frame stuttering. Cyberpunk 2077 also ran on the Surface Laptop 7th Edition but at around 26fps on average at low settings on 1080p resolution. The Witcher 3, Baldur’s Gate 3, Control, Rocket League, and Minecraft all worked out of the box, too.

The biggest issue here is that most anticheat services use kernel drivers that aren’t supported by emulation. BattlEye, a widely used anticheat service, is one of the rare exceptions that supports Windows on Arm, but it seems games like Destiny 2 that utilize this anticheat software will need to be updated to run properly here. Thankfully, there is a dedicated website that tracks which games are supported and run well. I’m not holding out much hope for Arm-powered gaming laptops anytime soon, though.

Screenshot by Tom Warren / The Verge
A lot of games use anticheat technologies that aren’t supported on Windows on Arm.

Another thing I’ve run into is apps just refusing to install. Google Drive is the big one here, as it throws up an error about the Windows architecture of Copilot Plus PCs not being supported. Google’s Drive app on Windows integrates into the shell much like Dropbox, which is something Microsoft didn’t originally support on Windows on Arm. There is, however, a native version of Dropbox that integrates into File Explorer, so hopefully Google is able to deliver a similar experience soon.

There are compatibility issues with external devices, too. I’ve seen reports about Brother printers and scanners not working well on Arm or simply that generic printer drivers don’t support all of the features you’d expect. There’s no easy quick fix for accessories that require driver support, and that’s only likely to come based on the sheer amount of people using these new Copilot Plus PCs. I’m less concerned about the driver issues here because I think most people will be able to plug in the type of accessories (webcams, printers, storage drives) you use on a laptop and have them up and running with the built-in drivers in Windows 11.

VPN apps are still an issue on Windows Arm, too. Bitdefender, NordVPN, and Private Internet Access don’t work. VPN developers use TAP and TUN virtual adapters and devices and need a signed driver from Microsoft to work correctly. Fortunately, Android Authority reports that those three VPN developers are all working on ARM64 versions.

That’s encouraging because the last time I used Windows on Arm regularly in 2019, I said, “Most of the apps I use on a daily basis haven’t been recompiled for ARM and probably never will be.” Now, it feels like app compatibility on Windows on Arm is changing on a daily basis, which is a scenario I wasn’t expecting to see five years ago.

While we’re in this transition point, you may need to use beta versions or download special builds of Windows apps that are ARM64 native — much like the macOS transition. That means the Windows Store versions of apps aren’t always ARM64, and you might be able to find the improved version on the web before the app store version is updated. That was the case initially with Slack earlier this month before the store version got updated.

For everything else, Microsoft does have some tools for power users that might improve app compatibility on Arm with existing unmodified x86 or x64 apps. There’s a program compatibility troubleshooter that can help enable or disable emulation settings, and you can also toggle these in the properties of an executable. You can control things like hybrid execution mode to force the use of x86-only binaries, disable floating point optimization that could impact performance, and much more. You can also modify how an emulated app uses multiple CPU cores, which might improve performance or compatibility in certain apps.

Ultimately, it’s down to app developers to focus on native ARM64 support for their apps. The sheer amount of native apps that are now available shows things are heading in the right direction. These new Qualcomm chips also provide the brute-force power to emulate apps a little better, alongside Microsoft’s Prism improvements. Day to day, I think most people won’t even run into app issues here because a lot of the key apps are already native or run well in emulation.

I’m confident a lot more ARM64 apps are still on the way. During my testing, benchmark tools and apps were updated to support ARM64, catching me by surprise. I’m willing to bet that we won’t be discussing Prism or emulated app performance as much in a year or two because native ARM64 apps will be as common as x64 apps are today after the transition from x86 began in the early 2000s. After 12 years of attempts to transition to Windows on Arm, it feels like Microsoft is finally about to succeed.

Read More 

Screenshot: Google

Stack, one of Google’s experimental Android apps, is due to disappear on September 24th.
A product of Google’s Area 120 incubator, Stack was one of those overlooked gems that was just, well, useful: it allowed you to create or import PDFs and store them in different categories that it called Stacks: tax, insurance, medical, etc. The app would pick up data from the PDF to suggest a title, add searchable details like the date of the document, the amount (if it was a receipt), or the organization that issued it, and there was also a field to add notes.
According to Google’s support page, Stack’s “functionality has been incorporated into the Google Drive app.” As someone who used Stack to save and then easily locate a number of PDFs that I needed quick access to, I could argue with that, but hey — it isn’t as if Google isn’t known for its graveyard of abandoned apps.
If you’ve been using Stack to create and track your PDFs, you’re going to want to move some — or all — of your documents. Google suggests that you move them to Google Drive and has incorporated a feature into Stack to make that an easy two-step process:

On the main page of the Stack app, tap on your person icon > Settings.
Select Export all documents to Drive.

Screenshot: Google
Stacks organized your PDFs.

Screenshot: Google
Any additional info will be lost in the export.

When you go to Drive, you’ll find a folder labeled Stack Export; all your PDFs will be within that folder. Unfortunately, if you’re a Stack user and you’ve been carefully categorizing and labeling your PDFs, you’ll find it was wasted effort — the folder will simply contain all the PDFs with none of the categorization or additional text included.
A few additional notes:

If you want to export individual Stack PDFs, go to that PDF and use the export icon at the top right to move it wherever you want it to go.
If you’ve been backing up your Stack content to your Google Drive from the beginning (Settings > Save PDF copy to Drive), then there’s a good chance that you won’t need to export your content since it will already be in Drive.

If you don’t want to keep your documents in Drive, the easiest thing to do is to send them all to Drive and move them out from there. However, you can also export them from Stack wherever you’d like; it’s just a more arduous method:

On the bottom of the Stack app, select All documents.
Long-press on the first document you want to export. You’ll see a checkbox in the upper-left corner; all other PDFs will also now have checkboxes available.
The PDFs will be ordered by date; you can select all the documents from a particular date by tapping the checkbox next to that date. If you want to send all your PDFs somewhere other than Drive, you can select all the dates in the list and use the export icon in the top-right corner.

Once you’ve moved your PDFs out of Stack, you can go to the app’s Settings > Delete page to get rid of all those carefully organized documents.
Finally, from someone who really liked this particular app: thanks a lot, Google.

Screenshot: Google

Stack, one of Google’s experimental Android apps, is due to disappear on September 24th.

A product of Google’s Area 120 incubator, Stack was one of those overlooked gems that was just, well, useful: it allowed you to create or import PDFs and store them in different categories that it called Stacks: tax, insurance, medical, etc. The app would pick up data from the PDF to suggest a title, add searchable details like the date of the document, the amount (if it was a receipt), or the organization that issued it, and there was also a field to add notes.

According to Google’s support page, Stack’s “functionality has been incorporated into the Google Drive app.” As someone who used Stack to save and then easily locate a number of PDFs that I needed quick access to, I could argue with that, but hey — it isn’t as if Google isn’t known for its graveyard of abandoned apps.

If you’ve been using Stack to create and track your PDFs, you’re going to want to move some — or all — of your documents. Google suggests that you move them to Google Drive and has incorporated a feature into Stack to make that an easy two-step process:

On the main page of the Stack app, tap on your person icon > Settings.
Select Export all documents to Drive.

When you go to Drive, you’ll find a folder labeled Stack Export; all your PDFs will be within that folder. Unfortunately, if you’re a Stack user and you’ve been carefully categorizing and labeling your PDFs, you’ll find it was wasted effort — the folder will simply contain all the PDFs with none of the categorization or additional text included.

A few additional notes:

If you want to export individual Stack PDFs, go to that PDF and use the export icon at the top right to move it wherever you want it to go.
If you’ve been backing up your Stack content to your Google Drive from the beginning (Settings > Save PDF copy to Drive), then there’s a good chance that you won’t need to export your content since it will already be in Drive.

If you don’t want to keep your documents in Drive, the easiest thing to do is to send them all to Drive and move them out from there. However, you can also export them from Stack wherever you’d like; it’s just a more arduous method:

On the bottom of the Stack app, select All documents.
Long-press on the first document you want to export. You’ll see a checkbox in the upper-left corner; all other PDFs will also now have checkboxes available.
The PDFs will be ordered by date; you can select all the documents from a particular date by tapping the checkbox next to that date. If you want to send all your PDFs somewhere other than Drive, you can select all the dates in the list and use the export icon in the top-right corner.

Once you’ve moved your PDFs out of Stack, you can go to the app’s Settings > Delete page to get rid of all those carefully organized documents.

Finally, from someone who really liked this particular app: thanks a lot, Google.

Read More 

Photo by Dan Seifert / The Verge

Apple’s putting renewed focus on its repairability efforts today. The company has extended its self-service diagnostics tool to Europe, giving customers in 32 countries an easier way to test products for potential issues.
But perhaps more interesting is the fact that Apple also published a new whitepaper — “Longevity, by Design” (PDF link) — that explains “the company’s principles for designing for longevity.” I’m sure there’ll be a lot of analysis about every page, but within the paper is actually some news: Apple says that later this year, it will extend more software features to third-party iPhone components.
True Tone, the feature that adjusts an iPhone display’s white balance to better match your environment, has typically been disabled whenever a third-party replacement screen is detected by iOS. But soon, Apple will allow customers to enable True Tone “to the best performance that can be provided.”

The company notes that the result might not be up to its usual standards since True Tone depends on server-side calibration that varies by each individual device and “accurate communication between a product’s display and light sensors,” which can’t be guaranteed with a non-OEM component. Customers “will be able to deactivate True Tone in Settings if the display does not perform to their satisfaction.”
Battery metrics are also coming to third-party batteries. Here’s the relevant section from the whitepaper on that, and I’ve emphasized the newsy bit:
Currently, battery health metrics such as maximum capacity and cycle count are not presented to consumers whose devices have third-party batteries. This is because the accuracy of these metrics cannot be verified by Apple. In fact, an Apple internal analysis has found that some third-party batteries sold as new are actually secondhand, with battery health metrics manipulated to appear as new. In an effort to improve support for third-party batteries, starting later in 2024, Apple will display battery health metrics with a notification stating that Apple cannot verify the information presented.
The whitepaper underlines how wary Apple is of third-party batteries. It includes data on many tests that led to failures — sometimes resulting in fire or explosions. “We encourage all consumers to confirm that the product meets stringent safety requirements,” the document reads.
Presumably “later in 2024” means that iOS 18 will be required to get this expanded software support for third-party replacement parts. The next big update is due to enter public beta testing sometime next month before a full release in the fall.
Beyond iPhones, Apple is addressing some other long-running criticisms of its self-service repair program. Crucially, obtaining parts in the first place will get much simpler. The whitepaper says that “customers and service providers will no longer need to input a device’s serial number to the Self Service Repair store to purchase a new part for most repairs.”

Photo by Dan Seifert / The Verge

Apple’s putting renewed focus on its repairability efforts today. The company has extended its self-service diagnostics tool to Europe, giving customers in 32 countries an easier way to test products for potential issues.

But perhaps more interesting is the fact that Apple also published a new whitepaper — “Longevity, by Design” (PDF link) — that explains “the company’s principles for designing for longevity.” I’m sure there’ll be a lot of analysis about every page, but within the paper is actually some news: Apple says that later this year, it will extend more software features to third-party iPhone components.

True Tone, the feature that adjusts an iPhone display’s white balance to better match your environment, has typically been disabled whenever a third-party replacement screen is detected by iOS. But soon, Apple will allow customers to enable True Tone “to the best performance that can be provided.”

The company notes that the result might not be up to its usual standards since True Tone depends on server-side calibration that varies by each individual device and “accurate communication between a product’s display and light sensors,” which can’t be guaranteed with a non-OEM component. Customers “will be able to deactivate True Tone in Settings if the display does not perform to their satisfaction.”

Battery metrics are also coming to third-party batteries. Here’s the relevant section from the whitepaper on that, and I’ve emphasized the newsy bit:

Currently, battery health metrics such as maximum capacity and cycle count are not presented to consumers whose devices have third-party batteries. This is because the accuracy of these metrics cannot be verified by Apple. In fact, an Apple internal analysis has found that some third-party batteries sold as new are actually secondhand, with battery health metrics manipulated to appear as new. In an effort to improve support for third-party batteries, starting later in 2024, Apple will display battery health metrics with a notification stating that Apple cannot verify the information presented.

The whitepaper underlines how wary Apple is of third-party batteries. It includes data on many tests that led to failures — sometimes resulting in fire or explosions. “We encourage all consumers to confirm that the product meets stringent safety requirements,” the document reads.

Presumably “later in 2024” means that iOS 18 will be required to get this expanded software support for third-party replacement parts. The next big update is due to enter public beta testing sometime next month before a full release in the fall.

Beyond iPhones, Apple is addressing some other long-running criticisms of its self-service repair program. Crucially, obtaining parts in the first place will get much simpler. The whitepaper says that “customers and service providers will no longer need to input a device’s serial number to the Self Service Repair store to purchase a new part for most repairs.”

Read More