The Verge

112 readers
220 users here now

News community for TheVerge. Will be deleted or retired once the Verge officially supports ActivityPub in their site.


This is an automated RSS-Feed community. If you dislike RSS Feed communities consider blocking it, or the bot.

founded 1 month ago
MODERATORS
1
 
 

In her ruling banning Apple from charging a commission on purchases made outside the App Store, Judge Yvonne Gonzalez Rogers says that an Apple executive “outright lied under oath.”

According to Gonzalez Rogers, Alex Roman, vice president of finance at Apple, gave testimony that was “replete with misdirection and outright lies.” The judge writes that Roman lied when testifying that Apple hadn’t decided on the 27 percent number for its fee on purchases outside the App Store until January 16th, 2024.

“Contemporaneous business documents reveal that on the contrary, the main components of Apple’s plan, including the 27 percent commission, were determined in July 2023,” Gonzalez Rogers says. “Neither Apple, nor its counsel, corrected the, now obvious, lies.”

Gonzalez Rogers says that she is referring the case to a US attorney for possible criminal contempt proceedings against Apple and Roman.

“Apple willfully chose not to comply with this Court’s Injunction,” Gonzalez Rogers says at the end of the filing (emphasis hers). “It did so with the express intent to create new anticompetitive barriers which would, by design and in effect, maintain a valued revenue stream; a revenue stream previously found to be anticompetitive. That it thought this Court would tolerate such insubordination was a gross miscalculation. As always, the cover-up made it worse. For this Court, there is no second bite at the apple.”


From The Verge via this RSS feed

2
 
 

World’s orbs have scanned millions of peoples’ eyes to verify their identity on a blockchain.

Over the past few years, what sounds like an episode of Black Mirror has been unfolding in real life: millions of people around the world have agreed to let metallic orbs scan their eyes in exchange for receiving blockchain-based ID and a cryptocurrency called Worldcoin.

Beginning this week, Worldcoin (WLD) will be available in most of the US for the first time, including via exchanges like Coinbase. Those who scan their eyes at a World orb will receive 16 WLD. The Sam Altman co-founded startup behind the project, Tools for Humanity, is also working with Visa to release a debit card later this year that converts WLD to fiat currency at checkout. Altman, who is currently the chairman of Tools for Humanity, announced the news alongside CEO Alex Blania at a press event in San Francisco on Wednesday, dubbed "At Last."

World's orbs will be available for iris scans at physical "World Space" locations in Atlanta, Austin, Los Angeles, Miami, Nashville, and San Francisco, as well as in Razer gaming stores. Tools for Humanity updated the orb's physical design last year, and now it's releasing a smaller "orb mini" that's designed to be more portable.

Since it was founded by Altman, Blani …

Read the full story at The Verge.


From The Verge via this RSS feed

3
 
 

E**pic Games v. Apple judge Yvonne Gonzalez Rogers just ruled that, effective immediately, Apple is no longer allowed to collect fees on purchases made outside apps and blocks the company from restricting how developers can point users to where they can make purchases outside of apps.

The ruling was issued as part of Epic Games’ ongoing legal dispute against Apple, and it’s a major victory for Epic’s arguments. Rogers also says that Apple “willfully” chose not to comply with her previous injunction from her original 2021 ruling. “That [Apple] thought this Court would tolerate such insubordination was a gross miscalculation,” Rogers says.

The judge also referred the case to the US attorney to review it for possible criminal contempt proceedings.

As part of the ruling, Judge Yvonne Gonzalez Rogers says that Apple cannot:

Impose “any commission or any fee on purchases that consumers make outside an app”Restrict developers’ style, formatting, or placement of links for purchases outside of an appBlock or limit the “use of buttons or other calls to action”Interfere with consumers’ choice to leave an app with anything beyond “a neutral message apprising users that they are going to a third-party site”

Epic Games CEO Tim Sweeney says that, following this ruling, the company will bring Fortnite back to the US App Store “next week.” Sweeney is also offering a “peace proposal” from Epic: “If Apple extends the court’s friction-free, Apple-tax-free framework worldwide, we’ll return Fortnite to the App Store worldwide and drop current and future litigation on the topic.”

We will return Fortnite to the US iOS App Store next week.Epic puts forth a peace proposal: If Apple extends the court's friction-free, Apple-tax-free framework worldwide, we'll return Fortnite to the App Store worldwide and drop current and future litigation on the topic. https://t.co/bIRTePm0Tv

— Tim Sweeney (@TimSweeneyEpic) April 30, 2025

In many cases, Apple takes a 30 percent cut of purchases made in its apps, and Rogers’ 2021 ruling forced Apple to allow developers to point to alternative payment options. But Apple instituted a policy that demanded developers pay Apple a 27 percent commission on those purchases, which many companies, including Epic, were unhappy about.

Apple didn’t immediately reply to a request for comment.

Developing…


From The Verge via this RSS feed

4
 
 

Senator Elizabeth Warren (D-MA) wants to know whether Amazon founder Jeff Bezos received any “promises or favors” from President Donald Trump in exchange for his “subservience” after dropping all plans to show customers the cost of tariffs. In a letter sent on Wednesday, Warren questions Bezos about the details of his conversation with Trump and raises concerns “about the potential for tariff-related corruption.”

On Tuesday, Trump called Bezos to complain after a Punchbowl News report said that Amazon planned on showing the cost of tariffs right next to a product’s total price. Amazon later issued a statement saying that displaying tariff prices “is not going to happen,” adding that the company only “considered” showing the fees on its Haul store, which ships products directly from China. Trump then told reporters that Bezos is “very nice” and “solved the problem very quickly,” as reported by The New York Times.

Warren says Bezos’ conversation represents “another example of Big Tech working together with President Trump to seek special favors or support his policies.” She also criticizes Amazon’s move not to show the impact of tariffs on its products.

“If Amazon had followed through on any plans to provide transparency on tariff costs, it could have provided important information for consumers, allowing them to find out for themselves some of the true costs of President Trump’s broad and chaotic tariff policies,” Warren writes.

In her letter, Warren presses Bezos for more details about Amazon’s plans to show the cost of tariffs before his conversation with Trump, as well as when exactly the company decided to change plans. Warren also asks if Trump threatened Bezos with “adverse policy consequences” if Amazon followed through on showing the impact of tariffs, and if the president promised any concessions or exemptions from tariffs upon reversing the plan.

Warren sent Apple CEO Tim Cook a similar letter after Trump announced tariff exemptions for smartphones and other electronics.


From The Verge via this RSS feed

5
 
 

The Meta AI app could soon get a paid tier, similar to the ones offered by rivals like OpenAI, Google, and Microsoft. Meta CEO Mark Zuckerberg described the plan during a Q1 2025 earnings call on Wednesday, saying there’s an opportunity to offer a “premium service for people who want to unlock more compute or additional functionality” in Meta AI.

As part of Meta’s efforts to compete with ChatGPT, the company launched a standalone Meta AI app this week, allowing you to interact with the chatbot and generate images from within the app. The chatbot, which Meta says now has nearly 1 billion users, was previously only available within apps like Facebook, Messenger, and WhatsApp.

OpenAI’s ChatGPT, Google’s Gemini, and Microsoft’s Copilot all offer paid subscriptions that give users access to more advanced features and compute. Meta reported earning $42 billion in revenue over the past few months and revealed that it now expects to invest up to $72 billion on AI, rather than the up to $65 billion Zuckerberg announced previously.

Additionally, Zuckerberg mentioned incorporating “product recommendations or ads” within Meta AI. It’s not clear when ads, or a paid tier, might roll out, as Zuckerberg said, “I expect that we’re going to be largely focused on scaling and deepening engagement for at least the next year before we’ll really be ready to start building out the business.”


From The Verge via this RSS feed

6
 
 

Following a court order that blocks Apple from taking a commission on purchases made outside the App Store, Epic Games CEO Tim Sweeney says on X that the company plans to bring Fortnite back to iOS in the US “next week.”

The app hasn’t been available on iOS in the US since August 2020, when Apple kicked it off the App Store for implementing Epic’s own in-app payment system in violation of Apple’s rules. Since then, Apple and Epic have been embroiled in an ongoing legal battle, including a ruling more in Apple’s favor in 2021 and today’s ruling that is a major victory for Epic.

We will return Fortnite to the US iOS App Store next week.Epic puts forth a peace proposal: If Apple extends the court's friction-free, Apple-tax-free framework worldwide, we'll return Fortnite to the App Store worldwide and drop current and future litigation on the topic. https://t.co/bIRTePm0Tv

— Tim Sweeney (@TimSweeneyEpic) April 30, 2025

Fortnite has returned to iOS in the EU, but only via Epic’s own mobile app store, which isn’t available in the US.

Sweeney also offered a “peace proposal” from Epic to Apple in his post on X. “If Apple extends the court’s friction-free, Apple-tax-free framework worldwide, we’ll return Fortnite to the App Store worldwide and drop current and future litigation on the topic.”

Apple didn’t immediately reply to a request for comment.


From The Verge via this RSS feed

7
 
 

The WiZ HDMI sync box can make your WiZ lights dance with movies and music.

Budget smart lighting brand WiZ has announced a new HDMI sync box that syncs with its line of Wi-Fi-based LED lights to match what’s playing on your TV.

WiZ’s HDMI Sync Box with TV backlight will be available next month. It comes in two models: one with a TV backlight strip for screens 55 inches to 65 inches in size for $89.99, and the other for screens 75 inches to 85 inches in size for $109.99.

The new HDMI box is similar to that of WiZ’s sister brand, Philips Hue. But Hue’s HDMI Play Box starts at $249.99, with light strips sold separately. (Hue and WiZ are both owned by Signify, but their bulbs are not compatible with each other.) WiZ bulbs also use Bluetooth and Wi-Fi, whereas Hue relies on Bluetooth and Zigbee (via a hub).

Unlike camera-based light syncing systems from companies like Nanoleaf and Govee, WiZ and Hue’s sync products read the signal directly from the video source, which can help avoid latency issues.

This makes WiZ’s option a great budget alternative for those put off by Hue’s high prices. However, there are some concessions. The WiZ box only has one HDMI input, whereas the Hue Play supports four. This means if you have multiple devices connected to your TV, you’ll have to manually switch between them.

The WiZ Sync Box supports HDMI 2.0 and video formats up to 4K at 60Hz with HDR10 Plus and Dolby Vision. Hue’s latest Play HDMI Box supports up to 8K video, HDMI 2.1, and refresh rates up to 120Hz for 4K for console gaming. It also costs $350

The WiZ box also features a built-in microphone that captures audio, allowing the lights to change in sync with music as well as movies. The box can be controlled with the WiZ app, with voice via Google or Alexa, or the WiZ remote control.

According to the company, the Sync Box can connect to any color-capable WiZ light that supports both Wi-Fi and Bluetooth, including its new Gradient Light Bars ($59.99) and Gradient Floor Light ($89.99), which can sit alongside your TV to add more ambient light. As with all of WiZ’s lighting products, these support the Matter smart home standard, so are compatible with most smart home systems.


From The Verge via this RSS feed

8
 
 

Kia is opening up preorders in the UK for its PV5 Passenger electric van at a starting price of £32,995, or about $44,000 with a standard range battery.

Kia says preorders will open up tomorrow, May 1st, with initial deliveries in the UK planned for “late 2025.” It will come in a five-seat configuration (seven seats will come later) and customers can choose between the base “Essential” package or the “Plus” version that comes standard with the long range battery.

The standard range 51.5kWh battery can power the PV5 for 149 miles on a single charge based on the European WLTP spec, while the long range 71.5 kWh pack gets 249 miles of range. Kia says pricing for the larger battery will be revealed later.

The Plus upgrades the Essential’s 16-inch steel wheels to alloy ones, and adds an optional heat pump. The higher trim also includes heated front seats with power adjustments, a heated steering wheel, a powered tailgate, Vehicle-to-Load AC outlet adapter support, wireless phone charging, safety additions like cross traffic collision avoidance, power folding mirrors, and more.

Kia says it’s keeping options “simple and straightforward,” so you can’t select features between the two trims beyond upgrading the standard white paint to a premium option like “Cityscape Green” or “Runway Red.”

The PV5 is part of Kia’s lineup of all-electric PBVs (Platform Beyond Vehicle) that were introduced at CES in 2024. These modular EVs have a flexible architecture where the bodies can be swapped into a minivan, a full-sized van, or a small truck, depending on needs. The larger PV7 is coming in 2027, and the PV9 in 2029.

On the PV5, both batteries get a single electric motor with a max output of 120kW for the long range. The PV5’s 400-volt E-GMP.S architecture offers up to 150kW charging, which Kia says can charge from 10 to 80 percent in less than 30 mins.

Kia will also offer the PV5 in commercial Cargo, Crew, and Chassis Cab versions in 2026, when it starts shipping the vehicle to additional markets. The automaker intends the PV5 to also be used in fleets for deliveries and ride hailing. Uber has already signed on as Kia’s first customer.


From The Verge via this RSS feed

9
 
 

The Nintendo Switch just got a new update that adds Virtual Game Cards as a way to share digital games with other people. But the update also closes a loophole that would let you play shared games online at the same time as somebody you were sharing the game with, as reported by Eurogamer.

Before the introduction of Virtual Game Cards, the only way Nintendo allowed you to share digital Switch games was by logging into your Nintendo Account on a second Switch and downloading the game to that console. Using that system, which is now available under “Online License Settings,” it was possible for you and a friend to play that game that you owned online simultaneously.

Following the new Switch update, that digital sharing system is still an option – you don’t have to use Virtual Game Cards. However, with the old system, two people can’t play a shared game online at the same time anymore, which could be a disappointment if you relied on it ahead of this update.

If you do opt to use Virtual Game Cards, Nintendo won’t just force you to manage those on a Switch or Switch 2; the company will also let you manage them from a website, which could be handy if you don’t have your Switch nearby but want to share a game.


From The Verge via this RSS feed

10
 
 

Alongside all the rumors lately about a big visual redesign for iOS and a more Mac-like iPadOS experience, there’s one interesting one that’s been bubbling up: the iPhone could get a desktop mode of sorts.

Majin Bu, who has a mixed history with leaks, wrote last week that unnamed sources had said iPhones with USB-C will offer “a Stage Manager like interface” when connected to an external display. It wouldn’t be quite a full desktop experience, but would “allow users to extend their screen space, great for presentations, editing, or enhanced viewing.” They added that there “may be some limitations possibly in resolution or the number of apps displayed at once.” Given Bu’s history, it’s hard to put too much stock in that all on its own.

Then again, in a new episode of the 9to5Mac’s video series iOS Decoded, Jeff Benjamin says that he’s seen Stage Manager working with an iPhone and, tantalizingly, it had “full-on external display support.” In an article accompanying the video, he writes that 9to5Mac has been able to test the feature using the iOS Xcode Simulator. He doesn’t think it’s likely Apple will roll that out to iPhones in the real world, but it’s notable that Apple has been interested enough in the idea to test it out.

Apple wouldn’t be the first to try this — there were attempts that predated even Samsung’s DeX — but it’s an intriguing idea all the same. iPhones already have limited external display support, using cables or wirelessly through AirPlay, but that support is generally limited to either screen mirroring or on a per-app basis — like casting a YouTube video to an AirPlay-enabled TV.


From The Verge via this RSS feed

11
 
 

Sundar Pichai, Google's CEO, took the stand on Wednesday during the remedies phase of the company's search antitrust trial, and offered a simple message. The US government's plan to rectify Google's search monopoly, he said, would be so crushing to Google Search that it might be hard to justify continuing to build a search engine at all.

Pichai was called by Google in its defense, the second witness in its portion of the trial after the Department of Justice finished more than a week of its own questioning. John Schmidtlein, one of Google's lead attorneys, first led Pichai through a tour of Google's R&D investments, asking him how much the company has spent on Search, AI, and other projects. (The answer: about $49 billion just last year.) Then he asked Pichai about the government's proposal to require Google to share much of its search data, and its search index, with competitors at a "marginal cost."

Pichai said the data-sharing proposal would be a disaster. He called it "far-reaching" and "extraordinary," appearing on the stand practically flummoxed by the very idea of the proposal. He said that requiring Google to give up all the data in its search index, and the ways it rank …

Read the full story at The Verge.


From The Verge via this RSS feed

12
 
 

Within weeks of President Trump stepping into office, key health and environmental resources that doctors and farmers rely on started disappearing from federal websites. Trump was also quick to dismantle the US Agency for International Development (USAID), cutting off funding - as well as the flow of data that people around the world use to prevent famine and issue warnings ahead of natural disasters.

"As we all watched the websites being pulled down, as we all watched data disappearing, we were all concerned - because that's truth. There's truth in data," says a former contractor who was granted anonymity to speak freely without fear of repercussions. "Now it's more easy just to say something and force it as the truth. But there's no way to back it up."

The US collects a vast amount of weather and climate data, an essential resource for humanitarian efforts around the globe. It guides efforts to predict where droughts and crop failures might lead to food shortages, where people are at risk from flash flooding, and how to prepare for the Atlantic hurricane season. Even if these systems eventually get back up and running, erratic changes under the Trump administration are already …

Read the full story at The Verge.


From The Verge via this RSS feed

13
 
 

BBC Studios is using AI to recreate the voice and likeness of late detective story author Agatha Christie for the purpose of featuring it in digital classes that teaches prospective writers “how to craft the perfect crime novel.” A real life actor, Vivien Keene, is standing in for Christie, with her appearance augmented by AI to resemble the author.

The new class, called Agatha Christie Writing, is available today on BBC Maestro, the company’s $10-per-month online course service that usually gives you access to content from living professionals teaching things like graphic design, bread making, time management, and more.

Deepfaked Agatha Christie’s teachings are “in Agatha’s very own words,” her great-grandson James Prichard said in a press release. It uses insights from the real Christie and is scripted by academics — so the actual content appears to be human-made and not generated from a model that’s been fed all of her work. BBC collaborated with Agatha Christie Estate and used restored audio recordings, licensed images, interviews, and her own writings to make this all happen.

Live now, the class has 11 video lessons with 12 exercises for prospective writing students, including how to “structure an airtight plot” and “build suspense.”


From The Verge via this RSS feed

14
 
 

photo of Kia cars on a lot New Kia vehicles at the Port of Seattle in Seattle, Washington, US, on Wednesday, April 16, 2025.

It started last week with Tesla, followed quickly by General Motors, Mercedes-Benz, and Volvo. Automakers across the spectrum are pulling their guidance for the year because they can’t figure out how to accurately plan for the future thanks to President Donald Trump’s ever-shifting tariffs.

The auto industry is paralyzed by uncertainty. Stellantis, the parent company of Jeep, Dodge, and Ram, also recently scrapped its profit guidance for the year, with Chief Financial Officer Doug Ostermann telling analysts, “Most of us are in a period of waiting for a bit more clarity,” according to Reuters.

The lack of clarity was even more stark this week, as Trump signed a new executive order walking back some of the tariffs he had just imposed. Now car companies that pay a 25 percent tariff on auto imports won’t have to pay other levies, like on steel and aluminum, or on certain imports from Canada and Mexico. And yet the rules don’t appear to protect automakers from tariffs on steel and aluminum that their suppliers pay and pass on to them.

Car companies hate this stuff. These are global, multibillion-dollar companies that like to plan years — if not decades — ahead, weighing investments in factories and new models based on sales predictions and forecasts. The foundation of all this is market certainty.

The auto industry is paralyzed by uncertainty

But Trump’s tariffs have been anything but certain. And most analysts are predicting more chaos to come. The president’s recent modifications now require automakers who assemble their vehicles in the US to file for partial reimbursements on 25 percent tariffs levied on auto parts for two years. Basically Trump is dropping a whole mess in the laps of these car companies, who will now be forced to navigate the complexity of tariff compliance and reimbursement, while also trying to relocate as much of their supply chains to within the US.

For a better snapshot of how bizarre this is, just look at GM. The nation’s largest automaker was expecting a pretty robust year of profits until Trump blew everything up with his tariffs. This week, GM pulled its guidance, explaining that any prediction of profits at this point would be “a guess,” according to The New York Times. The company also postponed its conference call with financial analysts to discuss its first-quarter results by a couple days, while it assesses the impact of the latest curveball from the White House.

Ultimately, most analysts are sticking with their predictions of higher prices for new cars that will almost certainly be passed along to consumers. Wedbush’s Dan Ives said Trump’s most recent tweak to the auto tariffs “sounds good on paper,” but still won’t prevent disaster from descending on the industry. He said that car companies across the board are facing “a brutal situation.”

“A US car with all US parts made in the US is a fictional tale not possible today and many factories/production hubs could take 4–5 years to build in the US… and this speaks to the massive frustration from the industry as the rules of the US tariff game are untenable in our view,” Ives wrote in a note.

Automakers have been desperately lobbying the administration for relief from the tariffs, arguing that two years isn’t enough time to completely reorganize their complex manufacturing processes. And given the mercurial nature of Trump’s approach to tariffs, it’s not clear whether any of these massive investments are worth making. They could wake up tomorrow under an entirely different edict.

Even Tesla CEO (and “First Buddy”) Elon Musk said he was doing his best to nudge the president into lowering tariffs on imported car parts. “I believe lower tariffs are generally a good idea for prosperity,” Musk said during Tesla’s earnings call last week. “But this decision is fundamentally up to the elected representative of the people, being the President of the United States. So, you know, I’ll continue to advocate for lower tariffs rather than higher tariffs, but that’s all I can do.”

The situation is fast approaching a crisis, with all major auto stock prices trading lower each successive day. Customers are panicking, dealers are scrambling, and new models are being put on hold. It’s like being trapped in purgatory, and there’s no immediate sign that anyone is getting out.


From The Verge via this RSS feed

15
 
 

TikTok is back in a Washington, DC courthouse discussing the US law that - at least on paper - effectively banned the app. But this time, it's serving as a witness for the government, not fighting against it.

On Wednesday, TikTok's head of operations and trust and safety Adam Presser testified in the Federal Trade Commission's antitrust trial against Meta, in the same courthouse where a panel of judges ruled that the government could expel TikTok from the country. Presser's role in the Meta trial was to explain the ways in which TikTok competes (or doesn't) with Meta's services in a market the FTC has defined as personal social networking - a category the FTC says contains only Meta's services, Snapchat, and a small app called MeWe.

Lawyers for both the FTC and Meta brought up filings TikTok made in its own litigation against the 2024 divest-or-ban law, which required Chinese parent company ByteDance to sell its US business. (President Donald Trump has used dubious legal measures to extend the deadline for enforcement twice, leaving TikTok's service providers facing the potential for billions of dollars in penalties while saying he's working on a deal.) The FTC pointed to a TikT …

Read the full story at The Verge.


From The Verge via this RSS feed

16
 
 

For several weeks, President Donald Trump and his administration have been grasping for evidence that a man his administration deported and imprisoned in error is a dangerous gang member, and the effort has now reached what may be an untoppable peak: the President repeatedly insisting in an in-person interview that an obvious text label that says MS13 in the Calibri typeface is an actual tattoo.

On Tuesday, Trump sat for an interview with ABC News correspondent Terry Moran. Moran asked Trump about Kilmar Abrego Garcia, who was sent to El Salvador's CECOT prison without due process and in violation of a protective order, and who remains in El Salvador despite a Supreme Court ruling demanding he facilitate the man's return. Trump said (contradicting his administration's previous public statements) that he "could" pick up the phone and order Abrego Garcia's return, but that he refused because Abrego Garcia was a member of the MS-13 gang. His evidence: "on his knuckles, he had MS-13."

That would be the sort of evidence the government ordinarily would present at an immigration hearing, but so far it's refused. Trump has posted what he claims to be an image of these tattoos to his Tru …

Read the full story at The Verge.


From The Verge via this RSS feed

17
 
 

OpenAI rolled back a GPT-4o update for ChatGPT that caused the chatbot’s default personality to be “overly flattering or agreeable – often described as sycophantic” and that “sycophantic interactions can be uncomfortable, unsettling, and cause distress,” the company says in a blog post.

The company introduced a GPT-4o update last week that included adjustments “aimed at improving the model’s default personality to make it feel more intuitive and effective across a variety of tasks,” according to the post. OpenAI says it starts shaping model behavior first with what’s outlined in its Model Spec and teaches the models how to apply the principles in that spec “by incorporating user signals like thumbs-up / thumbs-down feedback on ChatGPT responses.”

But with the rolled-back update, OpenAI says that “we focused too much on short-term feedback, and did not fully account for how users’ interactions with ChatGPT evolve over time.” That meant that “GPT‑4o skewed towards responses that were overly supportive but disingenuous.”

OpenAI designs ChatGPT’s default personality to “reflect our mission and be useful, supportive, and respectful of different values and experience,” the blog post says, but adds that “each of these desirable qualities like attempting to be useful or supportive can have unintended side effects.” The company says that “a single default can’t capture every preference” for its 500 million weekly ChatGPT users.

OpenAI will be “taking more steps to realign the model’s behavior,” including “refining core training techniques and system prompts to explicitly steer the model away from sycophancy” and “expanding ways” for users to give feedback. “We also believe users should have more control over how ChatGPT behaves and, to the extent that it is safe and feasible, make adjustments if they don’t agree with the default behavior,” the company says.


From The Verge via this RSS feed

18
 
 

The Google Play Store shed 1.8 million apps since the beginning of last year, according to an analysis from Appfigures and reported earlier by TechCrunch. That makes up around 47 percent of the 3.4 million apps it started 2024 with — a decline that Google spokesperson Dan Jackson says reflects the company’s focus on “delivering high-quality apps” and “commitment to continuous improvements for user safety.”

In recent years, Google has ramped up efforts to crack down on spammy and low-quality apps. In 2023, the company started carrying out more thorough app reviews and began requiring developers to test their apps with at least 20 people for two weeks. It began purging apps with “limited functionality and content” last year, and announced in January that it blocked 2.36 million policy-violating apps from being published to the Play Store.

As shown by data from Appfigures, nearly 200,000 of the titles removed from the Google Play Store were in the “games” category, followed by 160,700 in education, and 115,400 in business. However, Appfigures shows that app releases are up 7.1 percent globally compared to the same time last year.

“Every year, we continue to invest in more ways to protect our community and fight bad actors, so users can trust the apps they download from Google Play and developers can build thriving businesses,” Jackson said in a statement to The Verge.

Though the Google Play Store may have lost a significant number of apps since last year, Appfigures’ data shows Apple’s App Store making a small jump from 1.6 million to 1.64 million apps.


From The Verge via this RSS feed

19
 
 

Microsoft just posted the third quarter of its 2025 fiscal financial results. The software maker made $70.1 billion in revenue and a net income of $25.8 billion during Q3. Revenue is up 13 percent, and net income has increased by 18 percent. Microsoft was expected to report $68.42 billion in revenue, so the results have beaten Wall Street expectations.

Cloud growth appears to be the main driver of improved revenues, with overall Microsoft Cloud revenue reaching $42.4 billion, up 20 percent. Microsoft hasn’t split out its AI revenues, so it’s still unclear exactly how much that part of its business is growing, though.

With all eyes on the impact of the Trump tariffs and consumer spending, Microsoft also reported Windows OEM and devices revenue growth of 3 percent year-over-year. Businesses and consumers are upgrading PCs and laptops ahead of the Windows 10 end of life in October, but there hasn’t been a big upswing in Windows OEM revenue yet. Microsoft does say that Windows OEM growth was partly down to “inventory levels remaining elevated due to tariff uncertainty.”

Microsoft now combines Surface revenue with Windows OEM revenue, so it’s no longer clear exactly how its Surface line is doing. Microsoft hasn’t launched any new consumer Surface devices in the recent quarter, but the company did start selling Intel-powered versions of the Surface Pro 11 and Surface Laptop 7 in February for businesses. Microsoft is also rumored to be launching smaller versions of the Surface Pro and Surface Laptop soon.

On the gaming side, Microsoft warned last quarter that Xbox hardware revenue would decline and it has again. Xbox hardware is down 6 percent this quarter, in a period when Microsoft has been launching a variety of games into Game Pass. Overall gaming revenue at Microsoft is up 5 percent.

Xbox content and services, which includes Game Pass, is up 8 percent, though. Microsoft isn’t revealing any new Game Pass subscriber numbers though, and the last time we got an update was in February 2024 when Microsoft revealed there were 34 million subscribers, including the Xbox Game Pass Core (previously Xbox Live Gold) members.

Over the past year Microsoft has been pushing ahead with its strategy to bring more previously Xbox-exclusive games to PlayStation and Nintendo Switch. The move has proven to be a successful one, with preorders for Indiana Jones and the Great Circle and Forza Horizon 5 topping Sony’s own PlayStation Store last month.

Microsoft’s Xbox everywhere strategy — which I’ve been covering closely in my Notepadnewsletter over the past year — will also include titles for the Nintendo Switch 2. Microsoft is preparing a similar list of games to port over to the Switch 2 this year.

As is always the case for Microsoft’s earnings, the company’s Office and cloud businesses are once again the stars of the show. Microsoft 365 commercial products and cloud services revenue increased 11 percent year over year, and even office consumer is up 10 percent in revenue. Microsoft 365 consumer subscribers are up to 87.7 million now.

Microsoft bundled its Office AI features into Microsoft 365 Personal and Family subscriptions earlier this year and raised prices. Previously, Microsoft 365 subscribers had to pay an extra $20 per month to get Copilot inside Office apps like Word, Excel, and PowerPoint as part of a Copilot Pro subscription, but Microsoft now offers these AI features inside Microsoft 365 apps for an extra $3 per month. Existing subscribers have been able to opt out of the AI features and not suffer the price increase, but it’s not clear how many have actually decided to do that.

Microsoft’s revenues in what it calls intelligent cloud have increased by 21 percent year-over-year to $26.8 billion, which is exactly double the revenue generated by Windows, Xbox, and search and news advertising (more personal computing). Server products and cloud services revenue increased 22 percent, with Azure growth of 33 percent.

Microsoft will now hold an earnings call at 5:30PM ET and we’ll update this article with any relevant comments from Microsoft CEO Satya Nadella and CFO Amy Hood.


From The Verge via this RSS feed

20
 
 

Meta is making a few notable adjustments to the privacy policy for its Ray-Ban Meta smart glasses. In an email sent out on April 29th to owners of the glasses, the company outlined two key changes. First, it’s giving Meta AI a more frequent view of the world. “Meta AI with camera use is always enabled on your glasses unless you turn off ‘Hey Meta,” the email said, referring to the hands-free voice command functionality.

So unless you turn that convenience-minded feature off, Meta will frequently be analyzing whatever’s captured by the built-in camera. If you simply want to use the Ray-Ban Metas as a “normal” camera without any artificial intelligence thrown in, you’ll have to disable “Hey Meta” and stick to the physical controls.

Second, Meta is taking after Amazon by no longer allowing Ray-Ban Meta owners to opt out of having their voice recordings stored in the cloud. “The option to disable voice recordings storage is no longer available, but you can delete recordings anytime in settings,” the company wrote.  In its voice privacy notice, Meta states that “voice transcripts and stored audio recordings are otherwise stored for up to one year to help improve Meta’s products.” If the company detects that a voice interaction was accidental, those recordings are deleted after a shorter 90-day window.

The motivation behind these changes is clear: Meta wants to continue providing its AI models with heaps of data on which to train and improve subsequent results. Some users began noticing these policy changes in March, but at least in the United States, Meta says they went into effect as of April 29th.

Earlier this month, the company rolled out a live translation feature to the Ray-Ban Meta product. And just yesterday, Meta rolled out a standalone Meta AI app on smartphones to more directly compete with Open AI’s ChatGPT, Google Gemini, Anthropic’s Claude, and other AI chatbots.

The company is reportedly planning a higher-end pair of Ray-Ban Meta glasses for release later in 2025. The current glasses lineup starts at $299, but the more premium version could cost around $1,000. Meta is set to report its Q1 2025 earnings later on Wednesday, and the company is likely to address the tariff chaos that has roiled markets in recent months.


From The Verge via this RSS feed

21
 
 

A blurred video thumbnail on YouTube. A small number of YouTube viewers will now see blurred thumbnails for certain search query results. | Screenshot: YouTube

YouTube has announced a new experimental feature that’s “aimed at providing safer search experiences for all users.” The service will begin blurring the thumbnails in search results for some queries “that frequently include sexual themes.” According to a recent update in the Community section of YouTube’s Help Center website, the feature is currently being rolled out to a small percentage of viewers.

There are no specifics about what sexual themes may trigger a search to return a list of videos with blurred thumbnails, but YouTube says the video’s title, channel name, and its description will remain visible. Viewers will also have the option to unblur thumbnails.

The goal of the limited test rollout is to “understand whether this type of feature helps users avoid accidentally viewing content that follows YouTube’s Community Guidelines but may be sensitive in nature.” But unlike SafeSearch that can both blur and filter out results on Google’s search engine, YouTube’s new experimental feature won’t omit results. In its current form it’s instead designed to provide an extra layer of protection that prevents younger or unsuspecting users from immediately being presented with content that may be inappropriate.


From The Verge via this RSS feed

22
 
 

DOGE flag

I once wondered whether law enforcement might stop Elon Musk's power grab. About a month ago, I got my answer: no.

On March 17th, the Musk-led Department of Government Efficiency broke into the nonprofit, congressionally funded US Institute for Peace, according to court documents filed by the USIP's board of directors. By allegedly threatening to "cancel every federal contract" of a private security firm that had worked with USIP until a day earlier, they convinced it to let them into the building - where, told by institute staff they were trespassing, the fired security firm headed for the USIP gun safe. That's when the USIP called the cops. The DC police arrived to escort DOGE into the building. USIP head of security Colin O'Brien, along with two of USIP's lawyers, was detained.

Any reasonable reading of these court documents shows that a fired security firm broke into the offices of a nonprofit agency, under a dubiously legal executive order and what, as described, sounds like straightforward extortion. Police officers elected to help the people who were breaking and entering. USIP's board of directors has sued DOGE and assorted other parties in Trump's administration, but …

Read the full story at The Verge.


From The Verge via this RSS feed

23
 
 

The Boox Mira Pro Color monitor showing a color brochure while sitting on a desk. Boox has upgrade its Mira Pro E Ink monitor with the same color screen technology found in e-readers. | Image: Boox

Boox is debuting its first desktop monitor featuring a color E Ink screen. The new version of the Boox Mira Pro carries over the same functionality as the black-and-white version that launched in 2023, but using the same E Ink color display technology found in devices like the Kindle Colorsoft.

E Ink offers several advantages to display technologies like LCD and OLED, including a screen that reflects light instead of emitting it, so it’s potentially easier on the eyes for prolonged use. Still, large E Ink panels are expensive. The Mira Pro Color is only 25.3-inches in size but is priced at $1,899.99. Boox also says its Mira Series of E Ink monitors only ship from China and “some countries may levy tariffs on the imported goods.” If you’re having one shipped to the US, it could cost you a lot more.

A close-up of the Boox Mira Pro Color monitor with a YouTube video playing on screen.

While devices like the Remarkable Paper Pro use E Ink’s more advanced Gallery 3 screen technology that can display over 50,000 colors, its slow refresh rate makes it problematic for desktop computing tasks like browsing the web. The Mira Pro Color instead uses a 3,200 x 1,800 E Ink Kaleido 3 panel that can only display a limited palette of 4,096 different colors. Its refresh capabilities also can’t match the best LCD or OLED screens, but Boox offers four customizable display modes balancing image quality and performance that’s fast enough to watch videos.

The Mira Pro Color features front lighting similar to what’s found in e-readers with the option to adjust the color temperature to warmer hues at night. It’s also got a pair of speakers, a dedicated refresh button on the side to remove ghosting, and five ports including HDMI, mini HDMI, USB-C, and DisplayPort.

If you’re a gamer, photo or video editor, or someone who uses their desktop monitor for watching TV and movies, the Mira Pro Color isn’t for you. But if you spend your days editing text, writing, or crunching numbers in spreadsheets in a space with plenty of light, an E Ink monitor might be worth considering, particularly if you find yourself frequently dealing with eye strain.


From The Verge via this RSS feed

24
 
 

Google is close to striking a deal with Apple to integrate Gemini into the iPhone. During the search monopoly trial on Wednesday, Google CEO Sundar Pichai confirmed the company expects to strike a Gemini deal with Apple by the middle of this year and suggested it would roll out by the end of 2025.

The integration would presumably allow Siri to call on Gemini to answer more complex questions, similar to the integration that Apple launched with OpenAI’s ChatGPT. Apple senior vice president Craig Federighi hinted at plans to build Gemini into its Apple Intelligence feature last June, when the AI service was first announced. “We want to enable users ultimately to choose the models they want, maybe Google Gemini in the future,” Federighi said at the time.

Today’s update came in response to questions asked by Department of Justice lawyer Veronica Onyema. Pichai answered “Correct” when asked about these plans. He also said he had met with Apple CEO Tim Cook to discuss AI development. “He [Cook] was trying to understand our plans for how we’re evolving AI technologies, our roadmap, and as part of that, we talked about the Gemini app distribution as well,” Pichai said. He added that he was told by Cook that more third-party AI models would ship on Apple Intelligence later this year.

There are other signs that this integration is on the way. In February, MacRumors contributor Aaron Perris found references to “Google” as an Apple Intelligence model choice in an iOS 18.4 beta. With Apple’s ChatGPT integration, Siri automatically asks for permission to use the third-party AI model to answer more complex questions. ChatGPT can also analyze photos and generate images based on text.


From The Verge via this RSS feed

25
 
 

Duolingo is “more than doubling” the number of courses it has available, a feat it says was only possible because it used generative AI to help create them in “less than a year.”

The company said today that it’s launching 148 new language courses. “This launch makes Duolingo’s seven most popular non-English languages – Spanish, French, German, Italian, Japanese, Korean, and Mandarin – available to all 28 supported user interface (UI) languages, dramatically expanding learning options for over a billion potential learners worldwide,” the company writes.

Duolingo says that building one new course historically has taken “years,” but the company was able to build this new suite of courses more quickly “through advances in generative AI, shared content systems, and internal tooling.” The new approach is internally called “shared content,” and the company says it allows employees to make a base course and quickly customize it for “dozens” of different languages.

“Now, by using generative AI to create and validate content, we’re able to focus our expertise where it’s most impactful, ensuring every course meets Duolingo’s rigorous quality standards,” Duolingo’s senior director of learning design, Jessie Becker, says in a statement.

The announcement follows a recent memo sent by cofounder and CEO Luis von Ahn to staff saying that the company would be “AI-first” and that it would “gradually stop using contractors to do work that AI can handle.” AI use will now be evaluated during the hiring process and as part of performance reviews, and von Ahn says that “headcount will only be given if a team cannot automate more of their work.”

“Our vision has always been to use technology to teach as well as a human tutor, and because of AI, that goal is within our reach for the first time ever,” spokesperson Sam Dalsimer tells The Verge in response to questions sent following von Ahn’s memo. “We’ve already been moving in this direction, and it has been game-changing for our company. One of the best decisions we made recently was replacing a slow, manual content creation process with one powered by AI, under the direction of our learning design experts. That shift allowed us to create and launch 148 new language courses today.”

Dalsimer says that Duolingo has been “using and testing AI for years” and that it “would not be rolling out new course content and this AI-first company strategy if we were not confident in AI.” Duolingo is also “constantly testing and improving our models” and has “systems in place” to ensure that output from AI meets its standards for course content and is aligned with the Common European Framework of Reference for Languages, a standard for language proficiency.

Dalsimer says that “internally, many teams are already embracing and using AI in their work, and have been for years.” However, Dalsimer acknowledges that there have been “negative reactions” to von Ahn’s memo. Dalsimer also notes that Duolingo has “no intention to reduce full-time headcount or hiring” and that “any changes to contractor staffing will be considered on a case-by-case basis.”


From The Verge via this RSS feed

view more: next ›