The Verge

Valve has been building up to this for months, first in preview, then beta, and now finally in a full stable release: a new version of SteamOS that brings brand new features to the Steam Deck and supports third-party handhelds like Lenovo’s Legion Go and Asus’s ROG Ally series.

SteamOS 3.7.8 is the first stable release to add official support for the Lenovo Legion Go S, and the first to offer an official “recovery” image that lets you install SteamOS on other AMD handhelds like the ROG Ally and original Legion Go, as Valve promised us in January of this year.

Until now, those who wanted to try SteamOS on those Windows handhelds could shoehorn the Steam Deck’s original recovery image onto their device, or opt for a similar experience like Bazzite. But now, Valve explicitly provides instructions for getting that image working on a Legion Go or ROG Ally, even as it warns that “support for all devices that is not officially ‘Powered by SteamOS’ is not final.”

But even if you don’t care about rival handhelds, SteamOS 3.7 has a lot of improvements for the Steam Deck. I’ve been asking Valve to let me use my Bluetooth earbuds’ microphone since launch, and it’s finally happening in desktop mode (sadly not gaming mode yet) thanks to HFP/HSP profiles. You can turn on your Steam Deck LCD from across the room with a Bluetooth controller, a feature previously exclusive to the OLED model, which comes in handy when you’re docked to a TV.

Speaking of leaving it docked, you can now set a Battery Charge Limit to 80 percent for longevity’s sake if your handheld is always plugged in, you can frame-limit variable refresh rate displays, and control the P-state frequency of certain AMD CPUs.

It’s all underpinned by new versions of Arch Linux, new Mesa graphics drivers, a much newer version of the Plasma desktop mode, and more. Here’s the full SteamOS 3.7.8 changelog.

And, an updated SteamOS FAQ seems to suggest that Valve is ready to expand SteamOS beyond Lenovo and the Steam Deck. “We’re currently working with select partners on officially licensed Powered by SteamOS devices. Please reach out to us at poweredbysteamos@valvesoftware.com for more information about licensing SteamOS for your device,“ one answer reads.

But the industry is still waiting for Valve’s other shoe(s) to drop. Rumors are still heating up that we’re close to the launch of Valve’s wireless VR headset “Deckard,” and that it might be pointing the way towards a SteamOS-powered living room console too. Valve has been moving slow and steady, but it seems like a master plan for Steam Machines could finally be coming into focus.

From The Verge via this RSS feed

The Federal Trade Commission (FTC) has officially dropped its case against Microsoft’s acquisition of Activision Blizzard. The FTC filed an order to dismiss its complaint on Thursday, just days after it lost an appeal for a preliminary injunction to prevent Microsoft from finalizing its acquisition.

“The Commission has determined that the public interest is best served by dismissing the administrative litigation in this case,“ says the FTC in its filing. The filing brings an end to the FTC’s fight to try and block Microsoft’s $68.7 billion deal, nearly two years after it originally sought a temporary restraining order and injunction from a US federal district court.

Today’s decision is a victory for players across the country and for common sense in Washington, D.C. We are grateful to the FTC for today’s announcement. https://t.co/nnmUI76q0l pic.twitter.com/KgLxhZppx3

— Brad Smith (@BradSmi) May 22, 2025

Microsoft won its FTC fight to buy Activision Blizzard in July 2023, and the deal was completed several months later, in October, nearly two years after the acquisition was first announced. The FTC had appealed the ruling nearly two years ago, but an appeals court panel affirmed the denial of an injunction earlier this month.

“Today’s decision is a victory for players across the country and for common sense in Washington, DC,” says Microsoft vice chair and president Brad Smith. “We are grateful to the FTC for today’s announcement.”

From The Verge via this RSS feed

Warhammer 40,000: Boltgun – Words of Vengeance is a new typing game based on Warhammer 40,000 and it’s available now for free on PC via Steam. The dark sci-fi spin on a typing teacher was shadow-dropped during a series of announcements made during the latest Warhammer Skulls Showcase. Also shown was a trailer for Boltgun 2, a sequel to Auroch Digital’s first-person shooter the new typing game is based on.

Words of Vengeance follows in the footsteps of games like Typing of the Dead where you type on-screen prompts as quickly and as accurately as you can to defeat enemies and progress in on-rails levels. It’s set in the world of Warhammer 40,000: Boltgun and uses the same pixelated 3D aesthetic with plenty of blood and violence, except the blood is spilled by typing.

As a fan of mechanical keyboards, I personally love occasionally firing up typing tests to hear what the latest keyboard I’m obsessing over sounds and feels like. But as much as I love Monkeytype or Keybr, an actual typing game offers a different kind of fun. The thought of going full boomer-shooter with my proper grammar and punctuation exercises just has me more stoked than ever. Plus, it’s free.

Brother! It’s time to get those fingers moving. More blood for the blood god!

From The Verge via this RSS feed

The US Senate voted today to strip California of its authority to enact state limits on tailpipe pollution that are tougher than national standards.

Opponents called it an illegal move by Republican lawmakers that flies in the face of Senate rules and norms.

Senate Republicans went forward with the vote today anyway, revoking waivers that allow California to pursue its climate goals and improve air quality by reducing emissions from cars and trucks.

Opponents say it’s an illegal move

“This is the easy way to do what the fossil fuel industry wants,” Sen. Sheldon Whitehouse (D-RI) said on the Senate floor during deliberations that ended past 1:00AM ET. “They had this quick and dirty, sneaky maneuver that they could pull off so they didn’t have to negotiate, they didn’t have to legislate, and they didn’t have to use regulatory process.”

The US Environmental Protection Agency (EPA) has granted California waivers to set its own rules for car and truck emissions since the Clean Air Act was adopted in 1970. The standards the state sets for vehicle manufacturers can influence the entire industry because California is one of the biggest car markets in the world. Seventeen states and Washington, DC have also adopted all or part of California’s vehicle emission regulations.

The EPA had previously issued waivers to California approving its plans to require an increasing number of medium and heavy-duty vehicles sold in the state to be zero-emission, and all passenger vehicles sold to be zero-emission by 2035. Taken together, the resolutions passed today attempt to hobble one of the most ambitious plans to tackle climate change in the nation by rescinding those approvals. A third resolution also revokes a waiver the EPA had granted California to limit nitrogen oxide pollution from vehicles. House Republicans passed the measures a few weeks ago.

“California has used its waiver authority to push its extreme climate policies on the rest of the country,” Sen. Shelley Capito (R-WV) said in closing remarks on the Senate floor last night.

Environmental advocates, meanwhile, argue that GOP lawmakers attacked the state’s rights. “If other states don’t like California’s approach, they don’t need to follow it – but federal lawmakers shouldn’t be intervening to block states from providing cleaner air and a healthier environment,” Manish Bapna, president of the Natural Resources Defense Council said in a press statement today.

Both Senate and House Republicans used the Congressional Review Act (CRA) to revoke the waivers. The CRA allows Congress to overturn certain new rules with a simple majority vote and avoid a filibuster by the opposing party. But the Senate parliamentarian and Government Accountability Office — nonpartisan watchdogs — have previously found that the waivers aren’t considered recent rules within the parameters of the CRA.

The vote today sends the CRA resolutions to President Donald Trump to sign. Trump unsuccessfully tried to take away California’s authority to set its own tailpipe standards during his first term in office.

Auto trade groups haveopposed California’s plans to require more EV sales. “Disapproval of the rules is essential to ensuring a unified national vehicle marketplace that promotes continued progress on fuel economy while safeguarding economic growth and consumer interests,” Neil Bradley, executive vice president of the US Chamber of Commerce said in a letter to senators last week.

From The Verge via this RSS feed

The “One Big Beautiful Bill Act” that House Republicans narrowly passed early Thursday would strip state legislatures of AI oversight and scale back consumer protection and climate initiatives while funding border surveillance, among many other provisions.

The budget reconciliation bill still needs to be approved by the Senate, where some Republicans have voiced concerns with aspects of the text. But with President Donald Trump pushing for its passage into law, they could face an uphill battle in fighting for changes.

Here are some of the key tech and science provisions in the House version of the text:

Moratorium on state AI laws

States would be stripped of their power to enforce laws regulating artificial intelligence models and “automated decision systems” for 10 years under the budget package. That would likely preempt hundreds of AI-related bills being considered in 2025, as well as dozens that have passed into law — and on top of that, the broad “automated decision” language could nix regulating all kinds of computer systems not frequently classed as AI.

Republican supporters say the rule is necessary to let US companies innovate and keep up with rivals in China, and the idea has been promoted by OpenAI. More than 60 AI-related state bills have been enacted so far, according to the National Conference of State Legislatures (NCSL), many of which could be impacted by the proposed pause. The bills do everything from addressing algorithmic discrimination to regulating how AI can be used by government agencies.

Critics worry the definition could also hamstring laws covering all kinds of systems that feature automation or use machine learning. That might include rules championed by state-level Republicans, who have passed numerous social media regulations in recent years.

“Until we pass something that is federally preemptive, we can’t call for a moratorium”

A couple Republican senators have expressed concern over the moratorium. Sen. Marsha Blackburn (R-TN), eyeing a run for governor, spoke at a recent congressional hearing about her state’s AI law that seeks to protect a musician’s right to their voice’s likeness. “We certainly know that in Tennessee we need those protections,” Blackburn said, according to The Washington Post. “And until we pass something that is federally preemptive, we can’t call for a moratorium.”

Sen. Josh Hawley (R-MO), who publicly opposed Medicaid cuts in the House bill, also pushed back on the state law pause. “I would think that, just as a matter of federalism, we’d want states to be able to try out different regimes that they think will work for their state,” Hawley recently told Business Insider. “And I think in general, on AI, I do think we need some sensible oversight that will protect people’s liberties.”

The provision could also face a challenge in overcoming the “Byrd rule,” which bars “extraneous” additions in reconciliation bills.

Cuts to green energy tax credits

Biden-era tax credits for electric vehicles would be deprecated within two years if the House package is signed into law, and wind and solar energy credits would be phased out by 2032. The slashed credits include a $7,500 credit for purchasing eligible EVs, or $4,000 for an eligible used one, as well as credit for home refueling infrastructure.

Updates shortly before the vote also rolled back key climate programs from the 2022 Inflation Reduction Act even further than the bill  initially did, though they also pared down an effort to roll back credits for nuclear reactors.

Scaling back funding for consumer financial protection

The Consumer Financial Protection Bureau (CFPB), which had already been decimated by Elon Musk’s Department of Government Efficiency (DOGE), would see its funding capped further under the bill. House Republicans seek to cap the amount it can receive from the Federal Reserve at 5 percent of the system’s total operating expenses, rather than the current 12 percent. That would reduce the resources the consumer protection agency would have to respond to consumer complaints over things like imminent foreclosures and credit card fraud, and regulate digital payments services.

Border tech funding

The bill includes billions of dollars to lock down US borders, including $12 billion to reimburse states for border security. In addition to the $46 billion it would invest to build and “modernize” the wall between the US and Mexico, the bill would also provide $1 billion in funding for technology to detect drugs and other contraband being brought across the border. Another $2.7 billion would go toward surveillance systems that House Homeland Security Republicans described as “ground detection sensors, integrated surveillance towers, tunnel detection capability, unmanned aircraft systems (UAS), and enhanced communications equipment.”

Limiting gender-affirming care

Health care plans beginning in 2027 that are purchased through the Affordable Care Act marketplace would be barred from offering gender-affirming care, including hormone therapy and surgery. Similarly, the bill would prohibit Medicaid from covering “gender transition procedures” for minors and adults while requiring coverage for detransition.

From The Verge via this RSS feed

The mystery is the point, but it’s also indicative of where AI hardware is.

The last 48 hours have been a wild rollercoaster ride for AI hardware. On Tuesday, Google ended its I/O keynote - a roughly two-hour event with copious references to AI - with its vision for Android XR glasses. That included flashy partnerships with Gentle Monster and Warby Parker, as well as the first hands-on opportunity with its prototype glasses for the developers and the majority of tech media alike. On the ground, it was among the buzziest things to come out of Google I/O - a glimpse of what Big Tech thinks is the winning AI hardware formula.

A day later, Jony Ive and Sam Altman kicked down the door and told Google, "Hold my beer."

If you've somehow missed the headlines, OpenAI CEO Sam Altman revealed that the company was buying Ive's AI hardware startup for $6.5 billion. That alone was enough to set the tech media sphere ablaze. After all, Ive is the legendary figure behind the iPhone and Apple Watch's iconic design, revered for his relationship to Steve Jobs. Altman is not only the most recognizable figure in this new AI era, he's also frequently compared to Jobs himself. It's a narrative that writes itself. But for gadget nerds, the real nugget was the tidbit that Altma …

Read the full story at The Verge.

From The Verge via this RSS feed

Once Donald Trump won office for a second time, the writing was on the wall.

Trump ran on many messages, most of them confusing and contradictory, but one of his loudest and clearest messages was to end President Joe Biden’s “EV mandate.” It made no difference there never was a mandate, just a series of policies designed to encourage car companies to make more zero-emission vehicles and consumers to buy them — Trump was gunning for EVs.

As soon as he took office, he started signing a flurry of executive orders laying out exactly how he would start dismantling Biden’s legacy. And chief among them was an order to eliminate all of his predecessor’s electric vehicle policies, including weakening Biden’s tailpipe pollution rules.

But the orders were never going to be enough. It would take an act of Congress to unwind all the many tax credits and incentives designed to spur the sale of EVs by making them more affordable to a broader swath of the population.

So today, Congressional Republicans got to work. The House passed a bill to end the Inflation Reduction Act’s tax credits for clean energy, including the $7,500 EV tax credit. And the Senate voted — some would say “illegally” — to revoke California’s ability to set its own tailpipe emission rules, which are also followed by 17 other states and the District of Columbia.

Under the House proposal, most automakers would lose the EV tax credit right away — although those manufacturers who have yet to sell 200,000 EVs would get to keep the credit until the end of 2026. The bill would also eliminate the $4,000 tax credit for used EV purchases. And it would kill incentives for companies building solar, wind, and battery storage projects.

But wait, there’s more! If you already own an EV, the House bill would levy an annual $250 tax on your vehicle to help pay for road and infrastructure improvements. (Hybrid owners would have to cough up $100.) Internal combustion vehicle owners typically pay for road repairs through the federal gas tax of 18.4 cents per gallon — which, it should be noted, has remained exactly the same since October 1, 1993.

The House passed a bill to end the Inflation Reduction Act’s tax credits for clean energy, including the $7,500 EV tax credit

I understand the need for EV owners to pay for road repairs, but a flat tax is widely viewed by many smart people as the most regressive way to do it. The gas tax works because its a usage fee: the more you drive, the more gas you use, the more you pay into the Highway Trust Fund for repairs and improvements. A flat tax, which charges everyone the same regardless of usage, is much less desirable from a public benefits perspective.

A lot can still change as the bill now heads to the Senate for reconciliation. But it’s not looking good for our intrepid EV industry. Some lobbyists and advocates were holding out hope that Republicans with major clean energy projects in their district would push back against the attempt to kneecap the IRA. But as Heatmaps’ Matthew Zeitlin notes, the clean energy supporters were always the weakest faction amid a fractious GOP.

Where it goes from here is probably pretty bleak. The US was already woefully behind China and other developed nations in terms of clean energy investments. And now its likely to fall even further behind, perhaps permanently so. Major projects that were expected to take advantage of the federal incentives are now likely to die on the vine. That means fewer clean energy jobs, most of which were to be located in Republican-leaning districts.

It’s really a baffling political decision. Republicans are determined to cut off their own nose to spite their face, all because Trump ran on a false message of ending an EV mandate that never existed. In fact, Biden’s tailpipe emission rules were written in a way that acknowledges how passenger electric cars, specifically Tesla, have distorted the market thanks to their runaway success. So they crafted the rules so that passenger cars wouldn’t have to increase their miles-per-gallon numbers as rapidly as light trucks, as noted by Jalopnik’s Matthew Debord.

Whatever happens, the EV industry will persist. Automakers have poured billions of dollars into the shift to electric power, and they don’t want to let those investments go to waste. The focus should now be on the development of truly affordable EVs that can help spur mass adoption, as well as local infrastructure improvements to encourage more cycling and walking. But without hugely influential incentives from the federal government, it will be an uphill battle — spewing pollution and worsening the environment along the way.

From The Verge via this RSS feed

Nvidia has gone too far.

This week, the company reportedly attempted to delay, derail, and manipulate reviews of its $299 GeForce RTX 5060 graphics card, which would normally be its bestselling GPU of the generation. Nvidia has repeatedly and publicly said the budget 60-series cards are its most popular, and this year it reportedly tried to ensure it by withholding access and pressuring reviewers to paint them in the best light possible.

Nvidia might have wanted to prevent a repeat of 2022, when it launched this card's predecessor. Those reviews were harsh. The 4060 was called a "slap in the face to gamers" and a "wet fart of a GPU." I had guessed the 5060 was headed for the same fate after seeing how reviewers handled the 5080, which similarly showcased how little Nvidia's hardware has improved year over year and relied on software to make up the gaps.

But Nvidia had other plans.

Here are the tactics that Nvidia reportedly just used to throw us off the 5060's true scent, as individually described by GamersNexus, VideoCardz, Hardware Unboxed, GameStar.de, Digital Foundry, and more:

Nvidia decided to launch its RTX 5060 on May 19th, when most reviewers would be at Computex …

Read the full story at The Verge.

From The Verge via this RSS feed

Notepad’s new Write feature uses generative AI to create content for you based on a prompt. | Image: Microsoft

Microsoft is now testing a new feature in Notepad that can generate text for you using AI.  It’s part of a Windows 11 update being released to Windows Insiders in the Canary and Dev channels with Copilot Plus PCs. The update also includes new AI-powered features for Paint like a sticker generator, and improvements to the Snipping Tool that can automatically crop and resize screenshots based on what you’re selecting.

Notepad’s new Write feature can be used to “quickly draft text based on your prompt,” according to the Windows Insider Blog, “or build upon existing content with AI-powered assistance.” You can either right-click where you want new text to be inserted in a document or after selecting existing text you want to use as a reference.

After selecting the new Write feature in Windows’ Copilot menu, a prompt will appear where instructions can be entered. The AI-generated output will appear in the document but you’ll have the option to keep it, discard it if it’s not what you’re looking for, or refine the output by entering follow-up prompts. The feature was first discovered earlier this year in code included in test builds of Windows 11.

Write follows two other AI-powered tools Microsoft has been testing with Notepad. Summarize, first introduced last March, can generate a summary of the text in a document while Rewrite, which began testing last November, can adjust the tone of text, shorten or lengthen it, and rephrase sentences.

Joining AI features already being tested in Microsoft Paint like Generative Fill, Generative Erase, and the Cocreator tool that can create images based on text prompts and reference sketches, is a new Sticker generator. A new button in the Copilot menu opens a prompt where you can describe the sticker you want to create. Paint will then generate a small set of stickers based on your description that can be added to your canvas or copied over to other applications.

Paint is also getting a new Object select tool that relies on AI to create smart selections that takes most of the grunt work out of isolating and manipulating specific elements in a photo or painting.

To potentially make it easier to screenshot exactly what you need and immediately share it without edits, Windows 11’s Snipping Tool is gaining a new Perfect screenshot button on its toolbar. When framing a screenshot using the rectangle tool Perfect screenshot will use AI to “intelligently resize based on the content in your selection,” but you’ll still have the option to resize or reposition the selected region of your screen before capturing it.

It’s being joined by a new Color picker in the Snipping tool that lets you see the HEX, RGB or HSL color values below an eyedropper cursor, and it can be zoomed for more precision by either scrolling or using Ctrl +/- keyboard shortcuts.

You’ll need to be signed into your Microsoft account to use Notepad’s new Write feature which will use the same credits system as other AI-powered Windows 11 features do. Microsoft hasn’t yet announced what pricing for these credits will be if it eventually starts charging for them.

From The Verge via this RSS feed

House Republicans advanced a sweeping spending package that would roll back Biden-era tax credits for renewable energy projects. If the bill passes the Senate and makes it to President Donald Trump’s desk to sign, it could deal a serious blow to renewables, new nuclear technologies, and clean energy manufacturing across the US.

The rollbacks would undo much of the 2022 Inflation Reduction Act (IRA), which Democrats touted as the biggest investment in climate and clean energy initiatives. Losing these tax credits would slow efforts to build out enough new energy sources to meet rising electricity demand, as well as previous commitments the US has made on the international stage to help stop the climate crisis.

“This package is really economic malpractice,” says Brad Townsend, vice president for policy and outreach at the Center for Climate and Energy Solutions (C2ES). The bill that the House ultimately passed was even harsher on clean energy than a draft released last week. “The original version was bad. This version is worse.”

“This package is really economic malpractice.”

Based on the previous draft, C2ES and research firm Greenline Insights estimated that restrictions on which projects would be eligible for tax credits would cost hundreds of billions of dollars in lost GDP. An updated bill released overnight and passed early this morning could lead to even larger losses if the Senate ultimately passes it as-is.

Notably, the bill stipulates that projects must start construction within 60 days of it being enacted and placed in service by the end of 2028 in order to qualify for clean energy tax credits.

That would effectively make it impossible for new projects to qualify, given the long lead times needed to secure permits and financing before starting construction. During remarks on the Senate floor this morning, Senate Minority Leader Chuck Schumer (D-NY) called the provision a “clean job kill switch.”

“It’s one of the most devastating things added at the last minute in this bill snuck in the dark of night. And we in the Senate — and I hope our Republican colleagues will join us in this — are going to fight this every step of the way,” he said.

Nearly 977,000 jobs and $177 billion in GDP would have been lost as a result of requirements in the previous draft that stipulated that projects be placed in service by 2029 to qualify for credits, according to C2ES and Greenline Insights. Again, that draft was less stringent than the text that ultimately passed.

The bill seemingly includes a carveout for nuclear energy industry, to which some GOP members, including Secretary of Energy Chris Wright, have ties. Wright dialed into a meeting with Republican lawmakers on Wednesday night to discuss the tax credits, Politico reported. The bill subsequently says that new nuclear reactors would only have to commence construction by 2028 in order to qualify. But even though the provisions aren’t as strict for new nuclear projects to qualify, the bill still sets unrealistic goals. Next-generation nuclear reactors aren’t expected to be ready to deploy commercially until the 2030s.

The bill also ends an IRA policy that allowed renewable projects to transfer credits to one another, dealing another economic blow to developers outside of nuclear energy. It disqualifies projects owned by or receiving “material assistance from prohibited foreign entities.” Those restrictions are essentially unworkable, according to clean energy advocates and industry experts — considering that clean energy supply chains are still concentrated in China and that it could bar developers with investors from other countries. Restrictions on the involvement of foreign entities alone could lead to $237 billion in lost GDP, Greenline Insights and C2ES previously estimated.

Ironically, Republican districts stood to benefit the most from IRA incentives for new solar and wind farms and factories. Investments were concentrated in rural areas, and 73 percent of manufacturing facilities for clean power components are in red states, according to a recent industry report from the American Clean Power Association.

“Texas in particular is going to be hammered by the package as written,” Townsend says. His organization’s analysis found that Texas would lose the most jobs — more than 170,000 — from tax credit restrictions initially proposed in the bill.

“Texas in particular is going to be hammered.”

Fortunately, solar and wind power are already cheaper sources of electricity than fossil fuels in many cases and have been making steady gains in the US for decades thanks to falling costs. To be sure, developers now have to contend with new challenges posed by Trump’s tariff regime. But the industry has managed to make progress — now providing more than 20 percent of the US electricity mix — despite years of on-again, off-again credits prior to the IRA codifying incentives in a way that offered more long-term certainty for the industry.

What the tax credits in the IRA were supposed to help accomplish, however, was a dramatic ramp-up of carbon-free energy needed to stop the climate crisis. The IRA was expected to slash US greenhouse gas emissions by roughly 40 percent from peak levels by the end of the decade, according to independent analyses. That nearly got the nation to the goal that former President Joe Biden committed to under the 2015 Paris Agreement, which was cutting pollution by at least 50 percent by 2030. And since the US is responsible for more greenhouse gas emissions historically than any other country, the decisions that Congress makes now have consequences for the planet.

Trump, of course, has called climate change a hoax despite mountains of evidence showing how emissions from fossil fuels exacerbate floods, storms, droughts, fires, and other climate disasters.

Aside from worsening weather events putting pressure on the US’ aging power grid, the country is also grappling with a sudden rise in electricity demand from new AI data centers, crypto mining, electric vehicles, and increased domestic manufacturing. Electricity demand could grow by 25 percent by 2030, according to one forecast published this week by consulting firm ICF. By slowing the deployment of clean energy, the repeal of IRA incentives would lead to more pollution and raise household energy costs by up to 7 percent by 2035, according to a recent analysis by research firm Rhodium Group.

The Senate now has to wrangle with the entirety of Trump’s so-called “big, beautiful bill.” It also includes proposals to extend and expand income tax cuts, increase military spending, fund mass deportations, impose new restrictions on Medicaid and food assistance programs, and more. Even though the Republican-controlled Senate is likely to fall in line with Trump’s agenda, there’s still time for proposals in the bill to evolve.

In its current version, “Americans’ electric bills will soar. Hundreds of factories will close. Hundreds of billions of dollars in local investments will vanish. Hundreds of thousands of people will lose their jobs,” Abigail Ross Hopper, president and CEO of the Solar Energy Industries Association (SEIA), said in a press statement. But, Hopper added, “it’s not too late for Congress to get this right. The solar and [energy] storage industry is ready to get to work with the US Senate on a more thoughtful and measured approach.”

From The Verge via this RSS feed

Mozilla is shutting down Pocket, the handy bookmarking tool used to save articles and webpages for later. The organization announced that Pocket will stop working on July 8th, 2025, as Mozilla begins concentrating its “resources into projects that better match their browsing habits and online needs.”

Following the shutdown, you’ll only be able to export saves until October 8th, 2025, which is when Mozilla will permanently delete user data. Mozilla says it will start automatically canceling subscriptions as well, and will issue prorated refunds to users subscribed to its annual plan on July 8th.

It has also taken down the Pocket web extension and app as of May 22nd, 2025, but users who have already installed the app will be able to re-download it until October 8th.

Pocket — originally called Read It Later — launched in 2007 and grew in popularity as people used it to keep track of the articles, recipes, videos, and more that they planned to revisit. In 2015, Mozilla added Pocket to Firefox as the browser’s default read-it-later app, and then acquired it two years later.

Mozilla says it’s shuttering Pocket because “the way people save and consume content on the web has evolved.” Pocket’s email newsletter, called Pocket Hits, will continue under a new name, “Ten Tabs,” but it will no longer have a weekend edition.

In addition to shutting down Pocket, Mozilla is also sunsetting its fake reviews detector, Fakespot. “We acquired Fakespot in 2023 to help people navigate unreliable product reviews using AI and privacy-first tech,” Mozilla says. “While the idea resonated, it didn’t fit a model we could sustain.” Review Checker, the Fakespot-powered tool built into Firefox, is shutting down on June 10th, 2025, too.

“This shift allows us to shape the next era of the internet — with tools like vertical tabs, smart search and more AI-powered features on the way,” Mozilla says. “We’ll continue to build a browser that works harder for you: more personal, more powerful and still proudly independent.”

From The Verge via this RSS feed

At $479.99, the AirPods Max are now at their second-best price to date.

From savings on big-ticket items like LG’s C4 OLED TVs to smaller ones like the Garmin Forerunner 265, there are plenty of great Memorial Day deals to shop right now – many of which we’ve gathered in our roundup of the best Memorial Day deals. Here’s another good one: the AirPods Max with USB-C are currently available for $479.99 in multiple colors at Amazon, Best Buy, and Walmart – their best price since January, and just $30 shy of their all-time low.

The AirPods Max are among our favorite wireless headphones, delivering excellent sound quality (the recent addition of lossless audio makes them even better). Their active noise cancellation comes close to matching the best from Bose and Sony, while their natural-sounding transparency mode makes it easy to stay aware of your surroundings without having to take them off to hear what’s going on.

They’re also built to impress, with a stainless steel and aluminum design that feels as high-end as it looks, and is a step up from the mostly plastic builds of Apple’s rivals. We especially recommend them for those deeply embedded in Apple’s ecosystem, as they offer one-touch pairing, automatic device switching, spatial audio with head-tracking, and built-in Find My support – a combination of features you won’t find in any other pair of headphones.

Read our review of the AirPods Max with a Lightning port.

From The Verge via this RSS feed

Meta is beginning to test showing 3D photos on Instagram if you’re looking at your feed on a Meta Quest VR headset.

“Thanks to our AI view synthesis algorithms, we can transform the existing photos that show up in Instagram feed pixel by pixel – no fancy 3D cameras required,” Meta says in a blog post. “That means flat photos that weren’t originally captured in 3D will automatically be converted into an immersive format that gives 2D images a sense of depth when you view them on Quest.”

The test starts this week. “Not everyone will have access to 3D Instagram photos on Quest – and even if you’re in the test group, you may not see it right away,” Meta says. And it’s not the first time Meta has dabbled with 3D photos; the company rolled out a 3D photo feature for Facebook in 2018.

The test of 3D photos in Instagram is being introduced as part of Meta Quest’s v77 update, which includes a bunch of other features. One notable one, also in testing, is Navigator, which Meta describes as “a new home for your games and apps, friends, notifications, essential system settings, and more on Quest.”

You’ll get to the Navigator by pressing the Meta or Oculus button on the right controller. It offers access to recently-used apps, and you can pin up to 10 items to make them easier to get to. Meta has more information about Navigator in a video on its website.

The v77 update also adds experimental support for Bluetooth Low Energy audio devices, lets you select a window that can follow you as you move, and combines the Meta Quest Link, casting, and remote desktop into a single PC app, Horizon for PC.

From The Verge via this RSS feed

What Mission: Impossible - Dead Reckoning Part Onelacked in narrative cohesion, it made up for by leaning into the reality that larger-than-life spectacle and Tom Cruise's enthusiasm for doing his own stunt work have always been the franchise's main draw. Even though the film's artificial intelligence-focused plot felt a little shaky, its action was thrilling, and it was obvious that Cruise and director Christopher McQuarrie were trying to guide IMF agent Ethan Hunt's overarching story toward a conclusion that would satisfy longtime fans.

It was also very clear that Dead Reckoning was just the first half of something bigger and even more ambitious. And while Paramount's original plan to release a sequel shortly after Dead Reckoning's debut ended up being waylaid by the 2023 SAG-AFTRA strike, there is something kind of poetic aboutMission: Impossible - The Final Reckoningarriving at a time when the push to incorporate AI into almost every aspect of our lives has become far more intense.

The Final Reckoning is a shaggier and sillier film than its predecessor - one that feels like it has given up any pretense of being about intriguing spycraft in favor of big set pieces design …

Read the full story at The Verge.

From The Verge via this RSS feed

If you’ve ever wondered how the Internet Archive uploads all the physical documents on its site, now you can get a behind-the-scenes look at the process. The Internet Archive launched a new YouTube livestream that shows the digitization of microfiche in real time — complete with some relaxing, lo-fi beats.

Microfiche is a sheet of film that contains multiple images of miniaturized documents. It’s an old form of storing newspapers, court documents, government records, and other important documents. The Internet Archive uses these microfiche cards to digitize and upload documents to its online library.

The livestream shows a close-up look at one of the five microfiche digitization stations at the organization’s Richmond, California location, along with a look at the document that it’s working on. App maker Sophia Tung, who created a LoFi music livestream showing Waymo’s robotaxis returning to their parking lot, also set up the microfiche livestream for the Internet Archive.

“Operators feed microfiche cards beneath a high-resolution camera, which captures multiple detailed images of each sheet,” Chris Freeland, the Internet Archive’s director of library services, writes in a post on the site. “Software stitches these images together, after which other team members use automated tools to identify and crop up to 100 individual pages per card.”

From there, the Internet Archive processes the pages, makes them text-searchable, and then uploads them to its public collections.

The livestream runs from Monday through Friday from 10:30AM ET to 6:30PM ET. “During the day, you’ll see scanners working on custom machines to digitize all the microfiche in the world,” Tung says. “During the off hours, you can also see everything else that the Archive has to offer, like silent films in the public domain or historical pictures from NASA.”

From The Verge via this RSS feed

A source tells Bloomberg that Apple’s glasses will be similar to Meta’s but “better made.”

Apple is planning to debut its first pair of smart glasses next year, according to a report from Bloomberg. The upcoming glasses will reportedly come with cameras, microphones, and speakers, “allowing them to analyze the external world and take requests via the Siri voice assistant,” Bloomberg says.

The glasses would also be capable of taking phone calls, controlling music playback, performing live translations, and offering directions. They’ll also reportedly feature an in-house chip, though plans for incorporating augmented reality still “remain years away.”

The rumor comes just one day after OpenAI CEO Sam Altman announced the acquisition of io, the AI hardware design company owned by former Apple designer Jony Ive. Altman and Ive plan to launch their first AI devices next year, and rumors suggest that it will be a screenless device with built-in cameras and microphones.

A source tells Bloomberg that Apple’s device will be similar to Meta’s Ray-Ban smart glasses, “but better made.” Meta sold more than 1 million pairs of its Ray-Ban smart glasses last year, while Google just announced that it’s working with Xreal, Warby Parker, Samsung, and Gentle Monsters to create AI smart glasses on its Android XR platform.

In addition to ramping up work on smart glasses, Bloomberg reports that Apple has scrapped plans to create a smartwatch with cameras and AI features, like Visual Intelligence. Apple is still working on AirPods with cameras, Bloomberg says.

From The Verge via this RSS feed

On the bright side, at least this isn’t called Bard.

Google executives took the stage this week at I/O to unveil their latest AI technology: Deep Think. Or was it Deep Search? Then there's the new subscription plan, Google AI Pro, which used to be Gemini Advanced, plus the new AI Ultra plan. Then there's Gemini in Chrome, which is different from AI Mode in search. Project Starline is now Google Beam, there are Gems and Jules, Astra and Aura… you get the idea. The products overlap in confusing ways, the naming conventions are diabolical, and I'm begging Google to return some semblance of sanity to its product line before we all lose our DeepMinds.

In Google's defense, at least we're not calling any of these things Bard. That was Google's original name for its AI chatbot during the Great Chatbot Rush of 2023. OpenAI shipped ChatGPT, and apparently Google decided it had to ship something before there was time to consider not naming it Bard. The company corrected that mistake and went with Gemini, folding in Duet along the way. This was all a very good idea.

This week's Google I/O made it clear that the naming conventions are out of control again. There's Gemini 2.5 Pro Deep Think, which is designed to reason through complex math and …

Read the full story at The Verge.

From The Verge via this RSS feed

Blue Prince’s planned 1.10 update, which will add things like a color assist mode, controls remapping, and more, will be the “final major update” and the “definitive version” of the game, director Tonda Ros says in notes for the 1.04.5 patch that was released today.

“It had always been my dream to release a definitive version of the game at launch,” Ros says. “I’m a big fan of complete standalone experiences and it was never my plan to continually tinker with the game with regular content updates, room rebalancing, or DLC and the like. Considering this is my first game, I’m extremely proud of everything I managed to include for launch, but there are a few minor things that were not quite completed before our release date that I feel must be added to fulfill that original vision.”

Here’s what you can expect in that 1.10 update, according to Ros:

Among the planned inclusions in this final update: an overhyped arcade game, a curious house cat, accessibility features like color assist mode, controls remapping, cursor size/opacity settings, and widescreen support. Some UI/UX improvements, more variations for end-of-day manor descriptions and accompanying house illustrations, HUD display customization, final room/item balancing, and the addition of a handful of extra cinematics that I am still working on (the unlockable challenge modes actually have their own unique openings and endings!)

However, Ros notes that there will be bug fixing patches released “intermittently” in the “months leading up to 1.10.” Ros says that some updates may include some of the features listed for 1.1o; “we wouldn’t want to sit on accessibility features for months on end while we wait on cinematics!” And Ros says the team will be “continually” committed to fixing bugs “for far, far into the future.”

The update released today, 1.04.5, includes a fix for a save rollback issue on PS5.

From The Verge via this RSS feed

Say hello to Claude Opus 4 and Claude Sonnet 4.

Anthropic has introduced Claude Opus 4 and Claude Sonnet 4, its latest generation of hybrid-reasoning AI models optimized for coding tasks and solving complex problems.

Claude Opus 4 is Anthropic’s most powerful AI model to date, according to the company’s announcement, and capable of working continuously on long-running tasks for “several hours.” In customer tests, Anthropic said that Opus 4 performed autonomously for seven hours, significantly expanding the possibilities for AI agents. The company also described its new flagship as the “best coding model in the world,” with Anthropic’s benchmarks showing that Opus 4 outperformed Google’s Gemini 2.5 Pro, OpenAI’s o3 reasoning, and GPT-4.1 models in coding tasks and using “tools” like web search.

Claude Sonnet 4 is a more affordable and efficiency-focused model that’s better suited to general tasks, which supersedes the 3.7 Sonnet model released in February. Anthropic says Sonnet 4 delivers “superior coding and reasoning” while providing more precise responses. The company adds that both models are 65 percent less likely to take shortcuts and loopholes to complete tasks compared to 3.7 Sonnet and they’re better at storing key information for long-term tasks when developers provide Claude with local file access.

A new feature introduced for both Claude 4 models is “thinking summaries,” which condenses the chatbots’ reasoning process into easily understandable insights. An “extended thinking” feature is also launching in beta that allows users to switch the models between modes for reasoning or using tools to improve the performance and accuracy of responses.

Claude Opus 4 and Sonnet 4 are available on the Anthropic API, Amazon Bedrock, and Google Cloud’s Vertex AI platform, and both models are included in paid Claude plans alongside the extended thinking beta feature. Free users can only access Claude Sonnet 4 for now.

In addition to the new models, Anthropic’s Claude Code agentic command-line tool is now generally available following its limited preview in February. Anthropic also says it’s shifting to provide “more frequent model updates,” as the company tries to keep up with competition from OpenAI, Google, and Meta.

From The Verge via this RSS feed

Tamagotchi Paradise introduces a new dial control letting you zoom in on characters to the cellular level. | Screenshot: YouTube

There are still plenty of Tamagotchi devices that have you focusing on raising just a single digital character, but the new Tamagotchi Paradise puts you in charge of a planet full of needy dependents. To make it a little easier to navigate its digital world, the device includes a full color screen and complements its three button controls with a dial on the side for zooming from a view of the entire planet right down to the cellular level of your virtual pets.

Launching in Japan first on July 12th, 2025, for $45, the Tamagotchi Paradise experience starts with you hatching an entire planet in an event called the Egg Bang. From there you can zoom in on your world to one of three different unique ecosystems called Tama Fields that include land, water, and sky. Which ecosystem you start with is dependent on which of three colored versions of the Tamagotchi Paradise you buy, but you’ll eventually be able to unlock all three Tama Fields as you play.

Each Tama Field will affect how the Tamagotchi characters grow and evolve through unique activities and foods. The game includes over 50 different Tamagotchis separated into 12 different species that can be raised from eggs, but Tamagotchi Paradise expands that with a breeding mechanic so two compatible characters can create new Tamagotchi with 50,000 possible combinations available. Two Tamagotch Paradise devices can also be physically connected, letting characters from each interact to potentially start a family or become foes that prank each other and even get in fights.

If caring for a planet full of virtual pets isn’t hard enough, the device’s dial even lets you zoom in on Tamagotchis to the cellular level, which is how you care for and cure them when they get sick. It sounds exhausting, but unlike kids, dogs, or cats, if the responsibilities become overwhelming, you can simply pop out the device’s AAA batteries for some respite.

From The Verge via this RSS feed

Honor has announced the 400 and 400 Pro, two midrange phones that launch today in the UK and Europe. They’re capable-looking handsets in plenty of respects, but stand out mostly for the guarantee of six years of software support, bested only by Google’s Pixel 9A for the price.

In fairness, at £699.99 / €799 (around $900), the 400 Pro is really a flagship in its own right. It’s powered by 2023’s Qualcomm Snapdragon 8 Gen 3 chipset, and boasts a 6.7-inch AMOLED screen, IP68 and 69 ratings, and sizable 5,300mAh battery (with an even larger 6,000mAh cell outside Europe). The triple camera is impressive too, with a 200-megapixel main shooter, plus an ultrawide and telephoto.

I’m a bigger fan of the regular 400 though, and not just because at £399.99 / €499 (around $560) it’s substantially cheaper. It has straight sides, rather than the curved edges of the Pro model, and combined with a smaller 6.5-inch display it’s much more comfortable to use.

The 400 ships with the same main and ultrawide cameras as the Pro, only giving up on the telephoto. Its IP65 water-resistance is a little less comprehensive, and the Snapdragon 7 Gen 3 won’t offer quite as much power, but the combination of the same large battery and a smaller screen should give this great endurance. The main downside for me would be giving up the Pro’s wireless charging.

Importantly, Honor’s commitment to six years of OS version updates and six years of security patches is the same for both phones, and should see them through to Android 21 in 2031. That matches Samsung’s promise for its Galaxy A56, and falls just one year short of the seven years guaranteed for the Pixel 9A. Both 400 phones arrive running Honor’s MagicOS and Android 15, and include a unique AI image-to-video feature powered by Google’s Veo 2 model, currently not available on any other phones.

While I’m not entirely sold on the 400 Pro’s near-flagship price tag, the base 400 looks like a compelling alternative to the Pixel 9A and Galaxy A56. It’s cheaper than either, should last for about as long, and bests both on quite a few specs.

Photography by Dominic Preston / The Verge

From The Verge via this RSS feed

Verizon has asked the Federal Communications Commission to get rid of the rule requiring it to unlock phones after 60 days. In a letter to the FCC spotted by LightReading, Verizon claims the current unlocking requirement “benefits bad actors and fraudsters.”

The FCC first imposed an unlocking requirement following Verizon’s purchase of C-Block spectrum in 2008. It forced Verizon to allow customers to change to a new cellular carrier after purchasing a phone from the company, making it easier to switch away from than other providers.

After some pushback from Verizon, the FCC allowed the carrier to lock phones for 60 days after purchase in 2019. Verizon later extended the policy to TracFone following its acquisition of the prepaid provider in 2021.

But now, Verizon wants the 60-day period extended even longer, calling the FCC’s current requirement “outdated regulation that has become both burdensome and harmful.”  The company also says eliminating the rule aligns with the FCC’s recent initiative to get rid of “unnecessary” regulations.

It adds that “recent industry experience shows that even a lock of 60 days does not deter device fraud,” which is why the “industry standard” for providers who don’t have to abide by the 60-day unlocking rule is a minimum of six months.

“Waiving this rule will benefit consumers because it will allow Verizon to continue offering subsidies and other mechanisms to make phones more affordable,” Verizon says. “Waiving the rule also will benefit competition because it will eliminate the distorted playing field that currently exists.”

Last year, the FCC proposed rules that would require all providers, including T-Mobile and AT&T, to unlock phones after 60 days of activation.

From The Verge via this RSS feed

With access to Gemini-powered features on Nest Cams like this one, app developers could tap into more advanced automations.

Google is bringing Gemini intelligence to its Home APIs, allowing smart home developers and manufacturers to tap into Gemini’s AI-powered features and potentially making your smart home a lot, well, smarter. The company announced the news in a blog post during the Google I/O developers conference this week.

Last summer, Google added several new Gemini-powered capabilities to its own platform, and now it’s extending those to others. This includes access to AI-powered camera analytics from Google Nest cameras and creating smart home routines using conversational language.

The company also announced new triggers for automations, including date and weather. Developers with access to the APIs will be able to pull these features into their apps, meaning that, as a user, you might experience them without ever opening the Google Home app.

For example, access to the AI-generated descriptions and search capabilities of Nest cameras could allow a company like ADT (which is already using Google Home APIs) to bring these features into its own home security app. So a user could ask questions like, “Did the kids leave their bikes in the driveway?” In addition to the AI features, developers can also pull livestreaming, event history, two-way talk, and camera settings for Google’s Nest cameras into their apps.

With Gemini in Home APIs, developers can allow users to set up smart home automations and routines within their apps with conversational language; they can just say what they want to do. Gemini can also proactively suggest automations by analyzing the devices in a user’s home, according to Google.

All of this should enable companies to offer more advanced smart home control within their own apps, potentially eliminating some of the more complicated and confusing setup processes and allowing users to get more out of their smart home devices without spending lots of time manually configuring them.

Google originally launched its Home APIs at I/O last year as a way for developers to tap into Google Home-compatible devices and automations. This includes any devices connected through Matter, which boosts the number of devices Google Home can control to over 750 million, according to the blog post.

Some companies, including ADT, Yale, and Eve, got early access to those APIs. Another, First Alert, recently launched a smart smoke alarm that uses Home APIs to interconnect with Google’s Nest Protect smoke alarm. Google has stopped manufacturing the Protect, but as it can continue to work for at least 10 years, the First Alert device could help homeowners maintain interconnected smoke alarms as they upgrade their system.

This week, Google announced that smart lighting manufacturer Cync is integrating Google Home APIs, allowing users to control any device they have synced to Google Home in the Cync app. Smart home platform Tuya Smart is also working with the APIs, and Google said iRobot is leveraging the APIs to integrate Google’s Home and Away feature into its app. This would allow select Roomba vacuums to automatically start cleaning when you leave home — an option Roombas have had with Amazon Alexa.) for a while.

Developers who use the Home APIs will be able to control those devices and experiences through the Gemini smartphone app and Nest smart speakers, with support for smart displays, Google TV, and the Pixel Tablet coming later this year, according to Google. They also teased a new Home Summary widget for the Pixel Tablet.

Most of those devices are controllable by voice, and while Google Assistant has been officially retired on smartphones in favor of Gemini, it’s still around on smart speakers and displays. However, the company has begun rolling out a “new experience” powered by Gemini to its Nest Audio and Nest Mini 1st and 2nd-gen speakers for users in its Public Preview program who have a Nest Aware subscription. It sounds like the rest of the hardware lineup won’t be far behind.

But don’t expect to see these new Gemini-powered features in loads of third-party apps soon. According to Google, the Gemini in Home APIs will be available via an early access program “to a small group of partners” later this year.

From The Verge via this RSS feed

When Microsoft CEO Satya Nadella ran into Meta's former engineering chief, Jay Parikh, at a conference last summer, he had the future of AI top of mind. The pair have known each other for around 15 years, but this meeting was different, and Nadella called Parikh shortly after bumping into him to dig into what was really on his mind.

"We were chatting about the future and chatting about all the stuff he needs to do here and that the team needs to do around AI," Parikh, now head of Microsoft's CoreAI team, tells The Verge. "That's when he said, 'Hey, why don't you come join and help me transform the company around all of this AI stuff?'"

Nadella regularly talks about Microsoft being part of a new AI era, but he now wants the company to overhaul how it builds software to meet this new era head-on. Parikh, who transformed Facebook engineering teams, now leads a transformation that he describes as building an AI "agent factory" for Microsoft's customers.

"I described this agent factory idea to Bill [Gates], not knowing that he and Paul [Allen] described Microsoft 50 years ago as the software factory," Parikh says. "Just like how Bill had this idea of Microsoft being a bunch of softw …

Read the full story at The Verge.

From The Verge via this RSS feed

If you’ve been wanting to grab a subscription to Nvidia’s GeForce Now cloud game streaming service, now’s a good time to commit. That’s because six-month subs to its mid-tier Performance subscription — which offers ad-free 1440p gameplay, short queue times, plus ray-tracing on supported games — is 40 percent off. Instead of $49.99, you’ll pay $29.99, a reasonable price if you don’t have a gaming PC, or just want a reliable way to play many (but maybe not all) of your Steam, Epic Games Store, and other titles away from your PC.

Each tier of GeForce Now offers different session lengths; the Performance tier caps you at six hours of playtime before freeing up your virtual machine for another player. However, you can simply start a new session immediately after. This deal will be active through July 6th, and after your six months are up, you’ll be charged the full price unless you cancel beforehand.

This discount comes at an especially opportune time for Steam Deck owners, as the official GeForce Now should be launching soon for Valve’s handheld, although an exact date hasn’t been publicly confirmed. Now, it might seem kind of redundant to play Steam games via the cloud when you can simply load them onto a Steam Deck’s storage, but Nvidia’s virtual machines are more powerful. You’ll be gaining visual fidelity not possible on the Deck, but adding some input latency — a given with any and all cloud game streaming services.

From The Verge via this RSS feed