Feed aggregator

Production in Hollywood has been suffering. But it's unclear how a 100% tariff on movies produced outside the United States would work – or who it would help.

(Image credit: Mario Tama)

James.'/>

The 2025 Pulitzer Prizes were announced Monday afternoon. Percival Everett won the award for fiction for his novel James, a powerful re-imagination of Huckleberry Finn.

(Image credit: Henry Nicholls)

The question of OpenAI, its business, and intentions for the future of AI may finally be solved. In an open letter, OpenAI CEO and co-founder Sam Altman outlined plans to keep OpenAI running under the oversight of a non-profit. What's more, the profit side of the business is transitioning to a Public Benefit Corporation (PBC).

A PCB is notable because it means that while that portion of OpenAI will still be interested in making a profit, it will have a larger purpose, one that's intended to serve the good of society.

In more practical terms, Altman wrote, "We want to put incredible tools in the hands of everyone....We want to open source very capable models. We want to give our users a great deal of freedom in how we let them use our tools within broad boundaries, even if we don’t always share the same moral framework, and to let our users make decisions about the behavior of ChatGPT."

In recent years, former partner and OpenAI co-founder Elon Musk has sued OpenAI for leaving its non-profit roots behind, and others have voiced concern about OpenAI not open-sourcing key models. Altman previously admitted that he was on the wrong side of that argument, and Musk eventually lost his case.

Now, though, OpenAI and Altman seem to be moving in the direction Musk and the open-source critics want.

AI for the good

The change of heart comes as Altman admits that in the early days, "we did not have a detailed sense for how we were going to accomplish our mission" and also admitted that some at OpenAI back then thought AI "should only be in the hands of a few trusted people who could 'handle it'."

The perspective now, though, especially as Artificial General Intelligence (AGI) is on the horizon, is "We want to build a brain for the world and make it super easy for people to use for whatever they want," wrote Altman.

The go-forward plan is for OpenAI's non-profit to be "the largest and most effective nonprofit in history that will be focused on using AI to enable the highest-leverage outcomes for people."

Questions remain

Altman also wants to develop "beneficial AGI" and notes the importance of safety and alignment. "As AI accelerates, our commitment to safety grows stronger. We want to make sure democratic AI wins over authoritarian AI."

Altman's come quite a long way since he was suddenly ousted in late 2023 by, among others, Ilya Sutskever, formerly OpenAI's Chief Scientist and co-founder. He returned just days later. There's a sense in the new letter that AI and the coming AGI are bigger than one person, one company, and one AI like ChatGPT.

As for what this will mean for the future of OpenAI, ChatGPT, and AGI, it is unclear. The PCB may be focused on the public good, but it will still be interested in making a profit. How the non-profit overseer impacts that is unclear.

OpenAI says it will be talking to attorneys generals in California and Delaware, who helped it come to this decision, along with its biggest commercial partner, Microsoft (Copilot's base models are GPT-based), about the implementation of its new plan.

"We believe this sets us up to continue to make rapid, safe progress and to put great AI in the hands of everyone," wrote Altman.

We'll see.

You might also like

• I asked ChatGPT, Gemini, and other AIs how to combat aging, and only one did the right thing
• Goodbye, ChatGPT-4 – here's why the retired AI model ...
• The Godfather of AI is more worried than ever about ...
• I've tried Google's new AI Mode and now you can too

The annual Eta Aquarid meteor shower is set to peak early on May 6 and will be viewable in the dark predawn skies.

(Image credit: NASA/MSFC/B. Cooke)

After weeks of confusion, the Trump administration confirmed that it terminated visa records for thousands of international students because of past brushes with law enforcement, many of them minor.

(Image credit: Lisa Poole)

Come for the collectible, shark-mouth-shaped popcorn bucket, stay for the movie.

Here are hints and the answer for today's Wordle No. 1,417 for May 6.

Microsoft has announced that the pioneering online video calling service that's been around for more than two decades will go offline on Monday.

(Image credit: David Ramos)

The former-Microsoft exclusive is coming to Sony's PlayStation 5, so you can dive into a very, very mad world.

Improve your mood with these easy mental health habits.

• Google’s Gemini 2.5 Pro has officially completed Pokémon Blue
• The game ran as a livestream experiment by an independent engineer
• Gemini played the game with some light developer intervention, but mostly on its own

Google's Gemini AI may not have passed the Turing test yet, but it would be very popular in the schoolyard three decades ago after winning a game of Pokémon Blue. The Gemini 2.5 Pro is now both Google's most advanced AI model and a Pokémon Master, as demonstrated in a Twitch livestream called “Gemini Plays Pokémon” run by an engineer unaffiliated with Google named Joel Z. Even Google CEO Sundar Pichai joined the celebration, sharing a clip of the victory on X.

What a finish! Gemini 2.5 Pro just completed Pokémon Blue!  Special thanks to @TheCodeOfJoel for creating and running the livestream, and to everyone who cheered Gem on along the way. pic.twitter.com/E2pn3tpfEbMay 3, 2025

You might wonder why an AI model beating a thirty-year-old game drew so much attention. It's partly because of the spectacle, but also because of AI model rivalry. Back in February, Anthropic showcased the progress its Claude model was making in beating Pokémon Red. They used the game to show off Claude’s “extended thinking and agent training” and launched a “Claude Plays Pokémon” Twitch stream, inspiring Joel Z.

Before crowning Gemini as the one true AI Ash Ketchum, it’s worth noting a few caveats. For one, Claude hasn’t technically beaten Pokémon Red yet, but that doesn’t automatically make Gemini better, as they employed different tools, known as “agent harnesses.” The models don’t play the game directly like a human with a controller would. Instead, they’re fed screenshots of the game environment along with overlays of key information, then asked to generate the next best action. That decision is then translated into an actual button press in the game.

And Gemini hasn’t been going it entirely alone. Joel admitted he occasionally stepped in to make improvements, though he has made a point of doing so only to improve some of Gemini's reasoning. He also plans to continue working on the Gemini Plays Pokémon project to make further improvements.

Pokémon AI

(Image credit: Sundar Pichai/X)

What makes this more than a quirky internet stunt is what it implies about where AI is headed. Playing a game like Pokémon Blue isn’t about fast reflexes or memorizing controller inputs. It’s about long-term strategy, adapting to surprises, and navigating ambiguous challenges. These are all areas where AI usually needs improvement. That Gemini could not only hold its own but finish the game (with minimal nudging) suggests that models like it are getting better at extended strategy.

It's also the kind of milestone the average person can understand. You can intuitively understand what the AI is doing when bumbling through Lavender Town or misreading a battle tactic, and compare it to the choices you'd make in that context. Of course, you shouldn't overstate what this means. AI can now finish a game you probably beat in middle school, but it also highlights how much human effort still goes into making AI seem autonomous.

Whether or not Claude or Gemini become true Pokémasters doesn't matter so much as what they're playing means for AI's development. Showing that AI won't just crunch numbers or generate spam emails could change how people think of what AI can do, even with help. And if this is how AI models start learning how to operate in unpredictable, open-ended environments, well, beating Mewtwo might just be a stepping stone to something a lot more profound. Or at least, a bit more productive.

You might also like

• AI video tool recreates Super Mario Bros. but it's so glitchy
• Google Gemini could soon get a super-useful 'Power up' button – here's what it does
• Virtual versions of 11,000 real college football players are taking the field for EA's 'College Football 25'

The $49 Pride Edition Sports Band (and matching watch face) brings a unique element to the Apple Watch in celebration of the LGBTQ+ community.

Two people can share a subscription -- but this plan is only being tested outside the US for now.

• Nvidia’s RTX PRO 6000 offers 96GB of memory, enough for giant AI models and datasets
• With 24,064 CUDA cores, this GPU dominates deep learning, rendering, and scientific simulations
• At $8,200, this card is not for gamers; it’s built strictly for serious professionals

In the evolving world of professional computing, PNY Technologies has launched what might be the most powerful workstation GPU to date: the Nvidia RTX PRO 6000 Blackwell Workstation Edition.

Directdial reports the card is priced at a staggering $8,200, making it firmly aimed at professionals working in AI development, simulation, or high-end content creation rather than casual users.

At the heart of the GPU is Nvidia’s latest Blackwell architecture, delivering 24,064 CUDA cores to accelerate demanding workloads such as deep learning, real-time rendering, and scientific computing.

Blackwell architecture delivers massive power and memory

The RTX PRO 6000 Blackwell supports a record-breaking 96GB of GDDR7 memory, operating over a 512-bit bus with a bandwidth of up to 1.75TB/s.

This is achieved using 3GB modules configured as 16×2×3GB, enabling the vast memory pool necessary for handling massive AI models and ultra-high-resolution assets. ECC memory is also onboard to improve stability in mission-critical tasks.

Despite its performance, the card maintains a relatively modest 300W TDP and is considered energy-efficient for its class.

The GPU supports a wide range of APIs, including Vulkan 1.3, DirectX 12, and OpenCL 3.0.

Early PCB images suggest the absence of a 12V-6x2 connector, possibly pointing to a rear-mounted power input design more commonly found in servers or Max-Q setups.

However, a single 16-pin connector supports the current desktop version, which uses a PCIe 5.0 x16 interface and fits into a standard dual-slot, full-length layout.

Though technically a workstation GPU, the RTX PRO 6000 Blackwell opens new possibilities across a range of specialist fields. It can support up to four 8K displays and is engineered to meet the demands of professionals in VFX, CAD, and AI training environments.

You might also like

• Check out the best 3D modeling software for 3D printing and more
• We've rounded up the best portable monitors available now
• TSMC committed to Arizona chip plant ahead of potential tariff impact

It is a dark time for smartphone fans. The news seems glum. Apple is sinking deeper and deeper into trouble over its failure to deliver a satisfying Apple Intelligence package. Phone makers like Motorola launch brand new phones with an hour explaining the AI features, then forget to mention the phone itself.

Samsung finds itself holding hands with Google as it drops AI feature after AI feature – first it was just Circle to Search, but now Samsung has given Google Gemini the Bixby button?! Dark times indeed.

The worst part is that nobody asked for these features. I don’t want these AI features on my phone. I could already drop a screenshot image into Google Search, I didn’t need to draw a circle to search. I never looked at my iPhone 14 and thought, ‘Gee, I wish this phone could inaccurately summarize my notifications for me.’

For every great AI feature, like Google’s awesome call screening features, there are twice as many terrible AI features, like the image generators that are problematic on so many levels, or the news headline summaries that simply make up imaginary news.

Bad AI is distracting us from great phones

That’s sad, because if you took away this AI bloat, today’s phones are… really great?! Today’s Android phones have matured beyond most of the complaints I’ve held about Android: that it was too complex and lacked a coherent interface design.

To Apple’s defense, there are so many incredible features in iOS 18, especially the features that work between iPhones and bring iPhone people together, that it seems a shame Apple wasted so much of its billboard space on features that don’t even exist, yet, like the super-intelligent Siri that unfortunately failed to graduate in time.

The new Pixel 9 Pro is the most polished Pixel phone ever (Image credit: Philip Berne/ Future)

Take the Google Pixel 9 series, for example. Google has a great new design, and the phone is more durable than ever before. That means it’s less likely to break, and you’ll be able to keep it longer than phones in the past. To back that up, Google also gives you seven years of Android updates. The phone shipped with Android 14, and it should last through Android 21!

That’s just incredible. A few years ago, we were lucky to get any software update promised from an Android phone maker. Today, Google and Samsung both promise seven years of Android updates for their top phones, and even Qualcomm promises its Snapdragon 8 Elite chipset will be supported for the next eight years by the chipmaker.

If you haven’t considered Android, or if you left the platform years ago, it’s worth a new look. Google has removed most of the confusing customization options that cluttered the home screen and app drawer. The whole interface is clean and tidy, and easier to use than ever. If you want to get complicated, you can still download a third-party homescreen launcher app, but the basic Pixel version of Android is refreshing and simplified.

Apple is just as distracted as Android

If you haven’t tried Apple’s iOS in a while, there are amazing new features that let you share between iPhones. You can share your contact information, photos, or even music playlists to let friends add songs to the party mix. You do this just by bringing two iPhones close together, and the phones do the rest using a feature called AirDrop. It works like magic, and it even has a cool magical effect on the screen to show it’s working.

Apple has also added great safety features to the iPhone that let you check in with friends and family so they can know you’re safe. If you think that iOS is too simple, think again. Apple has made it easy to completely customize and rearrange your Control Panel, and the iPhone homescreen now has the same sort of widgets, folders, and layout options you’d expect from an Android phone.

(Image credit: Apple)

Best of all, Apple’s latest titanium build means the new iPhone is also more durable than ever before. Apple doesn’t promise seven years of iOS updates, but it has consistently delivered at least five years of iOS to every single iPhone, and recently, Apple has offered the latest software to iPhones that are even older.

While AI seems unavoidable, you can still ignore most of the latest AI features and just enjoy a great smartphone. Apple and Google are making top-notch phones in spite of their best AI efforts, so don’t let the AI marketing and buzzwords scare you away. The Pixel 9 Pro and iPhone 16 Pro are the best phones these companies have made. Hopefully, the quest for smarter AI doesn’t keep the hardware from improving as well.

You might also like

• Apple officially delays the AI-infused Siri and admits, ‘It’s going to take us longer than we thought’
• The Bella Ramsey Apple Intelligence ad that disappeared, and why Apple is now facing a false advertising lawsuit
• AI took a huge leap in IQ, and now a quarter of Gen Z thinks AI is conscious

Waymo plans to add 2,000 more autonomous vehicles to its fleet by 2026.

You don't need to call an electrician. Gradient's 120-volt All-Weather Heat Pump is easy to install on your own and plugs into a standard outlet.

As the pioneering video-calling service shuts down, Microsoft urges users to switch to Teams for similar features.

Apple is getting ahead of 2025 Pride Month celebrations with an early reveal of a wild new watch band and some vibrant wallpapers.

Pride Month, which happens every June and celebrates LGBTQ+ communities, is always a month that embraces an array of colors, but this month's Apple Pride Month Collection adds a twist and nod to the "individuality of all members of the LGBTQ+ community," says Apple in a release.

Apple explained that the rainbow colors within each Sport Edition band start as individual color stripes. The bands are assembled by hand and compressed into their final shapes. Apple claims that this means, just like people, no two bands will be alike.

(Image credit: Apple)

Those colors can also be found on the Pride Month Dynamic Apple Watch face and with special wallpapers for the iPhone and iPad. Colors and bands will move on the screen as the users and wearers move.

The watch face and wallpapers are free, but the band, available in small-to-medium lengths, medium-to-large lengths, and in 40mm, 42mm, and 46mm watch face widths, will cost $49.

The band goes on sale next week, and the wallpapers and watch face will arrive soon with platform updates in iOS 18.5, iPadOS 18.5, and watchOS 11.5.

This latest Pride Band and face arrive as Apple is celebrating 10 years of the Apple Watch, a wearable that has become more than just a timepiece but also an important platform for supporting health, wellness, and fitness (along with personal style and maybe some social consciousness).

Apple's decision to move forward with a Pride Month collection in the US is notable as some major tech companies (looking at you, Google and Amazon) have scaled back Diversity Equity and Inclusion (DEI) efforts with moves that impact the LGBTQ+ population.

Apple has not done so and seems to be signaling that it will continue its support of diversity, equity, and inclusion. The release includes a line noting that "Apple is proud to financially support organizations that serve LGBTQ+ communities."

You might also like

• I've been wearing an Apple Watch for 10 years, and it ...
• Here's why the Apple Watch Series 10 is now my favorite ...
• I've been an Apple Watch user for 10 years

• Google claims its Ironwood TPU is 24x faster than the El Capitan supercomputer
• An analyst says Google's performance comparison is “perfectly silly”
• Comparing AI systems and HPC machines is fine, but they serve different purposes

At the recent Google Cloud Next 2025 event, the tech giant claimed that its new Ironwood TPU v7p pod is 24 times faster than El Capitan, the exascale-class supercomputer at Lawrence Livermore National Laboratory.

But Timothy Prickett Morgan of TheNextPlatform has dismissed the claim.

"Google is comparing the sustained performance of El Capitan with 44,544 AMD ‘Antares-A’ Instinct MI300A hybrid CPU-GPU compute engines running the High Performance LINPACK (HPL) benchmark at 64-bit floating point precision against the theoretical peak performance of an Ironwood pod with 9,216 of the TPU v7p compute engines," he wrote. "This is a perfectly silly comparison, and Google’s top brass not only should know better, but does."

24X the performance of El Capitan? Nope!

Prickett Morgan argues that while such comparisons are valid between AI systems and HPC machines, the two systems serve different purposes - El Capitan is optimized for high-precision simulations; the Ironwood pod is tailored to low-precision AI inference and training.

What matters, he adds, is not just peak performance but cost. "High performance has to have the lowest cost possible, and no one gets better deals on HPC gear than the US government’s Department of Energy."

Estimates from TheNextPlatform claim the Ironwood pod delivers 21.26 exaflops of FP16 and 42.52 exaflops of FP8 performance, costs $445 million to build and $1.1 billion to rent over three years. That results in a cost per teraflops of $21 (build) or $52 (rental).

Meanwhile, El Capitan delivers 43.68 FP16 exaflops and 87.36 FP8 exaflops at a build cost of $600 million, or $14 per teraflops.

"El Capitan has 2.05X more performance at FP16 and FP8 resolution than an Ironwood pod at peak theoretical performance," Prickett Morgan notes. "The Ironwood pod does not have 24X the performance of El Capitan."

He adds: "HPL-MxP uses a bunch of mixed precision calculations to converge to the same result as all-FP64 math on the HPL test, and these days delivers around an order of magnitude effective performance boost."

The article also includes a comprehensive table (below) comparing top-end AI and HPC systems on performance, memory, storage, and cost-efficiency. While Google’s TPU pods remain competitive, Prickett Morgan maintains that, from a cost/performance standpoint, El Capitan still holds a clear advantage.

"This comparison is not perfect, we realize," he admits. "All estimates are shown in bold red italics, and we have question marks where we are not able to make an estimate at this time."

(Image credit: TheNextPlatform) You might also like

• AMD-powered El Capitan becomes officially the world’s fastest supercomputer
• Tech editor takes close snaps of world's fastest supercomputer
• Google Cloud unveils Ironwood, its 7th Gen TPU to help boost AI performance and inference