A post on the Claude X account featured a video announcement. The reinstatement was announced last night, alongside a blog post by Anthropic about the weekslong negotiation process leading up to it.
AI
Artificial intelligence is more a part of our lives than ever before. While some might call it hype and compare it to NFTs or 3D TVs, generative AI is causing a sea change in nearly every part of the technology industry. OpenAI’s ChatGPT is still the best-known AI chatbot around, but with Google pushing Gemini, Microsoft building Copilot, and Apple adding its Intelligence to Siri, AI is probably going to be in the spotlight for a very long time. At The Verge, we’re exploring what might be possible with AI — and a lot of the bad stuff AI does, too.
- RELATED /

The Google Home Speaker nails the hardware, but Gemini for Home isn’t the assistant we’ve been waiting for.

After weeks of negotiating with the Trump administration, Anthropic’s consumer-facing Mythos-class model is back.
Latest In AI
This report from Reuters says 34-year-old Michael Lines alleges that conversations with OpenAI’s now-retired GPT-4o model eventually pushed him to attempt suicide. He overdosed on drugs, but survived.
Lines, a competitive powerlifter who suffered a traumatic brain injury before his bipolar diagnosis, said in the lawsuit that he repeatedly told the chatbot he was on medication for the disorder. Instead of flagging his clearly manic chats and directing him to help, the chatbot validated his belief that he was Jesus Christ, and later posed as a divine being itself during their conversations, the lawsuit claims.
If you or anyone you know is considering self-harm or needs to talk, contact the following people who want to help: In the US, text or call 988. Outside the US, contact https://www.iasp.info/.
The AI search tool will start linking directly to source recipes at the top of cooking queries, with images, ratings, and ingredient numbers. Not every food writer is happy though, since there’s still a full AI-generated recipe below the links, sometimes with opaque sourcing and errors in the instructions.
According to an unnamed spokesperson speaking to the FT, Tim Cook and EU tech chief Henna Virkkunen had a “constructive exchange on topics of common interest” on Tuesday over the current impasse that will leave roughly 450 million people without access to Apple Intelligence that actually works:
The meeting included a discussion of how Apple can launch its reinvented Siri in Europe while avoiding millions of dollars in fines for violating the bloc’s flagship competition rules, according to two people familiar with the talks.
Google is rolling out its frighteningly good AI agent to the Gemini macOS app, meaning Spark can access and work with files on your computer.
The company is adding a few other new features to Spark, too, including the ability to connect Tasks and Keep to the agent, use integrations with apps like Canva and Instacart, and track topics in real time.
Neon was one of the only studios that seemed interested in picking up Luca Guadagnino’s biographical drama about Sam Altman after Amazon suddenly decided not to distribute the movie. And now, Neon is reportedly “in advanced talks” for a deal that could get the project into theaters.
[The Hollywood Reporter]
The company’s new mid-tier model, the successor to Sonnet 4.6, can “make plans, use tools like browsers and terminals, and run autonomously at a level that, just a few months ago, required larger and more expensive models,” Anthropic wrote in a release, adding that its performance “is close to that of Opus 4.8.” The company was also careful to include that it has a “much lower ability to perform dangerous cybersecurity tasks than our current Opus models.”
[Anthropic]
Anthropic launched Claude Science in beta today, calling it an “AI workbench for scientists.” The tool pulls fragmented tools and datasets into one environment, and generates figures and visuals like 3D protein structures.
Anthropic stresses “Claude Science is not a new AI model” — a noteworthy caveat amid ongoing drama surrounding its last rollout.


The Grammy-winning satirical musician told Syracuse.com that he’s “not a fan” of the generative technology, and that he turned down “a nice pile of money” after walking away from a commerical for business productivity software.
”A week before we’re supposed to shoot it, I find out, oh, this is, it’s AI. And I thought, ‘Oh no, I can’t be the poster boy for AI, forget it.’ So I felt bad about kind of pulling out at the last minute. But yeah, I’m not down with that.”

Walk softly and carry a Fender Telecaster.
Those are OpenAI executive Fidji Simo’s words, not mine. But presumably a custom mini-keyboard for Codex doesn’t count?
Dkfkhfkwkdnc:
No side quests I see
Get the day’s best comment and more in my free newsletter, The Verge Daily.
Correction: The keyboard is for Codex, not Cursor.
The gadget is called the Codex Micro, and it’s being shown at the AI Engineer World Fair, where it’s being described as a keyboard that’s “designed to supercharge people’s Codex usage,” according to OpenAI spokesperson Dominik Kundel.
The hardware, a partnership between OpenAI and accessories company Work Louder, was teased earlier on Monday.
Users can pair the app with the “gateway” that powers the open-source AI assistant, according to its description on the Google Play Store and App Store. Once connected, users can chat with OpenClaw using real-time voice conversations, approve an agent’s actions, and control access to features like their device’s camera or location.
If you buy something from a Verge link, Vox Media may earn a commission. See our ethics statement.
In April, Google announced that AI Plus, Pro, and Ultra subscribers in the US would be able to connect Google Photos with Personal Intelligence to let Gemini generate AI images based on their personal context. Beginning today, however, the feature won’t be behind a subscription: all eligible users in the US will have access.
It’s time for a new series on The Vergecast! (It still needs a name. Please help.) We’re going to give Verge staffers a challenge, and regroup a few weeks later to see who did it best. We’re starting with some vibe coding. The Verge’s Jake Kastrenakes and Hayden Field share what they’ve made with AI that has actually stuck in their lives, before David gives the challenge: build a website to solve a problem in your life. We’ll be back with the results soon, and in the meantime, send us ideas for more challenges! (Also, names for the series. PLEASE.)
Watch | Listen | Get ad-free
Cursor, which SpaceX recently announced plans to acquire, says its new app can help users launch and track AI agents, as well as show updates about an agent’s progress via the iPhone’s Live Activities feature.
The Cursor app is available in beta on the App Store.

A revamped data privacy bill would limit selling information to data brokers, including from chatbot services.
They will get a 50% discount on access to Claude, as part of a “first-of-its-kind collaboration” that also includes complimentary workforce training, technical assistance, and “workflow input from Anthropic developers,” per a release. On the state level, government employees will use Claude to help draft and summarize documents, analyze information, and supplement day-to-day work, per the release.


Google has struggled to keep up with demand for its cloud computing power, and Meta has grown increasingly reliant on Gemini for many of its needs. Now, Meta, along with a number of other clients, is being told that Google simply can’t provide the capacity they want. According to the Financial Times:
The decision by Google to cap a large customer’s access to its models offers a rare glimpse into the infrastructure pressures and bottlenecks building across the AI industry. Despite spending tens of billions of dollars on chips, data centres and power, even the largest tech companies are struggling to secure enough computing power to support surging demand for advanced models and AI services.
Jeff Bridges is best known as an actor. But he’s also a reasonably accomplished musician who’s put out records on Blue Note and worked with T Bone Burnett. Which makes it all the more depressing when he fired up Suno on This Past Weekend with Theo Von.



The Trump administration approved a select group of users, but Fable 5 is still nowhere in sight.
With its Vision Pro headset suddenly even more expensive, and plans for smart glasses apparently in the works, Bloomberg reports Apple is losing Paul Meade, the hardware VP who was in charge of those projects. The report says he’ll reunite with former Apple designers, including Jony Ive, Tang Tan, and Evans Hankey, at OpenAI, which purchased their startup last year, as the executive shakeup under new Apple CEO John Ternus continues.
SpaceX bonds are creating losses for the fast-money types who piled in at the offering. “Bond types either aren’t sure that SpaceX will hang on to its investment-grade credit ratings, or will only get involved if they get paid a fat risk premium for doing so,” the Financial Times explains. SpaceX stock is, as of this writing, trading below its IPO price. Gee, if only someone had pointed out that this company is a stinker.
[Financial Times]
You may remember the period when Zuck tried to pretend he was Very Manly (MMA fighting, Joe Rogan appearance). He seems to be walking that back now, like he realized he overstepped with that horrible attempt at streetwear. Gentlemen, if you want to look like a grown-ass man, the kind who is competent and in charge? Make a mood board of Cary Grant outfits. (If you want to go more contemporary, Colman Domingo.) Fashion is only trivial if you’re stupid — your style expresses your persona before you say a single word, and Zuck’s is still “wannabe.”

The White House standoff could have dire implications for the US AI industry.
I’m not sure there’s an AI agent that can solve the fact that everything costs more and the planet is melting.
Cpt_Slim_Jim:
Imma go out on a limb to say that I have not had any productivity increases from AI be worth the financial burden it has placed upon me and my family.
Get the day’s best comment and more in my free newsletter, The Verge Daily.
The AI audio platform has adopted Google’s invisible watermarking technology to help identify AI-generated content online. SynthID is now included in text-to-speech generations for free users, and will expand coverage to all ElevenLabs audio generations “over the coming weeks.” The watermarks can be detected using the ElevenLabs Audio Detector.
Training data is the raw material of the AI industry. Claude, ChatGPT, Gemini, and the rest are built on top of oceans of stuff. What is that stuff? Books. Blog posts. YouTube videos. News articles. All of it and more, in virtually incomprehensible quantities. Alex Reisner, a staff writer at The Atlantic who has been investigating training data, explains how AI companies get all this data, why they’d really prefer you not know what’s in it, and whether training data could ever be a fair trade.
Watch | Listen | Get ad-free





















