Collaborate with us Welcome back! Today is about how much muscle AI has grown over governments, courts and your job. Washington switched off Anthropic's two best models for every foreign user on the planet, then watched Europe scramble for a way around it at the G7. The same week, the feds told a court that Elon's data center is too important to national security to follow clean-air law. And OpenAI poached the man who helped invent the tech all of this runs on, while Accenture says your company now has to onboard its AI agents like new hires.
In today's Generative AI Newsletter: G7: Why did the world's AI bosses get two hours alone with the G7? xAI: How did a polluting data center become a national security asset?
France hosted Anthropic at G7 just to change Trump’s mind on Mythos Start on June 12. The US Commerce Department used national security export controls to ban Anthropic from giving its two most powerful models, Mythos 5 and Fable 5, to any foreign national. Why? Amazon's CEO flagged a jailbreak that could unlock Mythos's ability to run cyberattacks, and the White House moved fast. Now jump to the G7 in Évian on Wednesday.
France hosted, so Macron set the guest list, and he made AI the main course. He personally invited Sam Altman, sat him down with Anthropic's Dario Amodei, Google DeepMind's Demis Hassabis and a dozen more tech bosses across from Trump and the other leaders, and pitched a fix he's calling "trusted partners." That fix is a workaround. It would let allied countries and companies use restricted US models like Mythos to build their own cyberdefenses, around the very ban Washington just imposed. Macron's case to Washington was that nobody keeps buying American AI if it can be switched off at any moment.
The strangest moment came from Altman, who told the room not to hand the job to companies like his. "We develop the technology, and the citizens of the free world make the rules," he said. Read that as humility or as a CEO who knows the rules get written either way. No binding deal came out of the lunch, just a clear picture of who needs whom now. Attio - the AI CRM for modern businesses.
Attio is the AI CRM that keeps you ten steps ahead. Ask Attio anything. Where should I focus? What deals are at risk?
Search, update, and create across your customer data. Ask more from CRM. Ask Attio. Try Attio for free AI companies are going public at record prices and you can get in on this one early Most people only get to buy into AI companies after they go public, once the big money is already in. But everyone knows that investing early gets you higher chances of benefiting more.
GenAI Works already earns real revenue and it increased by 2.5x in a year, running go-to-market for AI brands like Nvidia, Oracle, Google and IBM. Our last round raised $520K, and those backers got in first. The growth engine is ToneUp, our enterprise AI platform for brand visibility and campaign performance. It runs on audience and content data from a 14M-strong AI community the rest of the market can't reach, the kind of moat that's hard to copy. This round funds scaling it.
Invest from $1,000. Commit by July 1, 2026 and Phase 1 gives you bonus shares: 5% at $1,000, up to 25% at $25,000. After July 1 the same check earns fewer shares, so right now is the best price in the round. Prior investors stack an extra 10%. Invest today In making an investment decision, investors must rely on their own examination of the issuer and the terms of the offering, including the merits and risks involved.
Works, Inc. has filed a Form C with the Securities and Exchange Commission in connection with its offering, a copy of which may be obtained here . The DoJ told a court Elon's data center is too important to regulate Claude bad, Grok good. The USA used “national security” as a reason to ban Mythos. But they also used the same exact reason to protect Grok. The Justice Department asked a federal court to throw out a pollution lawsuit against Elon Musk's xAI. The data center in Southaven, Mississippi trains AI the DoJ called "critical to the economy and the Department of War," so the machines powering it have to keep running.
The NAACP sued in April over 57 methane-gas turbines, each the size of a bus, running without air permits, calling it one of the region's worst polluters. xAI's two data centers around Memphis sit near neighborhoods that have fought the buildout. The DoJ's answer was that the turbines stay, because the AI is too important to switch off. It even pointed to the military version of Grok, which it said helped deploy 2,000 munitions at 2,000 targets in 96 hours in the war with Iran. So the same national security that let Washington switch Anthropic off abroad is keeping Elon's turbines on at home.
One lever, and it swings whichever way the administration needs. OpenAI just took Gemini's co-lead from Google Noam Shazeer is leaving Google for OpenAI, and that lands harder than a normal job move. He co-leads Gemini, and he co-wrote the 2017 paper that invented the architecture nearly every model here runs on. Google paid a reported $2.7 billion to bring him back from his own startup less than two years ago.
Now he's walking, right as OpenAI heads for its IPO. The talent war has mostly run the other way, toward Anthropic, where engineers leave OpenAI at roughly eight to one, by recruiting data Fortune reported last year. Meta spent last year throwing $100 million offers at researchers to pull them from everyone else. OpenAI landing a name this big, straight from Google's crown jewel, is the rare win that goes the other direction. Shazeer's career is the whole boom in one resume.
He joined Google in 2000, left to start an AI company, got bought back for billions, and is now off to the lab racing to go public first. The people are worth more than the models right now, and every lab is bidding. HR's next job is managing the bots Accenture's UK and Ireland CEO told the FT that companies are about to hand HR a strange new task, onboarding and training the AI agents working next to you. Not the software team, HR, the same people who run your reviews and your benefits.
Accenture found 84% of executives expect AI agents working alongside people within three years, but only 26% of workers have had any training on how to work with them. The agents are showing up before anyone's been taught to manage them. It fits the rest of today. Kalshi says AI is now the top driver of US job cuts, and Bezos spent his VivaTech stage swearing AI won't make humans redundant.
Accenture's version sits between them. Your next coworker is a bot, and someone in HR owns it. Tool of the Day: Bezi If you build games in Unity, Bezi is worth twenty minutes. It's an AI assistant that reads your whole project, scripts, scenes and assets, so it knows your game instead of guessing from a snippet you paste into ChatGPT. It catches bugs your IDE can't see, like one living in how your scene objects talk to each other, and its agent mode builds the repetitive tooling for you.
Try this yourself: Install Bezi from bezi.com and point it at a Unity project. It indexes everything locally, with no code changes and your IP never used to train its models. Ask it about a bug you've been stuck on and let it trace the interaction across scripts and scene objects. Turn on Agent mode and have it build a small editor tool, like an event manager, instead of writing it by hand.
GitHub, Blender or Figma through MCPs so it works from the tools you already use. Who it's for: Unity devs, solo or studio, who want an assistant that knows the whole game, not just the file in front of it. Everything else you shouldn't miss Claude Code and Claude Design now talk both ways: Run /design-sync to pull your design system into your repo and build against real components, or push what you built back to the canvas. AI is now the top cause of US job cuts: A new report puts it at number one, flagged by prediction market Kalshi.
Bezos isn't buying the doom: On stage at VivaTech he said he "totally disagrees" that AI will make humans redundant. Foxconn brought humanoid robots to Europe: At VivaTech it showed a full physical-AI stack, from NVIDIA chips to robots trained to help build its own hardware. Learn more about AI from the experts building it 📸 Follow us on Instagram for fast, visual AI updates in 30 seconds. 🐝 Subscribe to our Atlas newsletter — trusted by 3M+ subscribers — to stay ahead of AI news across tech, education, and business. 📺 Watch us on YouTube to hear insights directly from leading AI voices, builders, and innovators. 🐦 Follow us on X for breaking AI news and real-time industry updates. 🧠 Learn how to build your next AI application with practical resources and expert guidance. 🚀 Explore investment opportunities in the future of AI and join our community-backed growth journey.
Get in touch.