π‘οΈπ€ AI safety used to mean delaying capability.
Anthropic is playing a different game. They are racing their most dangerous model into defender hands before the public or attackers see it.
150 orgs. 15 countries. Power, water, healthcare, hardware.
This is what release strategy looks like when one model can find 10,000 zero-days in two months.
Cybersecurity just got a new clock.
Weβre expanding Project Glasswing. Weβve extended access to Claude Mythos Preview to approximately 150 additional organizations, based in more than fifteen countries.
Read more about this expansion and our future plans for Project Glasswing: https://t.co/QrtHSBdRbh
πΊπΈβοΈ The most consequential AI policy moment of 2026 happened in private.
Trump signed an order asking Anthropic, OpenAI, and Google to hand frontier models to the government 30 days before release. No ceremony. No executives present.
The original was 90 days. Musk and Zuckerberg called overnight to cut it down.
Voluntary on paper. The pipeline is built.
π€π₯ Microsoft just made OpenAI dispensable.
Project Polaris is Microsoft's own coding model, running on its own Maia 200 chips, replacing GPT-4 inside GitHub Copilot in August.
This is the same Microsoft that bet $13B on OpenAI in 2023. Now it owns the full stack and competes on model quality, not just distribution.
The seven-year exclusive partnership ended in April. Polaris is the proof.
https://t.co/crse48eu5l
ππ For years, AI safety was something companies said they did.
OpenAI just published a framework that maps internal safety practices directly to specific California and EU laws. Regulators now have a verifiable document to point at when things go wrong.
This is what AI accountability looks like in 2026.
https://t.co/R7fFgQa89T
βοΈπ€ If your business has an AI chatbot answering customer service questions, California just put you on notice.
AB 1609 cleared the Assembly on May 27 and crossed to the Senate on the May 29 deadline. 78 chatbot bills are alive across 27 states right now.
The compliance clock started ticking.
https://t.co/FAIYAx1qNR
π€πΌ Google's first true consumer AI agent costs $99.99 a month.
Gemini Spark runs 24/7 on dedicated cloud VMs and integrates with Gmail, Calendar, Docs, and connected apps via MCP. Even when your phone is locked, it keeps working.
This is the first time AI Ultra actually justifies the price.
Gemini Spark is now available to all Google AI Ultra subscribers in the U.S.
It can handle the heavy lifting and connect the dots across your digital ecosystem to take action where it matters most. Whether you watch it work or let it run in the background, Gemini Spark is always under your direction.
Give it a try at https://t.co/gTWOnOwQDE or in the app and let us know what you think.
π§ β The most underrated AI metric is not intelligence. It is honesty.
Claude Opus 4.8 is four times less likely than 4.7 to let code flaws slip past without flagging them. SWE-bench Verified jumped to 88.6 percent.
Models that admit uncertainty are the only ones safe enough to ship to production.
π @AnthropicAI's Claude Opus 4.8 is now generally available and rolling out in GitHub Copilot.
Early testing shows:
β’ It demonstrates a clear step forward in code understanding and generation across a range of real-world coding tasks.
β’ It handles complex problem-solving and large-codebase navigation with notable improvement to previous versions.
Try it out in @code or Copilot CLI.
https://t.co/Br33vOsEwy
ππ From $380B in February to $965B in May.
That is a $585 billion valuation jump in 90 days for a 5-year-old company.
Anthropic's annualized revenue went from $4 billion in July 2025 to over $30 billion now β an 80-fold surge in less than a year.
The AI capital cycle just broke physics.
We've raised $65 billion in Series H funding at a $965 billion post-money valuation, led by @AltimeterCap, Dragoneer, @Greenoaks, and @sequoia.
This investment will help us advance our research and expand our capacity to meet growing demand for Claude.
π€π’ KPMG just gave Claude to 276,000 employees across 138 countries.
Deloitte already gave Claude to 470,000. PwC and EY will not be the last to follow.
The Big Four are not running AI pilots anymore. They are embedding it into the same software they use for tax, legal, and audit work.
The professional services industry is being rebuilt in real time, and the people using it have not noticed yet.
https://t.co/3nRNSfc5vC
ππ€ The NBA didn't decide to use AI for out-of-bounds calls because the tech is finally ready.
They decided because one blown call in Game 5 cost the Spurs a Western Conference Finals game and Adam Silver was tired of explaining it.
Every league will end up here. Tennis already did. MLB already did. NFL is next.
The fastest way to deploy AI in the real world is to wait for the human version to fail on live television.
ππ° Anthropic just hit a $965 billion valuation, quietly passing OpenAI's $852B.
Same day they shipped Claude Opus 4.8, which beats GPT-5.5 on 12 of 15 benchmarks at the same price as 4.7.
Five years ago Dario Amodei was running speech experiments. Today he leads the most valuable AI company on earth.
The pecking order in AI just flipped.
ANTHROPIC π₯: CLAUDE OPUS 4.8 IS ROLLING OUT TO ALL USERS.
The release also includes an updated Thinking effort selector with Low, Medium, High, Extra, and Max options available.
> Switch to Opus 4.8 for your most ambitious work - and now you can set the effort level for thoroughness or speed.
A study of 100,000 people found that GPT-4 now outscores 72% of humans on creativity tests. Researchers included Yoshua Bengio from Mila.
But the top 10% of humans still beat every AI model tested. The gap is not closing at the top. It is closing in the middle.
Average creativity is now a commodity. The only advantage left is being genuinely exceptional at thinking differently.
Source: https://t.co/WgdcgF0oBO
ByteDance is quietly discussing $70 billion in AI spending for 2026. That is more than Tencent and Alibaba's capex combined.
Bloomberg says the TikTok parent would fund it through $50 billion in profit from last year. They also just bought millions of Qualcomm chips for agentic AI.
Most people still think of ByteDance as a social media company. It is becoming one of the largest AI infrastructure investors on Earth.
The supply chain story is bigger than the model story.
Source: https://t.co/IGsgNQgqrJ
OpenAI is not just selling models anymore. It is selling the people who install them.
150 engineers embedded inside your company on day one. 19 PE firms with 2,000+ portfolio companies as built-in customers. A 17.5% guaranteed annual return for investors.
This is not a product launch. It is a distribution machine disguised as a consulting firm.
Today weβre launching the OpenAI Deployment Company to help businesses build and deploy AI.
It's majority-owned and controlled by OpenAI. It brings together 19 leading investment firms, consultancies, and system integrators to help organizations deploy frontier AI to production for business impact. https://t.co/GnyjGFaLLA
ππ€― A memory chip company. One trillion dollars. Most people still think AI wealth is about chatbots. Meanwhile Micron's stock is up 700% in 12 months because every Nvidia GPU needs their chips to function.
The real AI money is in the supply chain nobody talks about.
π€―π₯ The Pope just told the world to "disarm AI." Not a tweet. An 83-page encyclical presented alongside the co-founder of Anthropic. This is the first time a Pope has dedicated an entire document to artificial intelligence.
The AI industry now has the Vatican watching its every move.
Anthropicβs Chris Olah at the Vatican press conference to present Pope Leo XIVβs encyclical on AI: βThere is a real possibility that AI will displace human labor at very large scale. If that happens, supporting those displaced will be a moral imperative of historic proportions.β
Pope Leo wrote in the encyclical, βThe pursuit of greater profits cannot justify choices that systematically sacrifice jobs, because the human person is an end, not a means, and the economic order must remain subordinate to human dignity and the common good.β
Precision in AI animation tools is usually either nonexistent or hidden behind complex parameters. Direct control over the output like this changes what becomes possible.
Omni-reference gives you full control over your AI-animated shows.
It lets you craft scenes with an insane level of precision.
Hereβs how it works π
π€ Not just looks real.. behaves real..
Gravity.. kinetic energy.. physics simulation.. all from a text prompt..
Edit your own videos just by describing what you want.. live today for all paid Gemini subscribers..
This is the Nano Banana moment for video..
Gemini Omni combines an intuitive understanding of physics with Gemini's real-world knowledge and reasoning. π
Now, the stories and outputs you create won't just look photorealistic β they'll behave like the real world.
#GoogleIO