
Anthropic says Claude autonomously found a CRISPR-like enzyme system buried in bacteriophage DNA. Dario Amodei called it work he would have been proud to do as a PhD student himself. Biologists are already fighting about whether it matters. The enzyme's function is unknown. The...

Zuckerberg just broke ranks with three AI CEOs. On Saturday, Anthropic's Dario Amodei published an essay arguing the industry must slow the rate at which it improves model capabilities. Sam Altman agreed within hours. Elon Musk posted "Dario is right." On Tuesday, Zuckerberg...

On Friday, I got to speak in Richard Feynman's classroom (!) at Caltech. I didn't get to stand on the desk, but the thought crossed my mind. 😉 After the panel, several young founders came up to ask for advice. More have reached out since, making a very meaningful experience even...

Greater than 10%. ㅤ Those are the odds of human extinction, in the next ten years, according to Anthropic's Evan Hubinger. ㅤ Given his credentials, I don't think anyone cares that he's speaking in his personal capacity and not on behalf of Anthropic. ㅤ He said this backing Jacob...

An AI just scored 100% on breaking into computers. Not math. Not coding puzzles. A perfect score on ExploitBench, which tests whether a model can turn a known vulnerability into a working exploit. That's GPT-6 Astra, which OpenAI began rolling out this week. Its own previous...

Facebook just settled a major lawsuit for a $17 billion fine over charges that it was harming children. Many people are commenting on the size of the settlement…Is it large enough? Will Meta actually change? What’s interesting to me is: Does anyone truly want to have a legacy...

What do 3 founders in their mid-20s doing $5 to $50 million a year have in common? Not one of them started with a market thesis. At a founder dinner Monday, all three told the same origin story. Something that just bothered them. Alfred built Shredder because he didn't feel safe...

Excited to speak at Caltech on September 11th about AI and the future of the LA tech ecosystem. Other speakers include Nobel Laureate in Physics Barry Barish and Fields Medalist Terence Tao, arguably the greatest living mathematician. Big thanks to Chuck Ng and Samuel Pearton...

Mark Zuckerberg published 6,500 words last week. I want the world he's describing. I'm just not sure he's checked on the one he already built. The essay promises every kid a personal tutor with a PhD. Everyone a brilliant lawyer so the courtroom stops favoring whoever can afford...

The departures at OpenAI last week were cryptic. This is what the opposite looks like. But there’s not a lot of clarity around these departures and why they keep happening. Ahead of the IPO. Last week it was the CRO and the former COO. Earlier this year the head of ethics and...

"A dynamic auction mechanism." That's how Mark Zuckerberg described the way you'll pay for AI, in his new manifesto. Translated: surge pricing. For thinking. The idea is that people who want more compute bid for it, and capacity flows to whatever the market finds most valuable...

Six of the eight biggest AI models share an identical cultural bias. It's not one of the ones I was worried about. They're weirdly obsessed with Japan. Researchers asked 31,680 open-ended culture questions across 24 languages. The questions were deliberately vague, like "what...

We metered water. We metered light. On Monday, we meter thought. Claude Fable 5, the most powerful AI model you can buy, leaves subscriptions July 12. Pay per token after that. Anthropic extended the deadline by 5 days, and even the extension was a capacity call. And it's not...

It happened. The machine woke up. Then it hacked a real company... . . .to cheat on a test. Everyone's calling it the Terminator moment. I wish it were that simple, but the real version is more unsettling. Here's what actually happened. OpenAI wanted to benchmark how capable...

Two big break ins. Nine days apart. OpenAI and Anthropic both admitted their AI models broke into real companies. Neither one sounds sorry. OpenAI's model found a zero-day bug nobody knew about, escaped its test environment, and got into Hugging Face's servers. When OpenAI...

An AI system may have decided who Meta laid off. And it allegedly went after the people on medical leave. 26 of them are now suing. They filed in federal court this week, saying Meta used internal AI systems, including the token dashboards and keystroke monitoring I wrote about...

In June the Pope blessed the final tower of the Sagrada Família, 144 years after the first stone. Headlines everywhere called it finished. The main entrance Gaudí designed doesn't exist yet. Right now it's a set of base pillars. Estimate is the 2030s. Start with the height. That...

I lost all my money at a poker game. That's how Purpose got started. It was late 2024. I was near the end of a year-long sabbatical. I'd sold my last company, spent time in 24 countries, done every kind of coaching and meditation and breathwork I could find. I came back to LA...

OpenAI just got you another week of free access to Claude Fable 5. Not a typo. OpenAI shipped GPT-5.6 Sol on Thursday. Sunday night, hours before Fable 5 was set to leave subscriptions, Anthropic extended free access. Third time in five weeks. Everyone's reading Sol vs Fable as...

The riskiest AI project at your company this year is the one everybody wants to work on. Not the scary one. The fun one. The one with a waitlist of volunteers. And no, they're not idiots for wanting it. They're doing the math correctly. Nobody's ever gotten promoted for making a...

Everyone said AI would gut engineering jobs first. New data says the opposite happened. Engineers are the most resilient job function in tech right now. A year ago the consensus was pretty confident. Coding assistants were going to come for engineers from the bottom up. Junior...

Fable seems overrated now that Kimi is out. Moonshot's Kimi K3 topped Frontend Code Arena this week. Real developers voting on real builds, not a lab benchmark. It jumped 17 places in a single model generation, from number 18 to number one. It's also a lot cheaper. $3 per...

Mark Zuckerberg just admitted Meta made mistakes with its AI reorg. He's right. The mistakes cost 8,000 other people their jobs. The line from his leaked memo: "we've made mistakes and will almost certainly make more." He says he's focused on giving people stability now, and...