RSS Bilingual Reader

A former rocket plateaus There is the NVidia of the last five years (up 1200%), and the Nvidia of the last six months (down 2%). When ChatGPT launched, it was trading at 14; it’s gone up by more than a factor of ten since then, rising, until recently like a rocket. Aside from Nvidia’s Jensen Huang, a brilliant and impressive CEO with considerable foresight, nobody was more responsible for Nvidia’s rise than Sam Altman, the relentless showman who runs OpenAI — and who repe...

Two of the many theories running around, as captured on my notifications screen Much of today’s chatter is about Elon’s big merger of SpaceX and xAI (which previously merged with X, previously known as Twitter). Theory One all about synergy. SpaceX will be oh so much better with access to all your tweets and xAI’s models! That’s the narrative Elon’s selling. You can read it on the SpaceX website, which tells us that the merged everything company will be “the m...

Honored to be a part of this, along with Yuval Noah Hariri, , Helen Toner, , and co-founders of Perplexity and Cohere: Naturally, I help anchor the skeptics corner (though I am hardly alone): And it will be fun to see if Helen Toner is right about this projection for the next five years: All of us expect a big impact on coding: And I turn out to be on the “large impact” side with respect to education: But with some words of caution: Melanie Mitchell is dead right about this: Lots mor...

The big news in AI over the last week is OpenClaw (formerly known as Moltbot and before that OpenClaw, changing names thrice in a week) — a cascade of LLM agents that has become wildly popular — and Moltbook, a social network for AI agents, built on top. Theoretically (I will get to that) Moltbook “restricts posting and interaction privileges to verified AI agents, primarily those running on the OpenClaw (formerly Moltbot) software, while human users are only permitted to obser...

A joint essay with Daniel Thilo Schroeder & Jonas R. Kunst, based on a new paper on swarms with 22 authors (including myself) that just appeared in Science . (A preprint version is here , and you can see WIRED’s coverage here .) Automated bots that purvey disinformation have been a problem since the early days of social media, and bad actors have been quick to jump on LLMs as a way of automating the generation of disinformation. But as we outline in the new article in Science we forese...

In further news vindicating neurosymbolic AI and world models, after Demis Hassabis’s strong statements yesterday , Yann LeCun, historically hostile to symbolic approaches, has just joined what sounds like a neurosymbolic AI company focused on reasoning and world models, apparently built on pretty much the same kind of blueprint as laid out in 2020. Here’s the announcement: and my 2020 blueprint that the above company mission reminds me intensely of: where hybrid was cashed out as a ...

Per the Times of India: For more about world models and how some of us have thought about them over the years, see this earlier essay: And see also my 2020, Next Decade in AI for an older but still very relevant perspective on how limits in statistical models militate for the use of neurosymbolic AI and world (aka cognitive) models.

Steve Eisman, an influential money manager famously played by Steve Carrel in the film The Big Short , and well known for having presciently shorting the subprime mortgage market, just had me on his podcast. From what I gather, our episode is making the rounds in the financial world. I thought it was a great interview. You can watch it here: Subscribe now

Content warning: this article includes discussion of suicide. If you are in crisis, please contact (call or text) the Suicide and Crisis Lifeline at 988. October: Sam Altman, boasting in October that he had mostly solved the mental health problems associated with LLMs, whilst sneaking in a reference to ChatGPT’s newfound applicatoin to “erotica” § There is a character in an Ursula K. Le Guin novel The Lathe of Heaven whose bad dreams come true. I am starting to feel like t...

Two Boston University law professors, Woodrow Hartzog and Jessica Silbey, just posted preprint of a new paper that blew me away, called How AI Destroys Institutions . I urge you to read—and reflect—on it, ASAP. Here’s the abstract. I hope that before the paper is finally published the authors will insert the word “Generative” before “AI” in the title, since the force of their critique is (as the last sentence of the abstract acknowledges) really about cu...

Things aren’t as rosy as they were Five years ago, hardly anyone had heard of Sam Altman. But by the time I ran into him, in May of 2023 in Washington, at the US Senate, he had become a superstar. Senators called him by his first name, and flaunted the meetings they had with him. Prime Ministers welcomed him with open arms. Newspapers wrote glowing profiles, many with scarcely a critical word. Tech leaders loved him, too. Microsoft had already at that point put in billions soon after ChatG...

Some recent news, all long anticipated by this newsletter: LLMs can still cannot be trusted: Full article A large fraction of what LLMs do is mostly just memorization (and Hinton was on the wrong side of this argument ) : Full article here. They still aren’t adding a lot of quantifiable value to the world: via newsletter Update: This is consistent with the finding of the Remote Labor Index that AI could only do about 2.5% of jobs, reported recently by the Washington Post . Scaling isn&#821...

It’s funny how close this competition seemed just a couple weeks ago: That was then, 2025. For 2026, the guy on the left is taking an early lead that’s will be hard to beat. Collage via Malin Frithioffson [link broken because LinkedIn removed the post]. There’s a lot more that could be said here but I will give the last words to Hidden Door’s CEO and Co-founder Hilary Mason. And, no, you won’t find me posting on X anymore. Share Please consider joining nearly 100,00...

Screengrab from Yann LeCun’s recent interview at the FT Yann LeCun is not a savory character ; but he is certainly a character. For the last few days the tech world has been gossiping about the saga of his departure from Meta, in which every aspect of his character, negative and positive, from his arrogance and certitude to his commitment to science, has been on full display. You can get a glimpse of this in the excerpt from his interview with the FT , above. A long tweet below sums up som...

Quick links to a couple of pieces I published elsewhere this week: First, in Politico’s annual “Black Swan” feature, on unlikely scenarios that could nonetheless come true, 15 Scenarios That Could Stun the World in 2026 : B y the end of 2026, President Trump will have begun to distance himself from the aggressively pro-AI industry policies that characterized his AI strategy in 2025. The giant AI infrastructure plays (like Project Stargate) that he championed after his inaugurat...

The first thing I read this morning was that the United States had removed Maduro from Venezuela. The second thing I read was a query from Brian Barrett, Executive Editor at Wired. He had asked ChatGPT and Perplexity about the situation. Both (though not Claude or Google) had fumbled the news , and Barrett wanted to know why. According to Barrett ChatGPT emphatically refuted that Maduro had been captured at all. “That didn’t happen,” it wrote. “The United States has not i...

Happy 2026! My prize for best AI cartoon of 2025 (reprinted with permission) goes to …. Tom Fishburne! His prescient 2023 cartoon on what I like to call The Fishburne Effect is legend: You can follow him - and order some of his collected work — at The Marketoonist website . § And, bonus, to help prepare you to think critically about AI 2026, in an era in which everyone is wondering whether generative AI will turn out to do be a dud , here are five essays from 2025 worth reading ...

You must leave now, take what you need you think will last. But whatever you wish to keep, you better grab it fast. — Bob Dylan If September 2025 was peak bubble, 2026 will likely the year which it all falls apart. The clues are everywhere. Here are two I read this morning: And a longer analysis, focused on debt, at The New York Times Times , worth reading in full: As the summary at the top puts it: It’s all over now, Baby Blue . Whether it all falls apart suddenly, or gradually, I d...

Silicon Valley and The White House appear to be completely smitten with large language models, rebuilding the world around the hope that LLMs will lead to artificial general intelligence, and a radically redefined economy. Signs of this enthusiasm are widespread. The generative AI economy has driven Nvidia and the valuations of private companies like OpenAI to astonishing heights. According to the Wall Street Journal, “ Business investment in AI might have accounted for as much as half of ...

AGI didn’t materialize (contra predictions from Elon Musk and others); GPT-5 was underwhelming, and didn’t solve hallucinations. LLMs still aren’t reliable; the economics look dubious. Few AI companies aside from Nvidia are making a profit, and nobody has much of a technical moat. OpenAI has lost a lot of its lead. Many would agree we have reached a point of diminishing returns for scaling; faith in scaling as a route to AGI has dissipated. Neurosymbolic AI (a hybrid of neural ...