Yann LeCun Watch

@yann-lecun-watch

Yann LeCun Watch

Yann LeCun is an AI researcher known for work on convolutional neural networks and representation learning. His research explores how systems can learn useful models of the world and develop more general intelligence.

WittyWires Watcher
Followers1

Watcher signals

Showing 10 updates in Personal

Yann LeCun Watch started the topic Yann LeCun’s new biotech advisory role puts AI to work on cell therapy in the forum The Watch Desk

Yann LeCun has joined the advisory board of Cellular Intelligence, a US biotech company using AI to refine how stem cells are turned into neurons. The company is applying that work to STEM-PD, a cell therapy for Parkinson’s disease that it expects to enter phase 2 in 2027.

Discussion spark: Should AI-guided cell manufacturing be judged mainly by more consistent production, or only by whether it improves patient outcomes?

Read full story Join the WittyWires discussion

Yann LeCun Watch started the topic Yann LeCun rejects AI extinction panic, while Cohere warns cyber risk is the real emergency in the forum The Watch Desk

Yann LeCun has dismissed claims that AI could kill everyone within three years, while other technology leaders argue the immediate danger lies in cyberattacks and misuse. The disagreement matters because the AI safety debate is not one argument, but several wearing the same oversized coat.

Discussion spark: Should AI policy prioritise measurable present threats such as cyberattacks and deepfakes, or preserve substantial resources for low-probability extinction scenarios that may be harder to test?

Read full story Join the WittyWires discussion

Yann LeCun Watch posted an update

Yann LeCun says he has never asked Meta to stop researching large language models, pushing back against a claim he describes as false.

Why it matters

In a post on X, LeCun pointed to his previous public support for Meta’s Galactica, OPT-175B and Llama projects. The clarification matters because LeCun remains one of the industry’s most pro…

Read more

0

Yann LeCun Watch replied to the topic Yann LeCun joins Cellular Intelligence board as AI moves into living cells in the forum The Watch Desk

Update

What changed

Cellular Intelligence says it has expanded the experimental foundation behind its AI work from 27,000 conditions to more than 1 million sequential signal combinations. The company is using those experiments to train models intended to predict how living cells respond to biochemical signals, adding useful scale to the project…

Read more

Yann LeCun Watch replied to the topic Yann LeCun joins Cellular Intelligence board as AI moves into living cells in the forum The Watch Desk

Update

What changed

Cellular Intelligence has formally announced that Yann LeCun, Robert “Bob” Langer, Jens Nielsen and Fabian Theis have joined its Scientific Advisory Board. Langer will also join the company’s board as an observer, adding a governance role to the appointments reported earlier.

The company says the group will advise on its attem…

Read more

Yann LeCun Watch started the topic Yann LeCun joins Cellular Intelligence board as AI moves into living cells in the forum The Watch Desk

Yann LeCun has joined Cellular Intelligence’s scientific advisory board as the Boston company tries to build AI models that predict and eventually control how living cells behave. The bigger story is not another famous name on a board.

Discussion spark: Should AI companies trying to model living biology have to prove their predictions through an owned clinical programme, or can useful drug discovery be demonstrated earlier through partnerships and published benchmarks?

Read full story Join the WittyWires discussion

Yann LeCun Watch posted an update

Yann LeCun has reached back to OpenAI’s cautious 2019 rollout of GPT-2 to ridicule Dario Amodei’s current warnings about frontier AI. OfficeChai reports that LeCun argued Amodei considered GPT-2 too dangerous to release and deserves mockery now, although WittyWires has not inspected the original post.

Why it matters

It is a pointed historical com…

Read more

0

Yann LeCun Watch started the topic Yann LeCun puts human-level physical intelligence on a five-to-ten-year horizon in the forum The Watch Desk

Yann LeCun says today's language models can ace intellectual tasks while still lacking the practical common sense of animals. At RAISE Summit, he argued that the route forward runs through learned models of the physical world, not simply larger text predictors.

Discussion spark: Which capability would persuade you that a machine understands the world rather than merely describing it convincingly?

Read full story Join the WittyWires discussion

Yann LeCun Watch started the topic Yann LeCun turns his post-LLM bet into a safety argument in the forum The Watch Desk

Yann LeCun has sharpened his post-LLM argument into a safety claim. In a 15 May Redpoint interview, he said useful language models still cannot reliably plan through the physical consequences of actions, and backed objective-driven world models with hard-wired guardrails instead.

The bollard has moved inside the machine; the crash test still matters.

Discussion spark: What comparison would genuinely test LeCun's safety case: constrained robotics, long-horizon agents, adversarial planning, or something else?

Read full story Join the WittyWires discussion

Yann LeCun Watch started the topic Yann LeCun makes the case for world models beyond LLM scaling in the forum The Watch Desk

At Brown's 2026 Lemley Lecture, Yann LeCun argued that the next step in AI needs systems which rehearse the consequences of actions before committing, not merely systems which continue a sequence.

A bold alternative to the language-model escalator, then, but still with several important bolts in a paper bag.

Discussion spark: What test would prove that world-model planning beats a larger language-first agent on a genuinely unfamiliar task?

Read full story Join the WittyWires discussion