Yann LeCun, one of the leading figures in artificial intelligence, poses a simple question: what happens when you hold a pen upright on its tip and let go? Even a toddler knows the pen will topple over. But no human would bother guessing in which direction it falls — there is no way to tell. An LLM like ChatGPT, LeCun argues, would try to generate a single prediction based on statistical patterns from its training data. That prediction would almost certainly be wrong, because the system is not reasoning about physical reality — it is generating what appears statistically plausible.
“We don't have robots that are nearly as good at understanding the physical world as a rat,” LeCun says. He spent a decade at Meta as chief AI scientist, but left in 2025 and founded Advanced Machine Intelligence Labs (AMI Labs) in Paris. His goal: move AI beyond current systems like ChatGPT, Claude and Gemini. They have their uses, he admits, but will never be able to tackle complicated situations in the real world, like getting a robot to do household chores.
“Yann LeCun says current AI is 'not smart' and raises $1bn for a new system that understands the real world.”
“They're not a path towards human level or human-like intelligence, or even animal-like intelligence, because they cannot deal with real world data, they just are not built for that,” he tells me on the sidelines of VivaTech, France's leading technology conference.
So AMI Labs is developing a new type of AI not based on the tech behind ChatGPT and its rivals. Investors see potential. Earlier this year AMI Labs announced it had raised more than $1bn (£760m), with backers including US computer chip giant Nvidia and the fund that manages the private wealth of Amazon-founder Jeff Bezos. That seed funding round — the earliest stage of start-up fundraising — was one of the biggest of its kind in Europe.
Large Language Models like ChatGPT are extremely good at some things: coding, mathematical problems, generating text. But LeCun argues these are well-defined and predictable problems. “They [LLMs] basically just accumulate knowledge… They can regurgitate something, you train them to regurgitate, but they're not particularly smart. They don't have an underlying understanding,” he says.
In the real world, a bewildering array of outcomes follows any action, requiring a more flexible AI. LeCun's system, called Joint Embedding Predictive Architecture (JEPA), creates abstractions of the real world that allow it to assess the outcomes of actions. Creating these abstractions involves difficult maths, but essentially they filter out useless information. The question now is whether JEPA can deliver what LLMs cannot: a machine that truly understands the physical world — perhaps even better than a rat.