Skip to content 
Mark Zuckerberg is catching fresh heat after Yann LeCun one of Meta’s most influential AI leaders publicly signaled he’s done with where the company is headed. The core accusation isn’t personal drama; it’s that Meta’s post-ChatGPT speed-run toward product and secrecy may be setting up a bigger AI mess than people want to admit.
Meta’s Yann LeCun seeks €3bn valuation for new AI start-up
Watch the clip. See what the real fight is about. It is not about if AI is strong. It is about what happens when we go too fast. They are asking if rushing to make AI bigger and faster and everywhere will make weak systems. Will it make systems that break easily? Will they be hard to control? Will they be easy to use in bad ways? LeCun’s action makes this fight bigger. He shows Meta’s plan as a gamble on getting huge. But he is betting on a completely different kind of smart machine.
Reactions online are split in a way Zuckerberg can’t easily shrug off. Supporters say this is what happens when elite researchers dislike product pressure and leave to build their own vision. Critics argue LeCun is pointing at a real danger: companies racing each other into deployment while the public deals with the consequences of cost, trust, and the physical strain of data centers on power and water.
A longer deep-dive talk featuring LeCun’s “beyond chatbots” vision (world models, autonomy, and why he thinks current approaches hit limits).
Meta’s Chief AI Scientist Yann LeCun talks about the future of artificial intelligence
This is more important than one boss leaving. It is like a test for Mark Zuckerberg’s whole plan for AI. His plan is to spend a huge amount of money now. The value will come later. If people listen to LeCun’s warnings then Meta’s AI growth will look different. It will not look like a sure thing anymore. It will look like a very expensive bet. Winning this bet depends on a few things. It depends on if the AI is reliable. It depends on if they can control it. It depends on if the public gets angry about the cost and the problems.
