Just weeks after receiving an award from King Charles at St James’s Palace, AI “godfather” Yann LeCun is stepping away from his role as Meta’s chief AI scientist to launch a new company centered on what he calls “advanced machine intelligence.” LeCun, a Turing Award winner and one of the foundational researchers behind deep learning, announced his departure following days of speculation, ending a 12-year tenure that included founding Meta’s influential FAIR research lab.
In a series of posts on Threads, LeCun thanked CEO Mark Zuckerberg and said FAIR’s impact on Meta, the AI community, and the broader tech world has been “spectacular.” FAIR helped push forward breakthroughs in machine learning, including translation and foundational research techniques. But Meta, like much of the industry, has increasingly concentrated its resources on large language models—the backbone of today’s generative AI tools.
LeCun has long voiced skepticism that LLMs can lead to truly human-level intelligence, arguing that their reliance on vast text datasets limits their ability to reason or learn like humans. His new venture will instead focus on training AI models through visual learning, mimicking how children and animals build understanding through interaction with the world rather than from static text.
As concerns mount about an “AI bubble” and the possibility of a market correction, LeCun’s move signals a bet on alternative approaches that could define the next era of AI development.
