According to a report from the Financial Times, Meta may soon lose one of its most prominent AI leaders. Yann LeCun, a chief AI scientist at the company, is reportedly planning to leave to build his own startup.
LeCun, who is also a professor at New York University and a winner of the prestigious A.M. Turing Award, plans to depart in the coming months. The report indicates he is already in talks to raise capital for a new venture that would focus on his work concerning world models. A world model is an AI system that develops an internal understanding of its environment, allowing it to simulate cause-and-effect scenarios to predict future outcomes. Other top labs and startups are also actively developing this technology.
This potential departure comes at a pivotal time for Meta. The company has recently changed its approach to AI development in response to concerns that it is being outpaced by rivals like OpenAI, Google, and Anthropic. Meta has reportedly begun revamping its AI organization after hiring over fifty engineers and researchers from its competitors to build a new unit called Meta Superintelligence Labs. In a significant move, Meta invested billions of dollars in the data-labeling vendor Scale AI and brought on its CEO, Alexandr Wang, to run the new division.
Sources indicate these decisions have created an increasingly chaotic environment within Meta’s AI unit. New talent has expressed frustration with the company’s bureaucracy, while the scope of the previous generative AI team has been limited. LeCun’s long-term research work under Meta’s Fundamental AI Research Lab, known as FAIR, has been overshadowed by CEO Mark Zuckerberg’s decisions to overhaul the company’s AI strategy. This shift occurred after Meta’s previous family of AI models failed to keep pace with competing models. Unlike the new Superintelligence Labs, FAIR is designed to focus on long-term research for techniques that may be used five to ten years in the future.
LeCun has been openly skeptical about how current AI technology, specifically large language models, is being marketed as a solution for all of humanity’s problems. He has publicly stated that AI systems still have a very long way to go, suggesting that the focus should be on creating more fundamental advances in intelligence.
Meta did not immediately return a request for comment outside of regular business hours.