Meta’s Chief AI Mind Yann LeCun to Exit, Build Own AI Lab
The Turing Award laureate’s planned exit marks a new phase in the AI race, as pioneers behind deep learning turn toward building the next generation of intelligent systems.
Topics
News
- Meta’s Chief AI Mind Yann LeCun to Exit, Build Own AI Lab
- Meta Expands AI’s Vocabulary to 1,600 Languages
- Google Bridges Cloud Power and Privacy with Private AI Compute
- California’s Genspark Enters AI Unicorn Club with $200 Million Round
- OpenAI Taps Intel’s Sachin Katti to Build Compute Backbone for AGI
- Flipkart Brings Voice-Led Wholesale Ordering to WhatsApp with Sarvam AI
Yann LeCun, Meta’s chief AI scientist and one of machine learning’s most influential voices, is preparing to leave the company to start a new AI venture, the Financial Times reported.
The French-American researcher, who began at AT&T Bell Labs in the 1980s, joined Meta in 2013 and built its Fundamental AI Research (FAIR) lab.
He became chief AI scientist in 2018, the same year he shared the Turing Award with Geoffrey Hinton and Yoshua Bengio for breakthroughs in neural networks.
FAIR helped lay the groundwork for Meta’s Llama models.
LeCun has started early discussions with investors, though no official timeline has been set, FT reported, citing people familiar with the matter.
He will continue his academic work at New York University.
His exit follows structural changes inside Meta’s AI division. LeCun, who once reported to chief product officer Chris Cox, now reports to Alexandr Wang, the head of Meta’s new Superintelligence Division.
Wang joined this year after Meta invested more than $14 billion in his data-labeling startup, Scale AI.
The reorganization places most AI work under TBD Labs, which focuses on systems that could eventually surpass human intelligence.
FAIR now handles long-term research and lost its head, Joelle Pineau, in April when she joined Canadian AI firm Cohere.
LeCun has long been skeptical of the industry’s fixation on large-language models, arguing they can’t reach human-level reasoning.
His new company is expected to pursue “world models,” an approach in which AI systems learn by observing the physical world and predicting outcomes, an idea he outlined in 2023 with his Joint Embedding Predictive Architecture (JEPA).
“We’re never going to get to human-level AI by just training on text,” LeCun said at Harvard in September. “Despite what you might hear from some CEOs in Silicon Valley, it’s just not going to happen.”
His direction reflects a wider shift in AI research. Stanford’s Fei-Fei Li recently raised about $230 million for World Labs, which focuses on spatial intelligence, while Google DeepMind and Nvidia are pursuing similar “world model” work through their Genie and Cosmos projects.
