LeCun Disputes Silicon Valley's Claim to AI Primacy
Meta's chief AI scientist lists foundational AI milestones born outside the Bay Area, from Montréal to Hangzhou.
Yann LeCun, VP and Chief AI Scientist at Meta, pushed back on what he called "BS" — a claim, attributed to investor Elad Gil, implying Silicon Valley leads AI development. In a rebuttal posted on X, LeCun catalogued where major AI milestones actually originated: the attention mechanism in Montréal, PyTorch in New York City, AlphaGo and AlphaFold in London, ESMFold in NYC, Llama 1 in Paris, and DeepSeek in Hangzhou. DINO and JEPA, two influential self-supervised learning approaches from Meta's own research, were developed across Montréal, Paris, and NYC. His closing line was pointed: "SV is 3 mos ahead on topics SV is singularly obsessed with."
The exchange matters for teams tracking where AI capability actually originates. Capital concentration in the Bay Area shapes which models reach production fastest, but LeCun's list is a reminder that the research foundations underpinning today's leading systems — architectures, frameworks, and frontier models — have emerged from labs distributed across North America, Europe, and China. Sourcing strategy, hiring pipelines, and partnership decisions that treat SV as the sole AI center may miss significant capability developing elsewhere.