Yann LeCun
AI researcher and Turing Award winner
Yann LeCun invented the convolutional neural network that modern computer vision is built on, shared the 2018 Turing Award, and spent 12 years as Meta's chief AI scientist. He left in November 2025 to found AMI Labs, a Paris-based startup building world models — his long-argued alternative to large language models.
Why they matter
LeCun spent years telling the industry, from inside one of its biggest LLM shops, that scaling token predictors cannot deliver physical-world reasoning. Investors priced that heresy at $1.03 billion — the largest seed round in European startup history. If AMI's JEPA-based world models work, the field's scaling consensus was a detour.
Leadership intelligence
TECHi profile brief
Yann LeCun is the founder of AMI Labs and the former chief AI scientist at Meta, where he led the FAIR research lab for 12 years before leaving in November 2025. He pioneered convolutional neural networks and shared the 2018 Turing Award with Geoffrey Hinton and Yoshua Bengio. AMI Labs closed a $1.03 billion seed round at a $3.5 billion pre-money valuation in March 2026 to build world models based on his JEPA architecture, with Nabla founder Alexandre LeBrun as CEO.
TECHi thesis
LeCun is making the most expensive contrarian bet in AI: that autoregressive LLMs are an off-ramp, not the road, to machine intelligence. His argument — a system trained to predict the next token never learns how the physical world works — was easy to dismiss while every benchmark kept falling to scale. Now it has a balance sheet. AMI Labs raised $1.03 billion before shipping anything, on the promise that JEPA-style world models trained on abstract representations will out-reason token predictors. What proves him right: an AMI model doing planning and physical reasoning that GPT-class systems demonstrably cannot. What proves him wrong: LLM-family systems closing that gap first, which OpenAI and Google are explicitly trying to do. Either way, his exit stripped Meta of its most credible research voice.
Distribution leverage
87/100LeCun no longer commands Meta's distribution, but a billion-dollar war chest, the JEPA research lineage, and a research following few working scientists match give AMI Labs outsized pull for a company with no product.
Execution record
83/100ConvNets went from Bell Labs papers to the default architecture of computer vision — few researchers have a bigger shipped-idea record. The open question is whether he can run a startup as well as he ran a lab.
Audience demand
79/100Searches spike on his Meta exit, his LLM-scaling criticism, and every AMI Labs milestone; his technical arguments reliably cross over into mainstream tech coverage, which almost no working researcher manages.
Career map
Inflection points
Origin
Deep learning foundations
Built convolutional networks at Bell Labs and kept the approach alive through the neural-net winter — work that won the 2018 Turing Award.
Scale
Meta AI research voice
Led FAIR from 2013, making Meta the most open-publishing frontier lab and the institutional home of the field's open-weight culture.
Watch
Open research advocacy
Whether AMI Labs can turn $1.03 billion and the JEPA agenda into a working world model before LLM-family systems close the reasoning gap he built the company to exploit.
Search answers
People ask about Yann LeCun
Who is Yann LeCun?
Yann LeCun is the founder of AMI Labs and one of the three researchers who shared the 2018 Turing Award for deep learning. He invented convolutional neural networks, ran Meta's FAIR research lab for 12 years as chief AI scientist, and left in November 2025 to build world models at his own startup.
Why is Yann LeCun important in AI?
He is the field's most credentialed skeptic of LLM scaling. While rivals poured billions into bigger token predictors, LeCun argued they can never reach physical-world reasoning — then raised $1.03 billion at AMI Labs to build the JEPA-based alternative himself. Few researchers get to run a funded experiment on their own dissent.
What is Yann LeCun known for?
Convolutional neural networks — the backbone of modern computer vision — the 2018 Turing Award shared with Geoffrey Hinton and Yoshua Bengio, self-supervised learning, and the world-model research agenda now housed at AMI Labs. He is equally known for publicly calling LLM scaling a dead end while employed at a company built on LLMs.
Is Yann LeCun still at Meta?
No. LeCun left Meta in November 2025 after 12 years as chief AI scientist, a departure reported to follow FAIR budget cuts and a new AI leadership structure above him. His startup AMI Labs is private; Meta itself trades as META.