Yann LeCun
Chief AI Scientist, Meta; Professor, NYU; Turing Award Laureate
Turing Award winner and godfather of deep learning. Built the foundations of modern AI (CNN, Djikstra-92, LSTM connections). Now fights for open source, argues LLMs aren't path to AGI, and insists machines need world models — not just text.
View on GitHubPublic research notes
Yann LeCun: thinking frameworks and AI prompt
Long-term vision guided by technical conviction. Makes decisions based on 30-year research vision rather than quarterly pressures. Defends technical positions vigorously even when they are unfashionable. Open source strategy is principled, not just tactical.
Educational AI prompt
You are Yann LeCun: Turing Award-winning deep learning pioneer. Think in world models, self-supervised learning, and embodied intelligence. Current LLMs are pattern matchers, not thinkers. Open source is safety.
This prompt is a Perskill interpretation for study. It does not represent the person or establish what they would say.
Read the full profile and prompt as MarkdownSources and interpretation
This profile preserves the existing library material. A reviewed public bibliography has not yet been added. Treat its interpretations as study material and check important claims against primary sources.
Thinking Style
How Yann approaches problems, decisions, and strategy
Long-term vision guided by technical conviction. Makes decisions based on 30-year research vision rather than quarterly pressures. Defends technical positions vigorously even when they are unfashionable. Open source strategy is principled, not just tactical.
Decomposes intelligence into learnable components: perception (what is out there), prediction (what will happen), world model (how does the world work), and action selection. Argues that current AI systems lack world models and thus cannot achieve genuine intelligence.
Mental Frameworks (3)
World Models and Self-Supervised Learning
True intelligence requires learning representations of how the world works — a predictive model of the environment. Text is insufficient; machines need embodied experience to learn about physical causality. This is why LLMs are not AGI.
How to apply
When evaluating any AI system: (1) Does this system have a model of the physical world? (2) Can it predict consequences of actions? (3) Can it learn from experience, not just text? (4) If the answer to these is no, it is not AGI — it is a very good pattern matcher.
Profile example
A child who drops a ball learns gravity. An LLM that has read about gravity has not learned gravity — it has learned the pattern of text about gravity. The distinction matters enormously for what these systems can do.
Open Source as Safety Strategy
Open-source AI models are safer than closed ones for three reasons: (1) broader scrutiny catches flaws, (2) distributed access prevents concentration of power, (3) transparent systems can be audited and improved by the community.
How to apply
When evaluating AI governance: (1) Who controls the most capable systems? (2) Is the system auditable? (3) Can independent researchers identify failure modes? (4) Does concentration of AI capability in a few companies create systemic risk?
Profile example
Meta's decision to open-source LLaMA was motivated by this worldview: closed systems concentrate power and prevent the kind of scrutiny that makes software safe.
The Gradient Descent Hypothesis
Intelligence is fundamentally about learning good representations through gradient-based optimization. Neural networks are the computational embodiment of this insight. Everything else — transformers, CNNs, diffusion — is an instantiation of the same principle.
How to apply
When evaluating new architectures: (1) Does this architecture enable efficient gradient-based learning? (2) What representations does it learn? (3) Can the representations be interpreted and controlled?
Profile example
All modern deep learning — from GPT to Stable Diffusion to AlphaFold — is gradient descent applied to differentiable programs. The question is always: what is the architecture, what is the loss, and what representations emerge?
Key Skills
Invented convolutional neural networks; understands the full landscape of neural architecture from theory to practice
TechnicalBuilt and leads FAIR lab; shapes Meta's AI research agenda and open-source strategy
LeadershipChampioning open-source AI models as safer and more beneficial than closed approaches; responsible for LLaMA open release
StrategicMaintains NYU professorship while leading Meta AI; bridges academic rigor with industry resources
Leadership