Yann LeCun

Domain expertise, not technical skills, is the critical factor for building successful AI applications in the era of democratized LLMs and foundation models.

“Meta's Chief AI Scientist acknowledges that every intelligence is specialized.”

OpenAI overstated GPT-5's mathematical capabilities by claiming it solved previously unsolved Erdős problems when it only found existing literature solutions.

“Meta's chief AI scientist criticized OpenAI's claims about GPT-5's math breakthroughs as overstated.”

“LeCun and Randall Balestriero proved the Platonic Representation Hypothesis, showing that separately trained image and text JEPAs converge to similar representations without paired data.”

“Proved with Randall Balestriero that 109 million parameter models can match understanding of 72 billion parameter models through joint embeddings, validating his years-long assertion that scale alone does not determine model understanding.”

“Developed JEPA, a self-supervised learning approach using joint embedding predictive architecture for abstract representation learning.”

“Founded AMI Labs, an AI research startup that raised about $1B in March.”

“Left Meta in November 2025 and admitted Llama 4 results were fudged, with different models used for different benchmarks, and expressed how the company lost confidence in AI leadership.”

“Shared 2018 Turing Award with Hinton and Bengio; withheld signature from letter calling for containment standards, arguing incidents stem from engineering shortcomings.”

“Did not sign letter calling for international AI containment standards; argued incidents reflect engineering failures.”

“Legendary AI researcher recruited by Zuckerberg to lead Facebook AI Research (FAIR); left Meta after being subsumed into new Meta Superintelligence Labs.”

“Computer scientist and AI researcher who argues there is no general intelligence and that missing world models are the key limitation.”

“Meta AI researcher who acknowledged company had fudged Llama 4 results by publishing internal build rather than public version.”

“Converging on world models as next major AI paradigm alongside Demis Hassabis and Fei-Fei Li.”

“Co-author of research on efficient video pretraining method LeVJEPA that matches Meta's V-JEPA 2 at up to 20x less compute.”

“Yann LeCun founded AMI Labs, a new AI research lab betting against LLM-based approaches to AGI and focusing on world-model architectures.”

“Joined Shaun Johnson and Oriol Vinyals to launch 224 Ventures, a technical venture firm investing in AI-native teams.”

“Co-author of 'Patch Policy' research paper on efficient embodied control for robot learning using visual representations.”

“Meta's chief scientist who agreed that concentration of power in AI is the biggest danger and compared closed AI labs to Ottoman empire banning the printing press; launched AMI Labs in Paris.”

“Yann LeCun's team built LeWorldModel, a stable end-to-end pixel-based world model with 15M parameters that plans 48× faster than foundation-model world models.”

“AI researcher and luminary who authored a paper proposing superhuman adaptable intelligence as a framework to replace AGI discourse.”

“Criticized Elon Musk's xAI as a failure, saying it cannot compete on the frontier of AI and warning of a potential bubble explosion in the industry.”

“Appointed to board of Logical Intelligence, left Meta in November, advocates against LLM-only approaches to AGI.”

“Former Meta AI chief who announced a $1 billion raise for AMI Labs, his newly launched AI venture.”

“Responded critically to OpenAI's false claims about solving mathematical problems.”

“Former Meta AI chief who left to found AMI Labs, which raised $1 billion.”

“Founded AMI Labs with $1.03 billion seed funding to develop world models for physical AI understanding beyond LLM paradigm.”

“Left Meta in late 2025 after 12 years citing disagreements over AI research direction and tensions with Zuckerberg; founded AMI Labs and raised €1.03 billion in seed funding.”

“Former Meta AI chief who left to found AMI Labs and raised $1 billion.”

“Leading research on alternatives to token-based thinking through JEPA and other architectures; left Meta in late 2025.”

“Co-author of LeWorldModel paper on joint-embedding predictive architectures for world models.”

“Founded AMI Labs after raising $1.03 billion and launched it in late 2025; pioneered latent World Models through JEPA and is skeptical that LLMs lead to AGI.”

“Meta's AI research division leader providing credible foundation for company's AI pivot away from metaverse.”

“Mentioned as having raised over a billion dollars and sharing a somewhat unclear vision of physical AI compared to Travis Kalanick's vision.”

“Founded AMI Labs and raised $1.03B betting on world models, specifically JEPA (Joint Embedding Predictive Architecture).”

“Co-authored research on temporal straightening for latent planning in world models to improve gradient-based planning stability.”

“Turing Award winner who co-founded AMI Labs to develop world models after leaving Meta.”

“Turing Award winner and ex-Meta Chief Scientist who launched AMI Labs with $1.03B seed funding to build world model AI instead of LLMs.”

“Launched Advanced Machine Intelligence (AMI Labs) with a $1.03B seed round to build AI systems based on world models with persistent memory, reasoning, and safety controls.”

“Co-authored research paper on multimodal pretraining beyond language modeling.”

LeJEPA introduces a simplified self-supervised learning method that eliminates many common training tricks while maintaining strong performance on benchmarks like ImageNet.

“Co-authored a paper introducing LeJEPA, a simpler way to train AI systems without labels using self-supervised learning.”

“Meta's chief AI scientist for 12 years who is leaving to start a new firm focused on advanced machine intelligence, criticized current AI approaches and called Alexandr Wang inexperienced.”

Soumith Chintala, creator of PyTorch and former Meta AI leader, has joined Mira Murati's Thinking Machines Lab startup, which is aggressively recruiting top AI talent and in talks to raise funding at a $50 billion valuation.

“Meta's chief AI scientist and founder of FAIR who is reportedly preparing to leave the company.”

“Pioneering AI researcher considering leaving Meta to launch a startup focused on world models, arguing that large language models are a dead end in pursuing superintelligence.”

Meta's chief AI scientist Yann LeCun is planning to leave the company to launch a startup focused on world models, amid organizational chaos in Meta's AI division following a major restructuring under CEO Mark Zuckerberg.

“Planning to leave Meta to build his own startup focused on world models, according to Financial Times report.”

Future generative AI advances will come from world models trained through sensory interaction rather than expanded large language models, enabling more capable and generally useful AI systems.

“Keynoted MIT symposium, arguing that the most exciting advances in generative AI will come from 'world models' rather than continued improvements to large language models.”