How I Doctor with Dr. Graham Walker Move Over LLMS! AI Legends Yann LeCun and Alex LeBrun Debut AMI Labs' Bold Ambitions for World Models in Healthcare
9 snips
Feb 12, 2026 
Yann LeCun, Turing Award–winning deep learning pioneer, and Alex LeBrun, entrepreneur and CEO building clinical AI tools, discuss world models for healthcare. They explore why next-word LLMs fall short. They debate patient simulation, sensor-rich data over text, reliability needs beyond 80% accuracy, and practical near-term wins like documentation and medical coding.
AI Snips
Chapters
Transcript
Episode notes
World Models Enable Planning Not Just Prediction
- World models simulate future states from actions, enabling planning rather than token prediction.
- Yann LeCun argues LLMs predict next tokens blindly, while world models imagine consequences to plan robustly.
Sensor Data Outweighs Text For World Learning
- Real-world sensor data vastly exceeds text in quantity and richness for learning dynamics.
- Yann LeCun highlights vision and multi-sensory inputs as essential for building human-like models.
Doctors Use Mental Patient Models To Plan Care
- Clinicians run internal patient models to imagine treatment outcomes and sequences.
- Yann LeCun suggests AI should replicate this internal simulation to predict effects of interventions.