What's Missing Between LLMs and AGI - Vishal Misra & Martin Casado
From AI + a16z by a16z
March 17, 2026 · 48 min · Episode 89
About this episode
Vishal Misra discusses his research on LLMs and the requirements for achieving AGI.
Vishal Misra returns to explain his latest research on how LLMs actually work under the hood. He walks through experiments showing that transformers update their predictions in a precise, mathematically predictable way as they process new information, explains why this still doesn't mean they're conscious, and describes what's actually required for AGI: the ability to keep learning after training and the move from pattern matching to understanding cause and effect.
People in this episode
Host: Martin Casado
Guest: Vishal Misra
Topics covered
- LLMs
- AGI
- transformers
- machine learning
- consciousness
- cause and effect
Keywords
- LLMs
- AGI
- transformers
- machine learning
- consciousness
- pattern matching
- cause and effect
More episodes of AI + a16z
- Building Search for AI Agents with Exa CEO Will Bryk · June 4, 2026 · 50 min
- AI Agents and the Fight for Customer Data · June 2, 2026 · 51 min
- Ben Horowitz on AI Infrastructure, Economics and The New Laws of Software · May 19, 2026 · 30 min
- AI Infrastructure, Distribution, and the Next Wave of Software · May 12, 2026 · 39 min
- From Vector Databases to Knowledge Engines: The Next Layer of AI · May 5, 2026 · 46 min
- Why We Need Continual Learning · April 28, 2026 · 19 min
Explore listener stats, chart rankings, contacts and more on the AI + a16z podcast page.