An interview with OpenAI researcher Noam Brown about multi-agent systems, AI solving the Navier-Stokes problem, the internal/external model gap, and more (Dwarkesh Patel/Dwarkesh Podcast)

Noam Brown discusses how multi-agent systems can coordinate to solve complex physics problems, including the Navier-Stokes equations, and explains the gap between internal and external AI models. He highlights the challenges of aligning agent goals with human intent and the need for robust oversight mechanisms. The interview also touches on recent advances in AI-driven research and the implications for future scientific discovery.