SignalTech/AI
OpenAI's reasoning father's latest interview! Mathematics is just the appetizer of the multi-agent era.
OpenAI reasoning chief interview signals multi-agent systems as next AI frontier.
Context
What happened: OpenAI researcher Noam Brown, a core author of o1, discussed multi-agent systems in a podcast, including an experiment in which 10,000 Agents spent 88 hours and produced 130 billion tokens to solve a Millennium Prize math problem.
Key numbers: 10,000 Agents; 88 hours; 130 billion tokens; Brown said multi-agent deserved at
Why it matters: Brown said the result came mainly from a general and very strong model, and that multi-agent may have received credit out of proportion to its contribution. He said coordination at the 10,000-Agent scale has not been rigorously measured.
What changed
OpenAI researcher Noam Brown, a core author of o1, discussed multi-agent systems in a podcast, including an experiment in which 10,000 Agents spent 88 hours and produced 130 billion tokens to solve a Millennium Prize math problem.
Why it matters
Brown said the result came mainly from a general and very strong model, and that multi-agent may have received credit out of proportion to its contribution. He said coordination at the 10,000-Agent scale has not been rigorously measured.