Generally Intelligent cover image

Generally Intelligent

Episode 27: Noam Brown, FAIR, on achieving human-level performance in poker and Diplomacy, and the power of spending compute at inference time

Feb 9, 2023
01:44:54

Episode guests

Podcast summary created with Snipd AI

Quick takeaways

  • Incorporating extra computation at inference time can lead to significant AI performance improvements in complex games like poker and Diplomacy.
  • Transitioning from successful poker bot development to Diplomacy AI tackled the complexities of a seven-player game requiring innovative modeling of human behavior.

Deep dives

Importance of Leveraging Extra Computation at Inference Time

Utilizing extra computation at inference time can lead to significant improvements in AI performance. By throwing extra computation at inference time, substantial gains can be achieved, as seen in techniques applied to poker and other games. This approach requires a shift from simply scaling up models to focusing on enhancing search algorithms during inference to unlock the full potential of AI systems.

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
App store bannerPlay store banner