20 Seconds of Thinking, 100,000× More Data
AI
generative AI
LLMs
reasoning
OpenAI
links
Noam Brown at TED AI: a poker bot thinking for 20 seconds matched scaling the model 100,000× and training it 100,000× longer.
“It turned out that having a bot think for just 20 seconds in a hand of poker got the same boosting performance as scaling up the model by 100,000x and training it for 100,000 times longer.” – Noam Brown
Originally posted on LinkedIn.
