Lex Fridman Podcast #344 – Noam Brown: AI vs Humans in Poker and Games of Strategic Negotiation
237 snips
Dec 6, 2022 Noam Brown, a research scientist at Meta AI, co-created systems that outsmart humans in poker and diplomacy. He discusses how AI achieved superhuman levels in these strategic games, highlighting the psychological dynamics of high-stakes poker. Brown explores the interplay between AI and human decision-making, emphasizing emotional engagement and trust in negotiations. He also delves into the future of AI in gaming, particularly its impact on non-playable characters and the ethical implications of replicating human-like decision processes.
AI Snips
Chapters
Books
Transcript
Episode notes
Libratus Poker Competition
- Noam Brown's team organized a 20-day poker competition, Libratus vs. four top human players, with $200,000 in prize money.
- Their bot, Libratus, aimed to approximate Nash equilibrium without explicit opponent adaptation, unlike earlier, less successful attempts.
Real-time Search in Poker AI
- Noam Brown realized that human poker players use real-time search and planning to improve upon their initial intuition.
- This inspired him to incorporate search into Libratus, significantly improving its performance by allowing for real-time strategy adjustments.
Libratus and Over Bets
- Libratus surprised Noam Brown by utilizing "over bets" (bets much larger than the pot), a strategy uncommon among human players at the time.
- These over bets proved effective in putting opponents in difficult situations, highlighting an area where AI strategy differed from human norms.
