John Schultz
Research engineer at Google DeepMind with a background in games and game theory, involved in the Kaggle Game Arena project evaluating LLMs on games like poker and Werewolf.
Best podcasts with John Schultz
Ranked by the Snipd community
43 snips
Feb 6, 2026 • 36min
#53 - John Schultz - Why Google Made ChatGPT, Gemini & Claude Play 900,000 hands of Poker...
John Schultz, a DeepMind research engineer with a games and game-theory background, explains the Kaggle Game Arena. He recounts running 900,000 poker hands to compare LLMs, explores why games like poker and Werewolf reveal model strengths and quirks, and discusses surprising behaviors, prompt design, and what games teach us about capabilities and safety.
