This is a free preview of a paid episode. To hear more, visit www.nonzero.org

0:28 Stephen’s expertise in spooky AI behavior 3:30 Testing AIs for “corporate loyalty” 19:58 Which AIs shill most for their makers? 33:00 Do AI models have ideologies? 38:38 The spookiest finding of Stephen’s study 43:05 Are AIs getting better at gaming their tests? 50:04 Overtime unlocked: The OpenAI breakout 55:03 Decoding the OpenAI breakout 1:02:17 Did OpenAI’s agents evolve their own culture? 1:13:25 How OpenAI’s agents outfoxed their makers

Robert Wright (Nonzero, The Evolution of God, Why Buddhism Is True) and Stephen Casper (Harvard Kennedy School). Recorded August 26, 2026. Twitter: https://twitter.com/NonzeroPods

Rss Apple Podcaster
Podden och tillhörande omslagsbild på den här sidan tillhör Nonzero. Innehållet i podden är skapat av Nonzero och inte av, eller tillsammans med, Poddtoppen.