(Overtime audio available to paid subscribers through the player at the top of the page and video version available below the paywall.)
0:00 Preamble: The OpenAI drama and this conversation
3:17 Where Dan falls on the AI doomer spectrum
6:13 Is safe AI possible without international cooperation?
12:09 Are we selecting AI for truth or engagement?
25:24 Using “psychology” to understand—and control—AI
30:39 The worrisome Darwinian dynamics of AI development
40:08 Life after an AI takeover
51:23 How structural incentives make AI safety harder
Robert Wright (Bloggingheads.tv, The Evolution of God, Nonzero, Why Buddhism Is True) and Dan Hendrycks (The Center for AI Safety). Recorded October 10, 2023.
Twitter: https://twitter.com/NonzeroPods