The Risk of Misgeneralization and Jailbreaking

0:00 How Eliezer Yudkowsky’s new book envisions AI takeover 8:24 Will we ever really know how AI works? 15:10 The “paperclip maximizer” problem revisited 26:13 Will advanced AIs be insatiable? 31:39 David’s alternative takeover scenario: gradual disempowerment 43:31 Can—and should—we keep humans in the loop? 51:46 Heading to Overtime

Robert Wright (Nonzero, The Evolution of God, Why Buddhism Is True) and David Krueger (The University of Montréal). Recorded September 24, 2025.

Twitter: https://twitter.com/NonzeroPods