The race is between getting strong interpretability and shaping motivations, and the AIs taking over in ways that are not perceived. The podcast is split into two parts: one about Carl's model of an intelligence explosion and its implications for alignment, and the other about the economics of AI. [summary] [comments]
Source link
