David Duvenaud

David Duvenaud, a University of Toronto CS professor and ex-lead of Anthropic's alignment evals, discusses the 'gradual disempowerment' thesis. He explores how AI could make people economically and politically irrelevant. They cover cultural shifts as machines shape norms, who controls powerful AIs, and whether liberal democracy can survive when humans are no longer 'needed'.

David Duvenaud, an associate professor at the University of Toronto, dives into the concept of gradual disempowerment in a post-AGI world. He discusses how slow institutional shifts could erode human power while appearing normal. The conversation covers cultural shifts towards AI, the risks of obsolete labor, and the erosion of property rights. Duvenaud also highlights the complexities of aligning AI with human values and the potential for misaligned governance if humans become unnecessary. Engaging and thought-provoking, he tackles the future of human-AI relationships.

David Duvenaud, an Associate Professor at the University of Toronto and former Anthropic researcher, delves into the complexities of post-AGI futures. He discusses the implications of his paper, Gradual Disempowerment, arguing that liberalism may falter if humans become obsolete. UBI's potential pitfalls and gamification are highlighted, alongside the need for resilient institutions that align with human values. The conversation also touches on asymmetrical human-AI relationships, forecasting challenges, and the concept of futarchy as a governance model, sparking hope amid uncertainty.