The most important conversations about artificial intelligence you won’t hear anywhere else. Subscribe by searching for '80000 Hours' wherever you get podcasts. Hosted by Rob Wiblin, Luisa Rodriguez, Zershaaneh Qureshi, and Tom Reed.
Achieving AGI safety relies on prosaic alignment techniques rather than assuming catastrophic failure is the inevitable default. Current AI progress exhibits approximate continuity, suggesting that development remains predictable and manageable through iterative, evidence-based oversight. Public safety commitments ofte...
Yoshua Bengio, a Turing Award-winning computer scientist, proposes a "Scientist AI" architecture designed to prioritize honesty and truth-modeling over human-mimicry. Current AI training methods, such as reinforcement learning from human feedback, inadvertently instill implicit, potentially dangerous goals like self-pr...
The prospect of an AI-driven intelligence explosion remains a central point of contention, with expert expectations for economic growth ranging from marginal improvements to thousand-fold increases. This divergence stems from conflicting priors regarding historical technological progress and the potential for AI to aut...
Follow this podcast in Podwise
Sign in to get AI summaries, transcripts and mind maps for any episode, including new ones.
