A razor-edged clash over whether AI spells extinction, transformation, or mass job loss. This 6-minute summary (from a 2-hour debate) distills Noah Smith and host Liron Shapira’s sharp trade-offs on extinction risk, job automation, and who controls compute. You’ll get the core probabilities Noah assigns (96% long-run transformation, 0.1% extinction by 2050, ~10% chance of massive population loss), Shapira’s push for skeptical short-term timelines, and their arguments about compute scarcity, economies of scale, persuasion, bio risk, and self-modifying agents. Learn why the difference between annihilation and “childless doom” matters, how resource constraints could preserve human comparative advantage, and which policy levers—alignment, biogovernance, and compute concentration—might steer outcomes. Hosts: Liron Shapira; Guest: Noah Smith. Keywords: AI, job automation, extinction risk, AI alignment, compute bottlenecks, biosecurity, political economy. Listen now to get the key ideas in minutes.