Ideas / Dwarkesh: Arguments About the AI Trajectory
AI 2027: A Month-by-Month Model of an Intelligence Explosion
Dwarkesh Patel with Scott Alexander and Daniel Kokotajlo
Scott Alexander and Daniel Kokotajlo walk through a scenario from coding agents to AI-assisted AI research, geopolitical competition, and misalignment.
Treat this as scenario, not prediction. The sequence moves from better coding agents through automated research, recursive improvement, international competition, misalignment, and possible post-AGI societies.
Patel approaches the model skeptically and tests its assumptions along the way. That friction makes the episode more useful than a countdown clock wearing a lab coat.