Ideas / Dwarkesh: Arguments About the AI Trajectory
Ilya Sutskever — We're Moving From the Age of Scaling to the Age of Research
Dwarkesh Patel with Ilya Sutskever
Sutskever argues that current models generalize far worse than people and that the next gains may require ideas beyond brute-force pretraining.
Sutskever draws a line between absorbing more data and learning well from experience. The conversation moves through generalization, learning during deployment, superintelligence, and alignment without pretending these are separate departments.
Placed after Karpathy, it sharpens the question. Both see missing capabilities; they differ in emphasis about what kind of research might supply them.