Ilya Sutskever on Why Models Still Can't Generalize Like Humans

research training interview agi inference

The former OpenAI Chief Scientist explains the gap between eval performance and real-world capability, and why the AI scaling era may be ending.

Read full article →