✨
AI Summary
- Ilya Sutskever declares 'age of scaling' is ending and marks return to 'age of research,' identifying severe lack of generalization compared to human learning as fundamental bottleneck
- Paradox highlighted: models 'seem smarter than their economic impact would imply,' possibly due to researchers inadvertently 'reward hacking' evaluations rather than genuine capability gains
- Future AI development requires continual learning systems that rapidly acquire knowledge like humans rather than finished minds, with emotions functioning as evolutionary value function for decision-making