Episode #98
Ilya Sutskever — We're moving from the age of scaling to the age of research
“These models somehow just generalize dramatically worse than people. It's a very fundamental thing.”
November 25, 20251h 36m10 chapters
What this episode covers
“These models somehow just generalize dramatically worse than people. It's a very fundamental thing.”
Where to start
SSI’s plan to straight-shot superintelligence
SSI’s plan to straight-shot superintelligence
Start at 35:45
“We are squarely an age of research company”
“We are squarely an age of research company”
Start at 1:18:13
Why humans generalize better than models
Why humans generalize better than models
Start at 25:13
People and topics
Key takeaways
- Explaining model jaggedness
- Emotions and value functions
- What are we scaling?
- Why humans generalize better than models
- SSI’s plan to straight-shot superintelligence