4 ms·
Sutskever is simply much more interested in doing research like https://openai.com/blog/introducing-superalignment https://openai.com/blog/introducing-superalig
by amai 3y ago
Sutskever is simply much more interested in doing research like https://openai.com/blog/introducing-superalignment https://openai.com/blog/introducing-superalignment than just earning money.
- Jensson 3y agoI can see why, being able to give hard rules the AI can't break wouldn't just solve alignment but also solve the hallucination problem. It is the kind of breakthrough that would be worth trillions.