Hacker News new | ask | show | jobs
by futureshock 318 days ago
I think that sounds very reasonable, but unfortunately these models don’t know what they know and don’t. A small model that knew the exact limits of its knowledge would be very powerful.