logoalt Hacker News

le-markyesterday at 11:16 PM0 repliesview on HN

Yes I have noticed this. I frequently have stronger models review weaker models. It’s very instructive to see what they get wrong.