AI eats the low-hanging fruit, so the next famous unsolved problems will be harder

yaroslavvb · x · 2026-11-21

Yaroslav Bulatov argues that AI is picking off the low-hanging problems, which means the next generation of famous unsolved problems will be harder by construction: anything new gets run through AI first, so whatever remains unsolved is what AI couldn't crack. He suspects these will be problems connected to the real world, like deriving friction from first principles, and links his own writeup on why friction is hard to model.

Original post →

More from AGI Musings

AGI Musings channel →