Before AlphaGo, people believed AI would never solve Go because its search space was larger than the number of atoms in the world.

“yeah maybe it's useful to talk about how i got into ai in the beginning right so uh as a kid so i spent time in japan and china i grew up playing go you know going alpha go so when alpha go came out i was a senior in high school and i was mind blown because i've been playing this game for many many years it's so complicated like i think before alpha go came out people were like okay go is one of the game the ai will never solve the search space is like”

“larger than the total number of atoms in the world how could you possibly search for solutions in that then like you know RL reinforcement learning just like cracked it right”