this post was submitted on 11 Jun 2025
129 points (99.2% liked)
chapotraphouse
13914 readers
895 users here now
Banned? DM Wmill to appeal.
No anti-nautilism posts. See: Eco-fascism Primer
Slop posts go in c/slop. Don't post low-hanging fruit here.
founded 4 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I mean they aren't large chess models. They can only do language tasks. They don't think, they predict words based on context and its similarity to the corpus they're trained on.
Is chess one of those problems that can be solved if you just memorize every single game ever played and continuously remember as they happen? Probably not. People have been trying that for centuries.
I think we're going to find a lot of things in life can't be solved by computers memorizing stuff and then doing stats on it to get an answer. Tech bros mold themselves after computers though. They think everything is just systems, algorithms, data structures, and math. And not the good math either, the mid-century diet-Rand game theory cold war shit they confuse with human nature.
Yes, chess can be solved by simply knowing every possible board state. However there's like 10^50 possible positions (we think, it's actually unknown how many possible legal positions there are) and storing that amount of information would require more than the sun's volume in hard drives
Even that might not be enough lol. There are more possible moves than there are atoms in the universe. If you get rid of what are likely illegal moves, it's (as you say) around 10^50. The space needed to even compute that, however, would be larger than our entire galaxy even with the most efficient computer possible that doesn't exist.
Go has over 10^170 moves, which is even more of a challenge to compute.