A Very Unlikely Chess Game | Slate Star Codex
Neither competitor has much to be proud of here. White has a poor opening. Black screws up and loses his queen for no reason. A few moves later, white screws up and loses his rook for no reason. Better players will no doubt spot other humiliating mistakes. But white does eventually eke out a victory. And black does hold his own through most of the game. White is me. My excuse is that I only play chess once every couple of years, plus I’m entering moves on an ASCII board I can barely read. Black is GPT-2. Its excuse is that it’s a text prediction program with no concept of chess. As far as it knows, it’s trying to predict short alphanumeric strings like “e2e4” or “Nb7”. Nobody told it this represents a board game. It doesn’t even have a concept of 2D space that it could use to understand such a claim. But it still captured my rook! Embarrassing! Backing up: last year, I wrote GPT-2 As Step Toward General Intelligence, where I argued that the program wasn’t just an essay generator, it wa
A Very Unlikely Chess Game | Slate Star Codex Home About / Top Posts Archives Top Posts Comments Feed RSS Feed Slate Star Codex Blogroll Economics Artir Kel Bryan Caplan David Friedman Pseudoerasmus Scott Sumner Tyler Cowen Effective Altruism 80000 Hours Blog Effective Altruism Forum GiveWell Blog Rationality Alyssa Vance Beeminder Elizabeth Van Nostrand Gwern Branwen Jacob Falkovich Jeff Kaufman Katja Grace Kelsey Piper Less Wrong Paul Christiano Robin Hanson Sarah Constantin Zack Davis Zvi Mowshowitz Science Andrew Gelman Greg Cochran Michael Caton Razib Khan Scott Aaronson Stephan Guyenet S
Explore this link on the map →related reading
- Chess-GPT’s Internal World Model | Adam Karvonenadamkarvonen.github.io
- Playing chess with large language modelsnicholas.carlini.com
- LLM Chess: Benchmarking Reasoning and Instruction-Following in LLMs through Chessarxiv.org
- Manipulating Chess-GPT’s World Model | Adam Karvonenadamkarvonen.github.io
- How Go Players Disempower Themselves to AI — LessWronglesswrong.com
- Chess Engines: A Zero to One Guidechessengines.super.site
- OthelloGPT learned a bag of heuristics — LessWronglesswrong.com
- Notes on Puzzles — Nabeel S. Qureshinabeelqu.co
- AlphaZero: Shedding new light on chess, shogi, and Go — Google DeepMinddeepmind.com
- Andrej Karpathy — AGI is still a decade awaydwarkesh.com
- Can chess be solved? – Rising Entropyrisingentropy.com
- Large Language Model: world models or surface statistics?thegradient.pub