Y
Hacker News
new
|
ask
|
show
|
jobs
by
nicowesterdale
113 days ago
I wrote a, I hope, amusing breakdown of the structural reasons why off-the-shelf Large Language Models physically cannot "see" a chess board, and continue to make illegal moves, and teleport pieces as seen in Gotham Chess' latest videos.
https://www.nicowesterdale.com/blog/why-llms-cant-play-chess