Improving in Chess Is Hard. I Built the Most Human-Like Chess AI to Help Me
Posted2 months agoActive2 months ago
mbuffett.comTechstory
calmpositive
Debate
20/100
Chess AIMachine LearningPersonal Projects
Key topics
Chess AI
Machine Learning
Personal Projects
The author built a human-like chess AI to help improve their chess skills, and shares their experience and insights on the project, with commenters discussing the technical aspects and potential applications.
Snapshot generated from the HN discussion
Discussion Activity
Light discussionFirst comment
6m
Peak period
2
0-1h
Avg / period
1.3
Key moments
- 01Story posted
Nov 2, 2025 at 9:40 AM EST
2 months ago
Step 01 - 02First comment
Nov 2, 2025 at 9:46 AM EST
6m after posting
Step 02 - 03Peak activity
2 comments in 0-1h
Hottest window of the conversation
Step 03 - 04Latest activity
Nov 2, 2025 at 3:49 PM EST
2 months ago
Step 04
Generating AI Summary...
Analyzing up to 500 comments to identify key contributors and discussion patterns
ID: 45790648Type: storyLast synced: 11/17/2025, 7:49:08 AM
Want the full context?
Jump to the original sources
Read the primary article or dive into the live Hacker News thread when you're ready.
> The models themselves (221MB for the 512-dimension model, 494M for the 768-dimension model), are too large to upload to GitHub, but if anyone wants them I suppose I’ll figure out a way to share them.
I have no real interest in spending this amount of time and effort on my chess hobby not actually playing chess. If it gets made available somewhere I'd try it out.
My approach is to play against a strong computer opponent in chess.com. I also have Lichess open at the analysis page with stockfish open. Each time I move I update Lichess with my opponents previous move to see what options I had and if my move was one of the good ones. Then I record my move and play again back in chess.com. Works for me.