what do you think of his idea?

i can annotate chess games from magazines , databases like TWIC , make them like a substack or publish them on blogs learn myself ,comparer with stockfish 19 :( ( a engine i hate cuz it makes feel outdated) and learn myself give my insights to others and they discuss and find my flaws or see where im right , and learn myself from blogs by like a FM i found named chessatlas and i like his blogs, so i can maybe make beginners learn and learn myself with deep analsyis!

I think that with AI that's kind of none productive.
Camille is a bot generator software not an analysis one, but when we plugged it to AI we found that doing statistical analysis of batch of game was very easy. For interpretation or finding pattern you can just ask GPT/Claude.

The future of chess analysis is probably some kind of data mining and putting in place software for doing precise analysis. GPT/Claude are just interpreting the data produce by said analysis. This being said just having GPT/Claude controlling a chess engine, analyzing a bunch of game and analyzing said data is already more powerful than most software available on the market...

@SF18NNUE-SFNNV12 said ^

i can annotate chess games from magazines , databases like TWIC , make them like a substack or publish them on blogs learn myself ,comparer with stockfish 19 :( ( a engine i hate cuz it makes feel outdated) and learn myself give my insights to others and they discuss and find my flaws or see where im right , and learn myself from blogs by like a FM i found named chessatlas and i like his blogs...

yes. do it.

@Himawari_Nohara said ^

You will soon become "jack of all, master of none" given the number of paths you think will help.

Elaborate please,I would like to know all I thought was of playing games,checking my statistics , analysing games thoroughly,doing puzzles,and annotating master game,reading books,watching videos what's wrong in doing all of that yk I'm not saying you're wrong but ,I don't see something bad in this, i think none of them are a waste of time, though i have priorities to arrange these and watching youtube videos is at lowest and correction- it was amazingAtlas no chessatlas my bad

@CamilleQueenOfTheBot said ^

I think that with AI that's kind of none productive.
Camille is a bot generator software not an analysis one, but when we plugged it to AI we found that doing statistical analysis of batch of game was very easy. For interpretation or finding pattern you can just ask GPT/Claude.

The future of chess analysis is probably some kind of data mining and putting in place software for doing precise anal...

i nevermentioned LLMs

also @Himawari_Nohara i think you earlier mentioned doing calculation training from agaards book, what do you think of this site i found? https://calculationtraining.com/

@SF18NNUE-SFNNV12 said ^

I think that with AI that's kind of none productive.
Camille is a bot generator software not an analysis one, but when we plugged it to AI we found that doing statistical analysis of batch of game was very easy. For interpretation or finding pattern you can just ask GPT/Claude.

The future of chess analysis is probably some ki...

Yes that's why I commented. I saw what you can do with hooking a simple chess engine with an AI. Very impressive.
Camille allow you to do batch analysis, in short you make a chess engine analysis the game and you check for elo variation between what the user played and the best move(s) . The you can classify using pattern ( fork, opening, isolated pawn ect ) .
This part is none AI .
But the AI is quite good at specific analysis. The main limitation would be to know what game to analysis and what question to ask

https://www.youtube.com/watch?v=pwKduhFg_LI

Until now those kind of tool were quite expensive. Now you could litteraly do it with the Claude/GPT app, give them the PNG, ask them to analysis with SF and to write you a report.

Here the main advantage of Camille is purely the wrapping, you have your library, a place to run the chess engine and a pre formatted report. But that's mostly cosmetic, information wise a paying GPT/Claude + the game as PNG or other format would be equivalent. And to use Camille this way, you need the paying GPT/Claude subscription anyway, the free tiers LLM would be a little too weak .

@CamilleQueenOfTheBot said ^

I think that with AI that's kind of none productive.
Camille is a bot generator software not an analysis one, but when we plugged it to AI we found that doing statistical analysis of batch of game was very easy. For interpretation or finding pattern you can just ask GPT/Claude.

The future of chess analysis is probably som...

is the LLM augemnted/hardcoded to follow stockifsh 19?, trusting LLMs is xtrmeley questionable in chess

@SF18NNUE-SFNNV12 said ^

I think that with AI that's kind of none productive.
Camille is a bot generator software not an analysis one, but when we plugged it to AI we found that doing statistical analysis of batch of game was very easy. For interpretation or finding pattern you can just ask GPT/Claude.

The future of chess analysis is ...

LLM are VERY bad at understanding chess, but LLM are very good at interpretating data coming out of SF. The video you see on YT of LLM being horrible at chess, are LLM playing chess.
If you ask the LLM to interprete data coming out of a chess engine, especially if you provide it in an organized fashion, the results are different.