CODA now has a released page.

[AndrewGrant]

But as Chris states above, it's completely clear there's no verbatim code copying.
So what's the issue?

Talkchess seems to fail to understand that even if you don't copy-paste chunks of code, you can still have a derived work of a GPLv3 project. If I line-by-line rewrite an engine in a different language, I am not magically free of the license obligations of said engine. What Coda has done is illegally consumed MIT/GPLv3/AGPLv3 content to produce a derivative work, and then 1) re-licensed it ( which he has no authority to do ), and 2) distributed it without proper notice to the source material.

Under talkchess' legal expertise, I should direct Chess.com to rewrite Stockfish in Rust. When we are done, we will have an identical engine, yet somehow "magically" be free of license constraints. Then we can proceed to improve Stockfish in tandem, and market ourselves as the #1 engine. How "exciting".

[Graham Banks]

But as Chris states above, it's completely clear there's no verbatim code copying.
So what's the issue?

Talkchess seems to fail to understand that even if you don't copy-paste chunks of code, you can still have a derived work of a GPLv3 project. If I line-by-line rewrite an engine in a different language, I am not magically free of the license obligations of said engine. What Coda has done is illegally consumed MIT/GPLv3/AGPLv3 content to produce a derivative work, and then 1) re-licensed it ( which he has no authority to do ), and 2) distributed it without proper notice to the source material.

Under talkchess' legal expertise, I should direct Chess.com to rewrite Stockfish in Rust. When we are done, we will have an identical engine, yet somehow "magically" be free of license constraints. Then we can proceed to improve Stockfish in tandem, and market ourselves as the #1 engine. How "exciting".

When others take ideas from open source engines without copy/pasting chunks of code, isn't that exactly the same, as long as they give credit where due?

[adamtwiss]

What Coda has done is illegally consumed MIT/GPLv3/AGPLv3 content to produce a derivative work, and then 1) re-licensed it ( which he has no authority to do ), and 2) distributed it without proper notice to the source material.

I'm here to have fun building a genuinely strong engine and to help push the field forward — and to do that in a way that adds to this community, credits people where it's due, and respects copyright/licensing. Everything I build on this project, I want to share back.

To get there, my Claude bots and I have studied a lot of engines to learn what good looks like - but let's be honest, most of us have looked at Stockfish's source at some point. That on its own doesn't make everything downstream a derived work.

If I've got something wrong on licensing or attribution, tell me and I'll review and fix it — I'd genuinely rather get that right.

Agentic coding and AI tooling open up real new opportunities and real new questions, and that isn't going away. I'd rather help work out how to do it responsibly and well, in the open, than pretend it isn't here.

[AndrewGrant]

But as Chris states above, it's completely clear there's no verbatim code copying.
So what's the issue?

Talkchess seems to fail to understand that even if you don't copy-paste chunks of code, you can still have a derived work of a GPLv3 project. If I line-by-line rewrite an engine in a different language, I am not magically free of the license obligations of said engine. What Coda has done is illegally consumed MIT/GPLv3/AGPLv3 content to produce a derivative work, and then 1) re-licensed it ( which he has no authority to do ), and 2) distributed it without proper notice to the source material.

Under talkchess' legal expertise, I should direct Chess.com to rewrite Stockfish in Rust. When we are done, we will have an identical engine, yet somehow "magically" be free of license constraints. Then we can proceed to improve Stockfish in tandem, and market ourselves as the #1 engine. How "exciting".

When others take ideas from open source engines without copy/pasting chunks of code, isn't that exactly the same, as long as they give credit where due?

It is not the same. The difference is a human vs an LLM.
Some even argue that even a human cannot view an open-source engine without incurring such obligations.