CODA now has a released page.

[Graham Banks]

First, thank you to Andrew Grant and others for the feedback on the 0.9.0 pre-release. You were right - there were things I got wrong on the licensing side, and I'm grateful it was raised.

Over the past few days I've done (with the help of Claude!) a proper audit of Coda's codebase and its dependencies, and a fairly big clean-up: removing or independently reimplementing anything we found that leaned too closely on another engine's code, and tidying up attribution. The full write-up is here: https://github.com/adamtwiss/coda/blob/ ... 6-07-13.md. I believe Coda is now in a much better place licence-wise. If you think I've missed something — or have any other feedback — please open a GitHub issue. There's been some lively debate here - and I really welcome civil/constructive feedback - but please remember I'm genuinely just trying to do the right thing here.

I've also put out 0.9.1 https://github.com/adamtwiss/coda/releases/tag/v0.9.1, which folds in those licensing fixes along with an updated network, a ~30% NNUE speed-up on ARM (aarch64), and a few important bug fixes — including a correction-history fix that adds some elo. This releases also fixes the EGTB bug loading from paths with spaces in that was reported here. This is still a very early release, so I expect there are more rough edges. Feedback/bug reports welcome.

Thanks again,
Adam

Is the net embedded in the exe, or do we need to download it separately and put in the same folder?

How often are you intending to make new releases?

[adamtwiss]

Is the net embedded in the exe, or do we need to download it separately and put in the same folder?

How often are you intending to make new releases?

The release binaries all contain the embedded net, as would any source builds built with ‘make’. Just download the release binary and it should work fine. If you have any probs let me know.

New releases will be as and when I have something worthwhile releasing. I’ve not really thought about it, but I expect every 2-4 weeks given that Coda is still rapidly evolving. I suspect external testers will uncover bugs that I haven’t.

[Modern Times]

New releases will be as and when I have something worthwhile releasing. I’ve not really thought about it, but I expect every 2-4 weeks given that Coda is still rapidly evolving. I suspect external testers will uncover bugs that I haven’t.

Sounds pretty normal for a new engine.

[AndrewGrant]

First, thank you to Andrew Grant and others for the feedback on the 0.9.0 pre-release. You were right - there were things I got wrong on the licensing side, and I'm grateful it was raised.

Over the past few days I've done (with the help of Claude!) a proper audit of Coda's codebase and its dependencies, and a fairly big clean-up: removing or independently reimplementing anything we found that leaned too closely on another engine's code, and tidying up attribution. The full write-up is here: https://github.com/adamtwiss/coda/blob/ ... 6-07-13.md. I believe Coda is now in a much better place licence-wise. If you think I've missed something — or have any other feedback — please open a GitHub issue. There's been some lively debate here - and I really welcome civil/constructive feedback - but please remember I'm genuinely just trying to do the right thing here.

I've also put out 0.9.1 https://github.com/adamtwiss/coda/releases/tag/v0.9.1, which folds in those licensing fixes along with an updated network, a ~30% NNUE speed-up on ARM (aarch64), and a few important bug fixes — including a correction-history fix that adds some elo. This releases also fixes the EGTB bug loading from paths with spaces in that was reported here. This is still a very early release, so I expect there are more rough edges. Feedback/bug reports welcome.

Thanks again,
Adam

So a little thought exercise for Adam here.

In your audit of the licenses, you identified that Integral was not under any such permissive open-source license. As a result, you have authored some commits which remove references to Integral in the comments. This suggests to me that you understand that having Claude consume content from Integral, to produce some output for Coda, is in violation of the right's of Integral's author. This is a tacit admission that you don't think Claude can consume private content, and have its output be free of obligations.

Okay. But then in your efforts to rectify the license issues, in particular for the AGPLv3 engines, you had Claude identify code sections that were clear derivative works of Reckless. You then instructed Claude to rewrite them. But this is not a clean-room design process, clearly. It is like asking an Intern to shuffle up the code so that it still does the same thing, but looks different than the original. That process very clearly does not free you of the obligations you picked up by publishing AGPLv3 content.

But you're acting in a contradiction to this a bit. On one hand, you pull back and say "Oh no, I've got work product based on Integral, a fully right's reserve entity, we have to remove things". On the other hand, you're saying "Well Claude consumed AGPLv3 code to make this, and since it still looks a bit similar, lets shuffle it".

Either you believe you have NO obligations to ANY of the engines that Claude produced Coda based off of OR you believe that you have obligations to ALL engines that Claude produced Coda based off of now and in the past.

I am a just going to open a Pull Request with the changes needed to rectify the license issues ( except for Integral, that can't be fixed )

[Graham Banks]

But this is not a clean-room design process, clearly. It is like asking an Intern to shuffle up the code so that it still does the same thing, but looks different than the original.

Isn't this fairly common practice? Rewriting code differently to use an idea garnered from elsewhere?

[AndrewGrant]

But this is not a clean-room design process, clearly. It is like asking an Intern to shuffle up the code so that it still does the same thing, but looks different than the original.

Isn't this fairly common practice? Rewriting code differently to use an idea garnered from elsewhere?

Well. Most famous example in Tech is probably IBM's BIOS. Which was copyrighted, and you could not legally copy it to build IBM-compatible PC's. Compaq ran a clean-room reverse engineering process. They had one team study the IBM bios, and draft a specification about it, which defined the behavior, but not the actual implementation. Then they had a separate team -- no shared members -- implement a new BIOS based on that spec. A company called Phoenix Technologies did the same, creating their own clean-room BIOS that was compatible with IBM. IBM tried and failed to sue both. Massive legal teams were involved in said clean-room process, to ensure everything was done proper.

Many other companies tried to clone the IBM BIOS, got sued, and ended up settling. They did not do this clean-room process. Or if they did, they did not do it well enough to stand up to legal scrutiny.

The problem with calling what Claude has done a clean-room, is that Claude has both reviewed the original material to create the specification, and been the one to implement the specification. MAYBE it was two separate instances of Claude, to get a little better, but that still raises red flags.

[mar]

Isn't this fairly common practice? Rewriting code differently to use an idea garnered from elsewhere?

I don't think you can copyright algorithms? You can either keep them secret (=don't go open source) or try to issue a patent.

sadly nobody asked the authors - I believe AI training should've been opt-in, not opt-out (even here you have no guarantee),
but it's too late to fix that
what we're witnessing is normalization of IP theft - and this goes way beyond silly computer chess

that being said I'm done with open source

[Rebel]

But this is not a clean-room design process, clearly. It is like asking an Intern to shuffle up the code so that it still does the same thing, but looks different than the original.

Isn't this fairly common practice? Rewriting code differently to use an idea garnered from elsewhere?

You got it.

Said in other words, what Claude does openly is happening in silence big time for more than a decade.

And nobody cares.

[chrisw]

But this is not a clean-room design process, clearly. It is like asking an Intern to shuffle up the code so that it still does the same thing, but looks different than the original.

Isn't this fairly common practice? Rewriting code differently to use an idea garnered from elsewhere?

You got it.

Said in other words, what Claude does openly is happening in silence big time for more than a decade.

And nobody cares.

It’s an eco system of silent complicity.

[leslies]

But this is not a clean-room design process, clearly. It is like asking an Intern to shuffle up the code so that it still does the same thing, but looks different than the original.

Isn't this fairly common practice? Rewriting code differently to use an idea garnered from elsewhere?

You got it.

Said in other words, what Claude does openly is happening in silence big time for more than a decade.

And nobody cares.

It’s an eco system of silent complicity.

thanks for this engine,it is a very strong monster.one thing though,i cant get any table bases to work,keep getting error message.