Stockfish: Our lawsuit against ChessBase

[Ras]

This is not the case when the original work is only used for adjusting weights.

That's a technical implementation detail. If large parts are just copied because the models are incredibly large, it doesn't matter whether it's stored in weights, as text file, as lolcode or whatever. Again, do you really think AI companies would strike expensive deals with news outlets if the legal situation were as clear as you think?

[smatovic]

I look forward to FSF or some other entity creating a very restrictive license under which your code cannot be fed into any AI training system. Has anybody attempted to create such a license?

Based on what? Copyright cannot be used to prevent others from training their own brain or training their AI system on your code.

Judge Dismisses Lawsuit Over GitHub Copilot AI Coding Assistant
https://yro.slashdot.org/story/24/07/10 ... -assistant

The judge ruled that the plaintiffs failed to establish a claim for restitution or unjust enrichment but allowed the claim for breach of open-source license violations to proceed.

Copyright traces how far back, Gutenberg ~600 years ago? If copyright does not fit the new technologies, then we need new laws? Like there was DMCA:

https://en.wikipedia.org/wiki/Digital_M ... yright_Act

--
Srdja

[chesskobra]

The judge ruled that the plaintiffs failed to establish a claim for restitution or unjust enrichment but allowed the claim for breach of open-source license violations to proceed.

--
Srdja

I had seen the news but had missed this part. So there is a little hope? Long discussions are here https://news.ycombinator.com/item?id=40919253

[smatovic]

[...]
I had seen the news but had missed this part. So there is a little hope? Long discussions are here https://news.ycombinator.com/item?id=40919253

Point is, the legal system is highly selective when it comes to corporate interests.

Are algorithms patent-able? A question I asked in my CS studies 2008, I still have no satisfying answer.

I guess this AI neural network training topic will take a long time till settled, in courts, in legislative, in society.

--
Srdja

[chesskobra]

I believe algorithms are not patentable. There was controversy about Karmarkar's algorithm (Linear Programming). The following is from the Wikipedia page https://en.wikipedia.org/wiki/Karmarkar%27s_algorithm The Wikipedia (I think) is ambiguous in the last line of the following paragraph. The patent was likely for the implementation. Also, the title of the patent is "Methods and apparatus for efficient resource allocation" (not just linear programming). I wonder if the wording of the title ('apparatus'...) was chosen by lawyers.

>> At the time he invented the algorithm, Karmarkar was employed by IBM as a postdoctoral fellow in the IBM San Jose Research Laboratory in California. On August 11, 1983 he gave a seminar at Stanford University explaining the algorithm, with his affiliation still listed as IBM. By the fall of 1983 Karmarkar started to work at AT&T and submitted his paper to the 1984 ACM Symposium on Theory of Computing (STOC, held April 30 - May 2, 1984) stating AT&T Bell Laboratories as his affiliation.[16] After applying the algorithm to optimizing AT&T's telephone network,[17] they realized that his invention could be of practical importance. In April 1985, AT&T promptly applied for a patent on his algorithm.

[syzygy]

Copyright traces how far back, Gutenberg ~600 years ago? If copyright does not fit the new technologies, then we need new laws?

Do we need new laws? Copyrights exist to stimulate the creation of new works. AI creates new works. Why should we want to stop the creation of new works?

If a particular AI-created work happens to include creative expressions from an existing copyrighted work, then obviously the copyright on that work can be invoked.

[smatovic]

Copyright traces how far back, Gutenberg ~600 years ago? If copyright does not fit the new technologies, then we need new laws?

Do we need new laws? Copyrights exist to stimulate the creation of new works. AI creates new works. Why should we want to stop the creation of new works?

Good question. If the big players are free to use data for commercial neural networks, the open source community is too.

If a particular AI-created work happens to include creative expressions from an existing copyrighted work, then obviously the copyright on that work can be invoked.

--
Srdja

[towforce]

Copyright traces how far back, Gutenberg ~600 years ago? If copyright does not fit the new technologies, then we need new laws?

Do we need new laws? Copyrights exist to stimulate the creation of new works. AI creates new works. Why should we want to stop the creation of new works?

If a particular AI-created work happens to include creative expressions from an existing copyrighted work, then obviously the copyright on that work can be invoked.

Yes: the ability of machines to create intellectual property lowers the value of intellectual property.

I am strongly tempted to say that no human has ever created anything that was not derivative from previous ideas, but saying that would provide scope for argument - so what I'm ACTUALLY going to say is: humans having ideas that are in no way derivative of previous ideas is rare. I don't think there's scope to sensibly argue with that.

[Uri Blass]

Copyright traces how far back, Gutenberg ~600 years ago? If copyright does not fit the new technologies, then we need new laws?

Do we need new laws? Copyrights exist to stimulate the creation of new works. AI creates new works. Why should we want to stop the creation of new works?

If a particular AI-created work happens to include creative expressions from an existing copyrighted work, then obviously the copyright on that work can be invoked.

There is an assumption that copyrights stimulate the creation of new works.
It is not something obvious.
Of course there are people who create new work only because of copyright but not all of them and it is possible that
there are people who do not create new work that is based on some old work in the first place because of copyrights.

You can claim that we are not going to have the old work in the first place without copyrights but it is not clear.
Suppose the only way to generate new free code is to make it public domain(and nobody can release it with GPL because you cannot put limitation to other people how to use it).
I guess at least part of the people who generate free code that is not public domain are still going to do the same when they prefer the code to be not public domain but if nobody can do it they still prefer to release the code.

[syzygy]

There is an assumption that copyrights stimulate the creation of new works.
It is not something obvious.

It is extremely obvious. Hollywood could not exist without copyright. There are tons and tons of other examples.

Of course there are people who create new work only because of copyright (...)

So you agree that it is obvious.