Hacker News Reader

Comment by mod50ack | Hacker News Reader

Any license that discriminates based on use case would not qualify as open source under the Open Source Initiative definition, nor as free software under the FSF definition. You also shouldn't expect for your project/code to be reused by or incorporated into any free or open-source projects, since your license would be incompatible.

You can release software under whatever license you want, though whether any restriction would be legally enforceable is another matter.

View on HN

> Any license that discriminates based on use case would not qualify as open source under the Open Source Initiative definition, nor as free software under the FSF definition.

Freedom 0 is about the freedom to run the software "for any purpose", not "use" the software for any purpose. Training an LLM on source code isn't running the software. (Not sure about the OSD and don't feel like reviewing it.)

Anyway, you could probably have a license that explicitly requires AIs trained on a work to be licensed under a compatible free software license or something like that. Conditions like that are comparable to the AGPL or something, adding requirements but still respecting freedom 0.

But that's not an "anti-AI" license so much as one that tries to avert AI-based copyright laundering.

Depending on how the courts weigh in on the role of fair use in AI training, it's possible that a "copyleft for AI" clause would end up either redundant with the existing GPL, or legally void. It would be crazy complicated to enforce if it does hold water though.

Agreed on all points, unfortunately. :(

If training AI is a copyright exemption, and it is likely to be the case, then the license is irrelevant.

If it is not then the trained AI is a derivative work, which the license should allow as long as it is publishable under the same license to be considered open source or free software.

In any case, I don't think an anti-AI clause would serve a meaningful purpose on open source software. You can however make your own "source available" license that explicitly prevents its use on AI training, and I am sure that some of them exist, but I don't think it will do much good, as it is likely to be unenforceable (because of copyright exemptions) and will make it incompatible with many things open source.

Laws cannot be changed retroactively. So if AI training is a copyright extension that can only happen starting sometime next year. So the consequences of these companies' choices are already set in stone, even if they're not known yet.

The GPL requires that all materials to reproduce any derivative work be made available at cost (and all models can reproduce linux kernel GPL data structures, including the private parts, character-by-character). So do I get access to OpenAI's full training data?

Or do I get to make and publish Mickey Mouse cartoons by training an AI on Disney movies then publishing the model output. Hell, I could even make better versions of old Disney movies, competing with half of Disney's current projects!

It seems to me one of these must be true. So which is it?

Um, no. Copyright puts specific restrictions on what you can do with work. Those restrictions are described by certain words. The question is whether the existing restrictions cover training AI. That's a matter of interpretation, but once an interpretation is accepted, it is understood as what copyright always meant.

Training AI is probably not a copyright violation because it never was one to begin with.

The comments of the (German) judge in this case seem to indicate the judge doesn't understand why any of the defendants even thought training AI wasn't a violation (at least not when taken to the point it can exactly reproduce and create derivative works to existing works. Maybe that's why OpenAI is trying to make that harder now. Still trivial to make it violate that rule though).

https://www.dw.com/en/openai-loses-song-lyrics-copyright-cas...

Note that OpenAI has now testified that they indeed used copyrighted works to train their models. The outcome of the case is that both training AI models using copyrighted work and providing AI model outputs that are derivative of some copyrighted work are copyright violations, and would mean model owners have to respect licenses (ie. compensating the authors)

The case can still be appealed, so it is not final. On the other hand, if I'm reading WTO copyright treaty rules correctly, this ruling applies in the US.

In the US things seem to be going in a similar direction: https://www.publishersweekly.com/pw/by-topic/digital/copyrig...

Seems to me this can still easily go the way the authors want it to in the US. And in theory, it doesn't even have to, OpenAI lost. Yes, it can be fought on appeal, but I've always heard that winning an appeal after losing a case is 10x harder than winning that case in the first place. And we'll know in early January if OpenAI fights it at all, so it's not like they have a lot of time left.

It would not be discrimination to mandate that weights of any model trained in the code need to be released under similarly open license.

Leaving aside the sentence case in the title, the author's post didn't capitalise open source: they clearly mean source which is open to be read freely, and from the context this can clearly be read.

I disagree. They said open source, so I’ll take them at their word that they mean open source. If they meant otherwise, they should’ve said that instead.

This is a highly nitpicky topic where terms have important meanings. If we toss that out, it becomes impossible to discuss it.

I've linked elsewhere to the Hippocratic License, which freely refers to itself as open source while specifically being built around refusing licensing based on ethical considerations. OSI don't own the term open source, and the simple and plain meaning of the term is clear to see. Otherwise, we wouldn't consider GPL software to be open source, because it attaches conditions on usage. That even applies to non-copyleft licenses like MIT which demand author attribution. The term open source is best read literally unless someone says "I want an OSI approved license".

Free Software and Open Source are similar, but not identical: https://www.gnu.org/philosophy/categories.html

The GPL places no restrictions on how you can run the software. All meaningful licenses place restrictions — or, conversely, limit the permissions they grant — on how the code can be used, distributed, integrated with other projects, etc.

But I disagree that the meaning of Open Source is malleable. As others here said, if we want to make a new definition, we should make a new term. In my opinion, in this case, we have. It’s Source Available, which is basically “look, but don’t touch”. And as with other brightly colored things in nature, it’s generally best to avoid it.

The OSD came out within months of the phrase "open source" first being used, and the phrase was coined as part of the same process of discussion that produced the OSD. It's not a natural phrase and does not have an obvious "simple and plain meaning". It's a term of art.

Indeed it is, but you get it now, right?

  > the author's post didn't capitalise open source: they clearly mean

You can't make this conclusion. A lot of people simply don't bother with capitalizing words in a certain way to convey certain meaning.

A random "initiative" does not have the power to redefine words. If the source is available, it's open source.

that would be “source available” software, and it’s not a random initiative

there is disagreement on exactly what “open source” means, but generally clear boundaries between open source and source available software in licensing and spirit of the given project. e.g. MIT and Apache 2.0 are open source, BSL is source available.

edit: PERSONALLY, I think if you don’t welcome outside contributions, it isn’t open source; see others’ responses for disagreement on this (it’s not a part of the standard definition)

> if you don’t welcome outside contributions, it isn’t open source

That isn't true. Open source refers to the ability to make use of the source code if you wish, not the ability to send pull requests. SQLite is open source (public domain even!), but does not accept contributions from outside.

argh I will re-edit my comment…sorta covered by the “disagreement” bit, and I disagree on this point (it’s not open to me if you don’t openly accept contributions), but you are right

Indeed, and it can also be free software and under a copyleft license (GPL AGPL etc) and not accept contributions. Otherwise, every project that shut down or was just a one off gist/blog post to begin with couldn't be called open source either!

Open source means the source is open, ie downloadable. It's not that complicated, that can't just be made up to mean something else

it’s not made up (well all language is made up but I digress), you’re just being flagrantly ignorant of the terms you’re using and their history. you can easily go read up on open source vs source available and the history of the terms/licenses

it’s also fine by me if you want to have your own definition; see other comments, I don’t personally 100% agree with OSI’s definition myself

One can find leaked Windows source code on the Internet. Is it open source?

> it’s not a random initiative

Arguably it is, in the sense that they didn't actually invent the term; there are many documented pre-OSI uses (including by high-profile folks like Bill Joy) saying "open source" to just mean "source available". And OSI's attempt to trademark the term was rejected.

> if you don’t welcome outside contributions, it isn’t open source

That isn't even part of the OSI's definition, so what are you basing this on?

> That isn't even part of the OSI's definition, so what are you basing this on?

edited my comment —- that is my personal belief/definition

I did mention there’s disagreement —- I haven’t read up on the history and whatnot myself in a while. will have to do some re-reading :)

> PERSONALLY, I think if you don’t welcome outside contributions, it isn’t open source

It's not a question of belief. Maybe words don't mean anything anymore, but certainly legal contracts and licenses do. "Open Source" is a class of licenses approved by the OSI. There are no spirits involved.

meh…I consider an open source license distinct from an open source project. obviously legal contracts can define their own static terms; language is dynamic

The OSI definition of open-source software is recognized by several governments worldwide as definitive and legally binding. What you're describing is source available and that's a very different thing.

Which governments? OSI wasn't even granted a trademark for "open source" in the US, the country they are based in.

That makes complete sense though? They don't hold the IP, I don't really see any way they could be granted a trademark on it.

As for the list, see [0].

[0] https://opensource.org/about/authority

They applied for a trademark and were rejected due to the term being too generic/descriptive. It has nothing to do with whether they hold IP.

That list doesn't appear to be "legally binding" in a general sense; to me, the way you worded that implies "there is a law saying OSD is the definition of open source in this country" which is very far from the case.

Instead that list appears to be specific cases/situations e.g. how some US states evaluate bids from vendors, or how specific government organizations release software. And many things on that list are just casual references to the OSI/OSD but not laws at all.

A trademark is literally a form of IP. Clearly you don't know what you're talking about.

I didn't say a trademark isn't a form of IP. I said their application for a trademark was rejected due to "open source" being too generic/descriptive, not due to the reason you directly asserted above ("They don't hold the IP, I don't really see any way they could be granted a trademark on it").

You can read more about this at https://opensource.org/pressreleases/certified-open-source.p... or https://www.techmonitor.ai/technology/open_source_initiative... among many other sources. Or a much longer blog post from a lawyer who is active on HN: https://writing.kemitchell.com/2020/05/11/Open-Source-Proper...

fwiw, a non-OSI attempt to trademark "open source hardware" was also rejected for the exact same reason. https://opensource.com/law/13/5/os-hardware-trademark-reject...

Not all phrases’ meanings are derivable from the literal definitions of the words that make them up.

But this one definitely is!

The didn't redefine the words, they defined them. Anyone using them for anything other than the purposes they were defined to cover is a dishonest parasite who is intending to trade on the goodwill of the people who adhere to OSI's guidelines. "Open Source," capitalized or not, was not a common phrase before they introduced it. I don't care if somebody in 1965 said "I decided I'm going to be open, and share the source code!" Somebody sitting right next to him probably said "I decided I'm going to be open, and tell people that I will never share the source code."

Because prefixing something with the word "Open" to imply that it would be completely transparent (in any context) wasn't even common before the term "Open Source" was invented. When people do that, they're hoping that the goodwill that Open Source has generated will be transferred to them, and they are judged on that basis. "Open" generally had a slightly different meaning: honest.

> A random "initiative"

And when you play stupid, nobody respects your argument. It's self-defeating.

As a personal anecdote, while I’ve heard smart people say they were using “open source” way back when, I had personally never heard it used in any way before starting to used Linux and the BSDs in the late 90s, when OSI came along and people started discussing it in that context.

I can’t say others weren’t using it before then. I can say say that I first heard of Open Source after I’d heard of Free Software.

Yes, "Open Source" is newer than "Free Software". The phrases was deliberately coined, yes in the 1990s, to cover different (and mostly broader) ground... because there was a desire for a clear distinction between them. Which there still is.

Well, OSI didn’t coin “open source”. Factually, the term existed before OSI started using it. People have shared examples of isolated usage before then. However, they definitely brought it, and the modern definition, into common usage.

Like, if people had collectively used the term 23 times through 1996, then 837,000 times in 1997 or whenever OSI popularized it, I’m fully onboard with saying it’s their term.

Open Source means OSI-approved license in the software context. Some government examples of this being explicitly mentioned:

- Canada/British Columbia: https://www2.gov.bc.ca/assets/gov/government/services-for-go... - European Union (this applies to all EU member states): https://eur-lex.europa.eu/eli/reg/2024/2847/oj/eng - search for "Free and open-source software is understood" in the text - Germany (the EU definition already applies here, but for good measure): https://www.bsi.bund.de/DE/Themen/Verbraucherinnen-und-Verbr...

Words have meaning!

Your first link (Canada/BC) offers guidelines for BC government usage of open source software. In this type of situation, the OSI's list of approved licenses (and OSD in general) is very helpful, since it avoids massive duplicative legal overhead of evaluating software licenses. But in my opinion, that has little bearing on whether or not people should strictly follow this definition in an international public forum.

As far as I can see, your second link (applies to all EU member states) makes no mention of the OSI whatsoever, and uses a definition that is far briefer and less specific than the OSD.

I cannot evaluate the third link (Germany) as I don't speak German and automatic translation may introduce subtle changes.

Reply on news.ycombinator.com