We now note that that we do not waive our attribution requirements for Spritely Goblins, which are present in Apache v2 (and for that matter, CC BY, CC BY-SA, and most FOSS licenses) https://codeberg.org/spritely/goblins/pulls/998/files
-
@copiesofcopies @bkuhn @trashheap My main concern is that the resulting legislative environment is as equitable for FOSS as it is for the proprietary world, whatever the outcome is. In many ways this is why I'm happy to see the vibe-laundered versions of Adobe's suite etc, I think that will push the issue.
@cwebber @bkuhn @trashheap unfortunately, we’re falling behind quickly, just in terms of the regulatory impact on who can afford to build competitive models. The CLAIM Act (just introduced) is basically designed to compel anyone who trains a model to either: (1) make it impossible for that model to accidentally cause harm, or (2) buy expensive insurance to cover the damage from inevitable harm.
-
@cwebber @bkuhn If I compile goblins and give the object file to someone, without attributing you -- as Apache v2 seems to allow sans NOTICE* -- and that person decompiles it, we've generated source code that is derived from goblins but does not attribute you. Seemingly legally.
Wouldn't the same apply to generating an LLM and using it no matter what you might ask in your README?
* ignoring moral rights
-
@cwebber @bkuhn @trashheap unfortunately, we’re falling behind quickly, just in terms of the regulatory impact on who can afford to build competitive models. The CLAIM Act (just introduced) is basically designed to compel anyone who trains a model to either: (1) make it impossible for that model to accidentally cause harm, or (2) buy expensive insurance to cover the damage from inevitable harm.
@cwebber @bkuhn @trashheap on the reverse-engineering side, we’ll see. Some interface and look & feel stuff is likely to be protected by copyright, and “clean room” efforts are probably misguided, because there’s no proving someone else’s model wasn’t trained on the target application.
-
-
@cwebber @bkuhn If I compile goblins and give the object file to someone, without attributing you -- as Apache v2 seems to allow sans NOTICE* -- and that person decompiles it, we've generated source code that is derived from goblins but does not attribute you. Seemingly legally.
Wouldn't the same apply to generating an LLM and using it no matter what you might ask in your README?
* ignoring moral rights
@joeyh @cwebber @bkuhn I find myself pondering what a (practical) "attribution-maximalist" license might look like. As I mentioned elsethread, for my fonts, I have a copyright notice that permits basically any use or modification, as long as 1. you not misrepresent the origin of the original (you don't get to claim you made my stuff), 2. you not misrepresent the origin of your modifications (you don't get to claim I made your stuff).
But there are some other considerations for non-font software, obviously. And I no longer share or distribute any other creative works than fonts (and those, I only send to friends).
-
@xinniw @cwebber I've drawn this comparison before, and I'm sure I will again, but the very first Easter Eggs in video games were... credits.
Why were they credits? Because Atari didn't want to include credit to the people who made the game. And you can't tell that story without it being intuitively obvious that Atari were being indefensible assholes. And now every video game has credits in it - most people won't pay too much attention to the contents, but the record is still important to the people who actually worked to make the thing, and to the people who love those people. The sky didn't fall from credits existing. The world got a little better actually.
But the capitalist incentive to alienate people from their labor, and consumers from the humanity of the people making their stuff, has never *stopped.* LLMs are just the next chapter of the story, the throughline of indefensible assholes that want to model labor as anonymous and interchangeable cogs, especially erasing minorities from history.
the capitalist incentive to alienate people from their labor, and consumers from the humanity of the people making their stuff
That seems like a collectivist incentive to me. The individual is merely a cog in the wheel of the collective.
Maybe it is just that assholes are assholes regardless of the economic structure they inhabit. -
Btw I have heard at least one prominent FOSS person (I'm choosing to not fill their notifications here by identifying them) that nobody cares about attribution requirements (and particularly not enforcing them). But to me, attribution is what I care about MORE than any other licensing requirements, more than copyleft "share-alike" style provisions, etc.
If you feel similarly, it would be good to hear on this thread. Perhaps even copying a similar notice into your project is a good idea.
@cwebber The ASF absolutely cares about the attribution requirements of ALv2. I am not involved with the ASF Legal team, but I am confident that they would be eager to enforce the attribution requirements against a clear intentional violation. I know that they have sent Stern Letters on such issues and as far as I can tell, those always work.
-
Btw I have heard at least one prominent FOSS person (I'm choosing to not fill their notifications here by identifying them) that nobody cares about attribution requirements (and particularly not enforcing them). But to me, attribution is what I care about MORE than any other licensing requirements, more than copyleft "share-alike" style provisions, etc.
If you feel similarly, it would be good to hear on this thread. Perhaps even copying a similar notice into your project is a good idea.
@cwebber "i just don't want someone to take my name off my code" is absolutely something i've said to people before.
-
Btw I have heard at least one prominent FOSS person (I'm choosing to not fill their notifications here by identifying them) that nobody cares about attribution requirements (and particularly not enforcing them). But to me, attribution is what I care about MORE than any other licensing requirements, more than copyleft "share-alike" style provisions, etc.
If you feel similarly, it would be good to hear on this thread. Perhaps even copying a similar notice into your project is a good idea.
@cwebber right; I still remember how proud I was when I first read a "Thanks to /me" in a changelog or bug report or something. and I try to follow this path myself.
(I do care about licenses but that's a different topic.) -
Btw I have heard at least one prominent FOSS person (I'm choosing to not fill their notifications here by identifying them) that nobody cares about attribution requirements (and particularly not enforcing them). But to me, attribution is what I care about MORE than any other licensing requirements, more than copyleft "share-alike" style provisions, etc.
If you feel similarly, it would be good to hear on this thread. Perhaps even copying a similar notice into your project is a good idea.
@cwebber Yes attribution is very important to me. Ideas and creations don't come out of nowhere. They evolve from prior work. Attributing reflects that reality. It helps to build trust and to connect people.
-
@cwebber having unwittingly become an expert in a certain cursed project, I say: attribution is SUPER IMPORTANT.
The cursed project when it was announced tried to hide they were based on ATproto, then launched and quietly went closed-source à la Truth Social.
And whenever they use other people’s tools built on ATproto, they are TERRIBLE at crediting them. Ouuuuuuf!
truthsocial is free software https://help.truthsocial.com/legal/open-source/ thanks to the @conservancy@sfconservancy.org who sued them, if i remember right.
CC: @cwebber@social.coop
-
Btw I have heard at least one prominent FOSS person (I'm choosing to not fill their notifications here by identifying them) that nobody cares about attribution requirements (and particularly not enforcing them). But to me, attribution is what I care about MORE than any other licensing requirements, more than copyleft "share-alike" style provisions, etc.
If you feel similarly, it would be good to hear on this thread. Perhaps even copying a similar notice into your project is a good idea.
@cwebber
#sharealike trained models should be qat trained models of all sizes as #openweights . Ttbh. -
@bkuhn @trashheap I appreciate the clarification on your position here.
I still do hold the positions that attribution are the glue that hold the commons together, and it's important that we fight to preserve it, and not give the impression that it is being voluntarily waived in the face of LLMs.
@cwebber
I absolutely agree attribution requirements are **not** waived. Also, IMO *copying* & *distribution* done to bring the software *into* a training set triggers all copyleft & non-copyleft requirements in #FOSS licenses.I worry, though, that FOSS community lacks sufficient funding to adjudicate that issue in Courts.
Fortunately, Big Tech & Big Content disagree on this. Big Content may get us good outcomes.
Big Worry: mandated compulsory licensing (as with music).
-
@cwebber @bkuhn @trashheap@tech.lgbt If bkuhn's right that permissive licenses make attribution hard to enforce, then the best way to accomplish that would be to use copyleft.
Anyway, if you keep Apache v2, it would certainly make sense to include a NOTICE file, since currently an Object derived from goblins does not need to attribute you at all. And as Apache v2 defines "Object", that seems to include AI models.
Yes, I do think copyleft is still going to be a tool that helps us to good outcomes in the age of LLM-gen-AI. I'm working on it.
Re: #Apache [sic] Software License: another reason to avoid it is that it's appropriated the name of North American native peoples.
Why anyone wants to use the #FOSS license equivalent of the “Washington Redskins” in 2026 is beyond me, especially when so many other licensing options abound.
Cc: @cwebber @trashheap
-
@cwebber @bkuhn @trashheap IMO the real challenge to FOSS enforcement against LLM producers is the same one everyone else has: if the training is fair use because the resulting model is transformative, the producer doesn’t need a license (and doesn’t have to respect the terms of any existing license). It’s still early early days in terms of court decisions, but this looks like the direction we’re headed in the US (though the outcome will likely vary depending on how much memorization a model can be shown to do).
∃ **so many** cases pending! It seems to me we can't guess the outcome at this point with reliability beyond like a 1-to-100 (or more) chance against the field.
Also, keep in mind the way most software developers use LLM-gen-AI now is to modify a work in front of them: `git clone` something, then work with the #LLMGenAI to modify it. I don't see how a fair use defense will allow those changes to dodge license requirements.
Also, FOSS licenses are contracts, too.
Cc: @cwebber
-
@copiesofcopies @bkuhn @trashheap My main concern is that the resulting legislative environment is as equitable for FOSS as it is for the proprietary world, whatever the outcome is. In many ways this is why I'm happy to see the vibe-laundered versions of Adobe's suite etc, I think that will push the issue.
The legislative environment will *never* be equitable for #FOSS under capitalism (& as such, it's never *been* equitable).
IMO, the advent of LLM-gen-AI is “just another day at the office” b/c I've spent ≈30yrs trying to get Big Tech to follow #copyleft requirements. They refuse at their whim.
Those refusals increased steadily (& occasionally exponentially, like now) since #GPLv2's release.
Meet the new infringer — same as the old infringer.
Cc: @copiesofcopies
-
∃ **so many** cases pending! It seems to me we can't guess the outcome at this point with reliability beyond like a 1-to-100 (or more) chance against the field.
Also, keep in mind the way most software developers use LLM-gen-AI now is to modify a work in front of them: `git clone` something, then work with the #LLMGenAI to modify it. I don't see how a fair use defense will allow those changes to dodge license requirements.
Also, FOSS licenses are contracts, too.
Cc: @cwebber
I expect the world looks very different from where you’re sitting than from where I am. Every time I look at my (FOSS-heavy) fedi feed, I’m struck by the contrast between the AI pessimism & skepticism here versus everywhere else I look.
LLMs have become part of everyday life for vast swaths of ordinary people. Skilled software developers routinely use them to produce code that’s far from slop. And probably most importantly, the AI industry is widely viewed as the most significant driver of U.S. economic growth and competitive advantage going right now.
All of those things surely sound negative to you, but they’re descriptively true. And against that backdrop, I see very little possibility that US courts will cut OpenAI et al off at the knees, even though we’ll surely see adverse rulings in cases of obvious misappropriation (Anthropic’s trove of pirated books, Sora reproducing entire songs from its training data).
-
I expect the world looks very different from where you’re sitting than from where I am. Every time I look at my (FOSS-heavy) fedi feed, I’m struck by the contrast between the AI pessimism & skepticism here versus everywhere else I look.
LLMs have become part of everyday life for vast swaths of ordinary people. Skilled software developers routinely use them to produce code that’s far from slop. And probably most importantly, the AI industry is widely viewed as the most significant driver of U.S. economic growth and competitive advantage going right now.
All of those things surely sound negative to you, but they’re descriptively true. And against that backdrop, I see very little possibility that US courts will cut OpenAI et al off at the knees, even though we’ll surely see adverse rulings in cases of obvious misappropriation (Anthropic’s trove of pirated books, Sora reproducing entire songs from its training data).
produce code that's far from slop
i would say perhaps code that doesn't look like slop or doesn't look sloppy, but if it is generated by an LLM, it is slop, and i think it is really important not to cede that point regardless of how impressive the output appears and even regardless of whether it happens to be good or correct in a particular instance. i will say more if you want, but my interpretation is based on what i've heard and read from linguistics professor emily bender (who is here on fedi too)
(in terms of the world looking different here than outside, i am probably proving your point and you might unfortunately be right though, but i also know that the media is pushing false and inflated narratives including about the prevalence of use, capabilities, and sentiments of users. it's not all false and many of us know people who are either true believers or deep down the pit, but i suspect it's not as widespread as we're being made to believe or to extrapolate based on a handful of personal anecdotes.)
-
produce code that's far from slop
i would say perhaps code that doesn't look like slop or doesn't look sloppy, but if it is generated by an LLM, it is slop, and i think it is really important not to cede that point regardless of how impressive the output appears and even regardless of whether it happens to be good or correct in a particular instance. i will say more if you want, but my interpretation is based on what i've heard and read from linguistics professor emily bender (who is here on fedi too)
(in terms of the world looking different here than outside, i am probably proving your point and you might unfortunately be right though, but i also know that the media is pushing false and inflated narratives including about the prevalence of use, capabilities, and sentiments of users. it's not all false and many of us know people who are either true believers or deep down the pit, but i suspect it's not as widespread as we're being made to believe or to extrapolate based on a handful of personal anecdotes.)
@tmw it’s always seemed to me that “slop” was a judgment about the quality of output. If it’s simply a value judgment about the source, i.e. code produced by an LLM is bad and “slop” and code produced by a human being is “not slop,” then I guess I’m not very interested in that conversation.
-
The legislative environment will *never* be equitable for #FOSS under capitalism (& as such, it's never *been* equitable).
IMO, the advent of LLM-gen-AI is “just another day at the office” b/c I've spent ≈30yrs trying to get Big Tech to follow #copyleft requirements. They refuse at their whim.
Those refusals increased steadily (& occasionally exponentially, like now) since #GPLv2's release.
Meet the new infringer — same as the old infringer.
Cc: @copiesofcopies
@bkuhn @copiesofcopies Of course the legislative environment has *always* been tailored to proprietary software. Without copyright (and related intellectual restrictions laws), we wouldn't have proprietary software!
What I mean is that Microsoft stepped in and set the tone with assumptions of "oh well it's okay to step all over and ignore the terms of the commons when we ingest them into the model, we would be upset if this happened to our stuff though"... for Microsoft, semi theoretical, but Disney and the record industries have actually done this: they're using AI for some projects, but also suing that their copyrights have been violated.
Copyright for me but not for thee type stuff.
(Which I know we agree about as a concern in this landscape.)
