Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The current state of the theory that GPL propagates to AI models

shujisado.org

Post reply on HN
Print view View original

301–310 of 314 posts

Re: The current state of the theory that GPL propagates to AI models

#301

Post by anticensor »

pclmulqdq wrote: post #124
iso1631 wrote: post #100

Earlier quoted context omitted.

How much do you pay them? If the GPL license didn't exist, and instead you just relying on copyright, then that's an injunction. You have to stop using the code you "stole" and pay reparations. In UK law, if you distribute copyright material in the course of a business you can be facing 10 years in prison and an unlimited fine. Sure you can't get them to agree to the GPL, they could simply stop distributing and then…

You pay them an amount determined by the court or your settlement, and you also have to stop using the code. This is how everything works. Corporations can't go to prison.

Corporations can have their business licence temporarily or permanently suspended.

Re: The current state of the theory that GPL propagates to AI models

#302

Post by dzaima »

CamperBob2 wrote: post #299
dzaima wrote: post #273

Earlier quoted context omitted.

The most difficult parts of getting readable code would be dealing with inlined functions and otherwise-duplicated code from macros or similar, and dealing with in-memory structure layouts; both pretty complicated very-global tasks. (never mind naming things, but perhaps LLMs have a good shot at that) That said, chatgpt currently seems to fail even basic things - completely missed the `thrM` path being possible here:…

You can't feed something like that to the free ChatGPT model and expect anything useful. Try these: https://chatgpt.com/s/t_6929f00ff5508191b75f31e219609a35 (5.1 Pro Thinking) https://claude.ai/share/7d9caa25-14f7-4233-b15c-d32b86e20e09 (Opus 4.5) https://docs.google.com/document/d/1C0lSKbLSZOyMWnGgR0QhZh3Q... (Gemini 3 Pro Thinking) All of them recognized the thrM exception path, although I didn't review them for co…

Those first two indeed look correct (third link is not public); indeed free chatgpt is understandably not the best, but I did give it basically the smallest function in my codebase that does something meaningful, instead of any of the actually-non-trivial multi-kilobyte functions doing realistic things needing context.

Re: The current state of the theory that GPL propagates to AI models

#303

Post by em-bee »

em-bee wrote: post #229
dragonwriter wrote: post #213

Earlier quoted context omitted.

> if AI generates something that is equal to existing code, then the license of that code applies. No, it doesn't, if the generation is independent of the existing code. If a person using AI uses existing code and makes a literal copy of it , then, yes, the copyright (and any license offer applicable in the circumstances) of the existing code may apply (it may also not, the same as with copies of portions of code mad…

if the generation is independent of the existing code well, that's the big question, isn't it? if the code is used for training AI and the AI reproduces the same code, is that really independent? i don't think so. Copyright protects against copying. It doesn't protect against someone creating the same content by means other than copying. if the code is the same, how do you prove it's not a copy? it's the same problem…

btw, this is the reason why people who at some point of time may have had access to windows source code are not allowed to work on wine. because if wine accidentally reproduces windows code, the only defense is that none of the contributors have ever seen that code before.

if AI has seen that code in training, then this defense is no longer possible.

Re: The current state of the theory that GPL propagates to AI models

#304

Post by CamperBob2 »

dzaima wrote: post #302
CamperBob2 wrote: post #299

Earlier quoted context omitted.

You can't feed something like that to the free ChatGPT model and expect anything useful. Try these: https://chatgpt.com/s/t_6929f00ff5508191b75f31e219609a35 (5.1 Pro Thinking) https://claude.ai/share/7d9caa25-14f7-4233-b15c-d32b86e20e09 (Opus 4.5) https://docs.google.com/document/d/1C0lSKbLSZOyMWnGgR0QhZh3Q... (Gemini 3 Pro Thinking) All of them recognized the thrM exception path, although I didn't review them for co…

Those first two indeed look correct (third link is not public); indeed free chatgpt is understandably not the best, but I did give it basically the smallest function in my codebase that does something meaningful, instead of any of the actually-non-trivial multi-kilobyte functions doing realistic things needing context.

Would be interesting to push the models with a couple of larger functions, if you have some links you'd like me to try.

I have paid pro accounts on all three, but for some reason Gemini is no longer allowing links to be shared on some queries including this one. All it would let me do is export it to Docs, which I thought would be publicly visible but evidently isn't.

Re: The current state of the theory that GPL propagates to AI models

#305

Post by m463 »

froh wrote: post #234
m463 wrote: post #167

Earlier quoted context omitted.

I don't think you understand the GPL. > "the user is allowed to do with the software" The GPL does not restrict what the user does with the software. It can be USED for anything. But it does restrict how you redistribute it. You have responsibilities if you redistribute it. You must provide the source code, and pass on the same freedoms you received to the users you redistribute it to.

first I thought you'd go into the nuance of gpl2 vs 3 or lgpl vs gpl vs agpl? patents, tivoization, cloud use? :-) I agree, I didn't make any statement what you can do with the software as long as you are licensed to use it you are allowed to build atomic bombs, nuclear power plants, tanks, whatever. but only as long as you comply i.e. give your downstream the freedom you've received. if you fail at that, you're no l…

> first I thought you'd go into

... I doubt that would clarify the clarity in clearness.

Re: The current state of the theory that GPL propagates to AI models

#306

Post by strogonoff »

klooney wrote: post #294
strogonoff wrote: post #240

Earlier quoted context omitted.

If LLMs are like human minds enough, then legally speaking we are abusing thinking and feeling human-like beings possessing will and agency in ways radically worse than slavery. What is missing in the “if I can remember and recite program then they must be allowed to remember and recite proframs” argument is that you choose to do it (and you have basic human rights and freedoms), and they do not.

We're halfway to Roko's Basilisk here

To be clear I’m not saying I believe that’s the case, merely noting that logically there are two options (either it’s like a human and then it can remember & recite but we can’t abuse it, or more likely it’s just a tool and then the freedom to “remember & recite” is simply not applicable to it, whatever it does is liability of its operator and user).

Re: The current state of the theory that GPL propagates to AI models

#307

Post by throwaway290 »

AnthonyMouse wrote: post #298
throwaway290 wrote: post #285

Earlier quoted context omitted.

like LLM or NFT or killer drones, malware isn't bad for somebody . it is always about who it is benefits the most. > the LLMs that aren't bad which LLM is not made by stealing copyleft code?

> like LLM or NFT or killer drones, malware isn't bad for somebody . Malware isn't bad for Russian crime syndicates, but we're generally content to regard them as the adversary and not care about their satisfaction. That isn't the case for someone who wants to use an LLM to fix a bug in their printer. They're doing the good work and people trying to stop them are the adversary. > which LLM is not made by stealing cop…

I see the mega wealthy in charge of LLMs who benefit the most from destroying copyleft and individual IP as adversaries)

> That isn't the case for someone who wants to use an LLM to fix a bug in their printer. They're doing the good work

they take advantage of temporary situation for good outcome but longer term they benefit those people doing shady stuff and concentrate power to them.

> Suppose you train an LLM only on GPL code, and all the people distributing and using it are only distributing its output under the GPL

That seems fair? but that's not what happens except by accident.

Re: The current state of the theory that GPL propagates to AI models

#308

Post by throwaway290 »

CamperBob2 wrote: post #296
throwaway290 wrote: post #285

Earlier quoted context omitted.

like LLM or NFT or killer drones, malware isn't bad for somebody . it is always about who it is benefits the most. > the LLMs that aren't bad which LLM is not made by stealing copyleft code?

You don't get to unilaterally make laws for the rest of us, which is what you are trying to do when you throw around terms like "stealing" in contexts where they have no legal meaning. Sorry. If the incumbent copyright interests insist on picking an unnecessary fight with LLMs or AI in general, they will and must lose decisively. That applies to all of the incumbents, from FSF to Disney. Things are different now.

The laws already exist. If you side with corrupt judges disrespecting these laws and interpreting them in favor of big tech corps, it's your choice)

Re: The current state of the theory that GPL propagates to AI models

#309

Post by boltzmann64 »

xorcist wrote: post #283
CamperBob2 wrote: post #146

Earlier quoted context omitted.

The GPL arose from Stallman's frustration at not having access to the source code for a printer driver that was causing him grief. In a world where he could have just said "Please create a PDP-whatever driver for an IBM-whatever printer," there never would have been a GPL. In that sense AI represents the fulfillment of his vision, not a refutation or violation. I'd be surprised if he saw it that way, of course.

The only legal way to do that in the proprietary software world is a clean room implementation. An AI could never do a clean room implementation of anything, since it was not trained on clean room materials alone. And it never can be, for obvious reasons. I don't think there's an easy way out here.

Google's engineers when they were copying Java API for Davlik (and later ART), they had access to and consulted Java source code. The infamous Oracle v. Google judgement siding Google set precedent at the highest level, SCOTUS that looking at the code is not an issue.

So, it doesn't matter if a AI can or cannot do clean room implementation. Unless it is a patent or trade secret violation, cleam room implementation doesn't matter.

Re: The current state of the theory that GPL propagates to AI models

#310

Post by gavinhoward »

em-bee wrote: post #173
gavinhoward wrote: post #142

Earlier quoted context omitted.

https://gavinhoward.com/2023/12/is-source-available-really-t...

just a comment on this article, that may be unrelated to the point you want to make: gavin makes a fatal mistake in interpreting RMS intent. he claims that he only wanted control over his hardware. that is not true. he also wanted the right to share his code with others. the person who had the code for his printer was not allowed to share that code. RMS wanted to ensure that the person who has the code is also allowe…

SAMS does do that. Read my article carefully; it also requires modification and distribution rights for users. See principles 0 and 1.
Post reply on HN