So, if I’ve got this right, AI tools attacked them, and this prompted them to buy AI tools?
Tech companies and creating the problems they sell the solution to, a pairing as old as time. Eg, I’ve always thought that Google probably surreptitiously attacks websites that don’t have Captcha, and then comes along and says Captcha would prevent the attacks.
“Attacking” is maybe too strong of an accusation, because for that they would need to be looking for weaknesses in websites and exploiting them. That’s too obvious, and as a monopolist Google has way more covert options at it’s disposal.
For instance, they may gently bias search results so sites that are not CAPTCHA protected appear more in queries made by bots, and less in those made by humans.
Another thing that has become way more prevalent since the AI bubble, is that scrapers indexing your site are sometimes indistinguishable from a DDoS attack. For this one, Google sells no solution, but Cloudflare does. Now, Cloudflare will not be DDoSing sites not covered by their service, but I have no doubt that they themselves are scraping all the public traffic, and selling it to all the AI companies. That way, scrapers know they don’t have to scrape protected sites, and focus their efforts on unprotected ones.
No, you can’t even comprehend the headline. They are switching to a different license. That is all. GPL not GPT
I’ve hated the prevalence of too-permissive licenses in the past 15 years. People need to go fucking brush up on their Stallman writings
In a perfect world we wouldn’t even need licenses, or just use BSD style ones, but alas… And thus moving to a copyleft license is the best move for sure.
AI bubble has shown that if you don’t have billions of dollars for lawyers, your IP will be copied and compressed into an LLM and you won’t be able to do anything about it, regardless of license. But at least Apple won’t be able to profit from Asahi team work and BSDs won’t be able to use drivers from Asahi. So much winning.
I really don’t understand why someone would use a permissive license in the first place. Especially a project this big.
Permissive licenses are what everyone should strive toward. A world where we can share advancements and technologies freely with one another and innovate based on the prior work of others is a world where a shit load of cool stuff gets done. The issue arises when bad faith actors take advantage of it and act against the spirit of collaboration. The problem isn’t caused by the permissive license, but by the bad actor, and the latter is what should be addressed
Well yeah, and the GPL was born exactly to preserve open source, sharing and transparency, and to fight against proprietary software
I don’t disagree - I’m not saying permissive licenses are necessarily the right call today :)
I mean stealing code was always possible, nowadays with AI and apparently copyright not existing anymore, nobody enforces anything anymore. They seem to only be enforced with actual copy pasting of code, which is not a “today” thing
By “today” I dont mean “post-ai” I mean “pre-revolution”
🤔🤔🤔🤔
A world where we can share advancements and technologies freely with one another and innovate based on the prior work of others is a world where a shit load of cool stuff gets done.
And how does a copyleft license get in the way of that? On the contrary, with the GPL, I can be sure that the technology and code will be freely shared. Freely as in freedom, that’s the whole point.
I misunderstood your initial point, I thought you were speaking against permissive/copyleft licenses generally. We agree with one another. Except in the “I can be sure that the technology and code will be freely shared” because just as bad actors can abuse a permissive license, bad actors (see: companies with enough money to just squash the legal case) can simply ignore a copyleft license. I think we 99% agree though and I love copyleft.
because just as bad actors can abuse a permissive license, bad actors
You are right, but with copyleft you at least have legal mechanisms to force the company to comply with the license. With MIT or BSD there aren’t any, because permissive licences don’t enforce any such restrictions.
I think we 99% agree though and I love copyleft.
Yup. And I love copyleft too.
Those legal mechanisms do exist in theory! But, as ever, if its you vs Microsoft, you can probably get fucked :(
When we get classless, moneyless, stateless society then sure, we won’t need licenses at all. But as long as we have a world run by lecherous corporate MBA’s, we need to use the current legal system to our advantage and keep our code safe from bad-faith actors who would steal from open source. The problem is not that corporations are greedy (they very well may be) but that the current relationship between the labor of writing code and the financial incentives of selling it is controlled by those bad-faith actors. They’re going to do everything they possibly can to leech off of open source and give nothing back because because it makes financial sense for them to do so, and until you change the rules of that relationship, you honestly can’t really blame them for it.
I agree somewhat with the sentiment of the first half of your reply. Though I don’t think that contradicts what I said. A permissive license may not be the best decision right now because you’re exceedingly likely to be taken advantage of, especially as LLM scrapers trawl the web for any scrap of code they possibly can. But I still think it is what we should be striving toward, revolution included.
As for the latter half, hmm no I think I can blame them for it actually. Obviously the system (see: capitalism) is the root cause but we don’t need to pretend that bad actors don’t carry any blame for their actions.
Isn’t AI and already rendering those licenses moot? AI data sets don’t care what type of license the code it scraped caries. and probably most AI code being generated for private companies is based on non-permissive licenses. yet nothing can be done about it.
I think the answer is “yes”.
This feels like a token response to a systemic unsolvable problem.
The article says itself talks about having models launder code so it can’t be considered a copy of the source.
to look at it another way: since anything an llm spits out is public domain, whoever commanded it do so is automatically in breach of the gpl unless they can show evidence that there is no copyleft material in the training data at all.
Afaik that’s not exactly so. I recall articles saying that significant input/prompt may influence that. Whether this is true or not I can’t say.
can’t be. if there is gpled code in the training data, all code must be gpl since the model is one work. and since it can’t, that’s a breach.
Yeah, in theory. In practice, good luck proving it.
I hate seeing this kind of doomerism around AI discussions. A future where we have no rights is not inevitable.
We know for a fact that OpenAI, Anthropic, Google etc trained on GPL code. They haven’t claimed otherwise, and this isn’t something they’re trying to hide.
We also know what the various open source models were trained on, since they’re open source.
We can pass laws that require LLM providers to disclose their data set in some way, eg through independent audits. Trump is the one who didn’t want AI regulation, and he basically just lost all ability to make that a reality. The final six weeks of the pedophile congress isn’t likely to accomplish anything on that front, and even if they do, the incoming congress can reverse it.
This is the time to energize people, not complain on the internet.
the question is if that theory is enough for a general ruling.
I recall articles saying that significant input/prompt may influence that. Whether this is true or not I can’t say.
That’s not true. Until these issues have run through the courts, we won’t know what is or isn’t legal. The biggest court case I’m familiar with atm is the New York Times’ case that has yet to be resolved, and even then the ruling there will probably be too narrow to decide anything for code licensing. Hell, OpenAI/Microsoft might even settle to avoid a bad ruling, leaving all those questions unanswered.
So the Tl;Dr is that anyone using LLMs to code (in any capacity) is taking a risk, since the legality of it is still uncertain. It’s not outside the realm of possibility that some court will rule that LLMs trained on stolen data are illegal, and the outputs of those models are infringing.
If you can compare the code and it matches you probably have a good case if the AI user can’t demonstrate it’s not coming from you.
There’s different kinds of datasets. Some respect, select and disclose licenses, others do not.
Some model trainers use any data they can, others are selective on what they train on.
It can also be a decision on principle rather than current or future enforcability. If you say “don’t copy this” and someone blatantly does, the morality is clearer than if you didn’t specifically say that.
There’s also a question of how court rulings and law may change the situation of enforcability and usage/ignorance risk.
MIT requires attribution too. GPL is, at the very least, a stronger statement of intention and demand. It can also be seen as protest or indication of a broader move of corporate AI exploitation leading to damage to the whole FOSS ecosystem, with additional restrictions being applied. Corps that would use AI now can’t use the MIT product anymore like before.
First, you can’t even compare the code if you don’t have access to it, you don’t have access to google’s or facebook’s code.
Second who can even legally challenge entities with pockets as deep as the ocean.
Third courts will probably classify AI generated code based of non-permissive licenses as fair use, so It’s a lost cause anyway.
Not being able to compare some proprietary is besides my point. Not all code is or remains hidden. I specifically wrote “if you can match” to remain neutral on that aspect.
Nothing is set. Until there’s a challenge in court, all you stated as certainties are just conjectures.
Why would it match? There’s thousands of ways to program any feature, the chance that it would be the same is tiny.
Copied code often matches, often look the same.
Why would a AI user copy the code though? They can just launder it through intelligence (brain and artificial both work) and the code will look different, but do the same.
For actual copying they wouldn’t need AI. They would just need “git clone”.
It would be interesting to see whether the BSD community finds a way to keep working with Asahi under the new terms.
They could go GPL too. :)
Asahi dual-licensing would be difficult, because then it’d be out there as MIT or BSD, which they don’t want. The BSD using custom-licensed parts - the BSD projects probably don’t want that.
I am not entirely sure why the GPLv2 license wouldn’t work for the BSDs, but either way I doubt that the problem was so big that it really required a license change. Everything in life is a trade-off, and I think the trade-off of not supporting the BSDs to prevent potential bad faith actors from “stealing the code without attribution” is a really bad trade-off.
I am not entirely sure why the GPLv2 license wouldn’t work for the BSDs,
The GPL is a copyleft license. Derivative and combined works must be published under the same license.
You can use BSD with GPL and then publish as GPL, but you can not publish as BSD.
I understand that the licenses are different. What I dont understand is why the BSD operating systems cannot make use of Asahi if it is GPL licensed instead of MIT licensed.
Because, as the comment you’re replying to says, if they include any GPL code, the whole combined work has to be licensed as GPL, and the BSD operating systems don’t want to be GPL.
“Sad to see Asahi out there burning bridges for no reason… relicensing m1n1 and drivers under GPL2-or-later, seemingly without the consent of all prior contributors, under the guise of “defending against AI bad actors”, seems entirely disingenuous and misguided.”
“no reason” but there’s reasons anyway and those are “entirely disingenuous and misguided”
Too bad [their] twitter posts don’t go into actual arguments. Why is it supposedly disingenuous? Why misguided? Are there better alternatives? Probably not.
Based
This won’t work. If they want to protect from AIs they need new license which probably won’t fit definition of open source by OSI. Also GPLv3 even won’t protect them.
I guess ai prompts in code comments can be… interesting?
Other options? Asking for a friend.deleted by creator
If you read the article, you’d see…
What the GPL changes is the commercial benefit. Under MIT, someone can take Asahi’s code, close it up, and sell a product, and that’s perfectly legal as long as the notice stays in.
Under the GPL, anyone who distributes a modified version has to release the source under the same license. You can’t legally build a closed, paid Apple Silicon distro on top of GPL code. And if a product’s source has to be public, the copying is easier to spot.
The article explains that the GPL license will enable legal action against future use along those lines - and that WINE did the same thing back in the 90s, and its worked for them.
it’s weird that they just phrase it as “has to release the source code”, it skips over the most important part. they have to release the source to any user who asks for it. if they want to be assholes about it they can request proof of purchase. unless the product is free, but then the question becomes “so what’s the problem again?”.
I deleted my comment a minute later because I didn’t want to discuss it any further.
But since you already replied: with ai rewriting, I don’t think that the switch to gpl changes anything. Even if the bad actors now publish their code, how would it change what they do?
Because it includes the modified code from the GPL licensed code and any derivative works, which in many cases may mean the source code to their entire application including what they may consider sensitive IP has to be provided or readily made available to customers who have paid for the software. Further to this any paying customer who receives this source code can make it all publicly available for free which can effectivelly kill the paid software.
