If you care about your users, you should always prefer some copyleft license.
Prefer permissive licenses only if you want to subsidize big business. Hopefully you have a plan for when they decide to take it closed-source and outcompete you. You probably don't, though, since that's not mentioned in the aggressive "use permissive licenses" marketing campaign they've been running for the last decade or two.
Permissive licenses can fit in a niche for software that has already been implemented many times. But the downsides remain, and why aren't you just using the existing code directly? The main occasion on which I will consider it is when everybody else is doing it wrong (and content to do so), and I as a user keep getting annoyed at hitting those limitations.
The second-worst reason to use permissive licensing is arguably "I don't understand how to dynamic linking".
> If you care about your users, you should always prefer some copyleft license.
There is no possible way for users to be hurt by a permissive license. Even if some company were to make a closed source fork, who cares? The original is still there, still just as free as it ever was.
> Prefer permissive licenses only if you want to subsidize big business. Hopefully you have a plan for when they decide to take it closed-source and outcompete you.
I prefer permissive licenses because I'm making a gift to the commons. It strikes me as hypocritical to say "I'm giving this to you, but only if you things the way I approve of". Thus, I make it completely free for all users, whatever they want to do with it. You may not see it that way, which is fine. But I'm sick of the false idea of "you're just subsidizing big business" being promulgated. That isn't true.
And why would I care if someone "outcompetes" me? My gift to society is still there to be used if people want, or not if they don't want. It doesn't diminish or harm my efforts in any way.
> Even if some company were to make a closed source fork, who cares?
They can add sufficiently popular functionality to said closed source fork and make the open source original a) obsolete and b) incompatible with the combined ecosystem, and thus deprive the users of a feasible free option.
If the closed fork functionality is superior enough to make the original de facto obsolete then the users have already collectively decided that the tradeoff is worth it.
And if the original can't compete it means the additional functionality was only going to exist because the financial model of the closed fork could pay for it.
> And if the original can't compete it means the additional functionality was only going to exist because the financial model of the closed fork could pay for it.
This completely disregards the fact that the "financial model of the closed fork" explicitly chose to build upon the permissively licensed original.
If the company chooses to build upon free software, they should be obligated to give back to the community from which they leech. Otherwise, they should just build their own thing from scratch with all the money they've hoarded, and keep it closed.
> If the closed fork functionality is superior enough to make the original de facto obsolete then the users have already collectively decided that the tradeoff is worth it.
Users cannot be trusted to make their own decisions about these kinds of things for the same reason that corporations cannot be trusted to be environmental stewards and children cannot be trusted to select a dinner menu or file taxes.
> There is no possible way for users to be hurt by a permissive license. Even if some company were to make a closed source fork, who cares? The original is still there, still just as free as it ever was.
The users of the closed source fork are harmed. They are unable to use the software to the full extent (e.g. inspect, modify, fix bugs). This is what copyleft intends to protect.
Closed source reduces the software into an appliance and the user into a consumer.
I have a different take on permission license. I offer up what little code I produce under the assumption that anyone can use it for whatever, just give me credit. Why? Because it might result in an interesting collaboration on a project I like. If someone uses my existing code to accomplish something neat and sees who I am, they can reach out to me for more, which I'm open to since it would likely result in some growth for me as an amateur. I'm not particularly interested in making money since I recognize that I am nowhere near that level, yet, but creating a future opportunity seems like a good idea since I enjoy coding.
> But I'm sick of the false idea of "you're just subsidizing big business" being promulgated. That isn't true.
All the permissively licensed code that came with my mac disagrees. This is even more true of iphone software: you can technically look at some of the code that comes with it, but you can't modify it.
> you can technically look at some of the code that comes with it, but you can't modify it.
This is equally true of the GPLv2. The attempt to close this loophole - Tivoisation as Stallman called it - only won him a lot of scorn from Linux land and a refusal to adopt the GPLv3!
Yeah, the idea here seems to be, iPhones would be open hackable devices if only more people would use copyleft licences, but in reality Apple would still lockdown their gadgets and just write the software themself if no permissive licence is avaiable.
> in reality Apple would still lockdown their gadgets and just write the software themself if no permissive licence [were] available
This is a common opinion, but seems uninformed to me. The free code helped them get started. They are Goliath now, but the benefitted from not having to do the R&D earlier.
I don’t know if they could have done the R&D. I definitely imagine it would not have been as good as the product they got from using BSD.
Anyway, I believe this kind of thinking is a shallow dismissal of the value that companies receive from starting from Libre software.
"Anyway, I believe this kind of thinking is a shallow dismissal of the value that companies receive from starting from Libre software."
It is a dismissal of the idea, that using a different licence will magically transform the whole industry.
If the code is open, the profit orientated companies don't need to do R&D from scratch. They just take the copyleft source code and make it closed, either by simply not publishing it, or if they care about whistleblowers by rewriting it. Which is easy if you have sources of something working. Even more so today with LLMs.
That is why I am a fan of permissive licences. It is a working compromise for companies who would never consider copyleft code in critical parts of their product.
Compromise for me is, the concept of open source gets more widewspread. There is no must in giving back.
And if you try to force it with copyleft .. you won't get far. I believe history is on my side here.
Android is based on copyleft code and it is far from an open hackable system. On some devices it possible to install an own build but it requires skills and patience so not an option even for most IT professionals. But for most devices there are locked loaders and closed drivers/firmware. I don’t see a copyleft license helping if a business is not willing to participate in open-source.
As a reminder, the GPL (or any other Free Software license, by definition - it's "Freedom 0" in the list) places absolutely no restrictions on users. Its provisions only kick in for those who want to go beyond merely using the software.
That's not quite true of AGPL (and for this very reason some companies, e.g. in the database provider space, use it for free versions of otherwise paid products)
Section 13 of the AGPL (the part that's not in the GPL) explicitly begins with "[notwithstanding,] if you modify the program [and allow users to access it over the network]". Merely using the upstream version is no burden.
> Never forget that Industry itself came to America by way of direct theft.
Two other examples come to mind: John Rolf arrived at the Jamestown settlement in 1612, having purchased tobacco seeds presumed to have originated in Spain (where their export was supposedly punishable by death, though I haven’t found sources confirming that).
William de la Barre adapted the rollers used in Minneapolis wheat mills from a European design. I have seen a few sources claim this was industrial espionage wherein Barre obtained employment in the mill and sketched its design out to copy it for his employers in Minneapolis, though Wikipedia makes no mention of this.
The American publishing industry also got its start by refusing to recognize British copyrights; this didn’t change until 1891.
This is to say nothing of the more explicit, above-board technology transfers involved in bringing European experts in to develop the American industrial plant. Similar things are currently going on in Saudi Arabia and China, and to a lesser extent in South America and Africa.
And then you have Operation Paper Clip and the looting of the German industrial plant after World War II.
> Prefer permissive licenses only if you want to subsidize big business.
Another reason to prefer permissive licenses is if you want anyone to use your project.
Most coding out there in the world is done in the service of a commercial enterprise and my experience is that companies won't use copy left projects. Companies are generally happy to use projects with permissive licenses or pay for SaaS or commercial licenses, but anything with the GPL is a non-starter.
There are a few, well known exceptions (Linux, GCC), but I can't think of any exceptions from projects started in the last 20 years.
I don't necessarily think this view is entirely rational or irrational, but I’m confident this attitude exists; companies using GPL projects are pretty rare. Anyone found differently?
There are alternatives. We can dual license software under a strong copyleft and a proprietary license, so that users (developers) who don't want their derived software to be copyleft can pay the proprietary license fee. IMO this is the approach that everyone should be taking - it rewards the developer rather than them subsidizing your derived works with their free time.
Obviously this has constraints. The developer must have full rights or ownership to redistribute the work under other licenses. This can be obtained from contributors via agreement, where every developer retains ownership of code they write, but grants permission to relicense it for the benefit the project. Ideally, contributors would also be rewarded for their time via bounty systems, contracts or direct employment.
It's possible to make non-derived works which utilize copyleft software. The license's conditions apply only to the derived works, and not directly to your code unless it unquestionably a derived work such as a fork. If your application is written in a way that it can function without dependency on a copyleft software, but can be augmented using copyleft software using a plugin system, then you're able to distribute your software under any license you chose, and only the plugin which uses the copyleft software would be required to be licensed under a compatible license. The "aggregate" would need to be distributed under the copyleft license - but there are also ways around that - simply distribute your software without the dependency on the copyleft work, and have it optionally download the copyleft plugins separately so that they're not part of the "aggregate".
Permissive licensing won't sustain itself unless the developers who depend on it start donating to the maintainers of the permissively licensed works they depend on. In some cases this happens, but mostly it's just the case that you can make money using software other people have given away and continue to maintain for free. What happens if they decide to stop maintaining it? Is your product still viable if you have to take over the task, or employ someone new to maintain it? Would it be viable if you, or your users, had to purchase a proprietary license fee for an alternative dependency?
You need to read the GPL licenses more carefully. Your code gets infected by GPL when you link. The principle difference between GPL 2 and GPL 3 being that the definition of "link" is much more expansive GPL 3. So separately downloaded packages don't really cut it unless the packages in question are LGPL-licensd (GPL with a link exception).
GPL 2.0 and GPL 3.0 (to a much more significant extent) disallow use of plugins. Or well, the do allow plugins, but at the cost of infecting the entirety of your own codebase with GPL 3.0 obligations.
And that's my principle objection to GPL. Linking to GPL code infects my code, requiring me to also license my code via a GPL license.
Permissive license stand a better chance of getting funding because they allow commercial users to use your software, and commercial users are more likely to give sustaining grants to projects they depend on than mere-ordinary users.
The GPL doesn't "infect" any code. You can write code which clearly calls GPL licensed code and license it under a MIT license.
The GPL comes into play when you distribute. If you distribute your MIT licensed code alongside the GPL licensed code, then the aggregate must necessarily be GPL licensed. This doesn't mean your code suddenly becomes GPL code and cannot continue to be licensed under MIT when not distributed as an aggregate. It is perfectly reasonable for someone to take the code, strip out the GPL licensed parts and replace them with MIT licensed alternatives, and license the whole thing as MIT.
Obviously for proprietary software this implies the aggregate must have its source code available in a GPL-compatible form.
But the point here is that you're separating the GPL work from your own code by means of a plugin system. If a plugin depends on GPL code, then the plugin itself must be distributed as GPL - but this does not infect the host program which loads the plugin.
That is, unless it can be reasonably argued that the host program is a derived work of the GPL licensed work. If the program basically does not function without dependency on the GPL licensed work then this might be a reasonable argument - but if the program runs fine, and is only augmented by a GPL licensed plugin, then it is not a derived work.
> GPL 2.0 and GPL 3.0 (to a much more significant extent) disallow use of plugins. Or well, the do allow plugins, but at the cost of infecting the entirety of your own codebase with GPL 3.0 obligations.
> And that's my principle objection to GPL. Linking to GPL code infects my code, requiring me to also license my code via a GPL license.
No. You release a host application with no dependence on GPL code.
The end user loads a plugin that is (or uses) GPL code. The end user doesn't distribute anything.
In very practical terms, what is the enforcement here? Who is gonna sue the end user?
Your logic seems to assume that there two disjoint groups of people:
1. Developers who work on the open source project
2. Developers who work on proprietary software based on the open source project
In reality these groups overlap. Developers contribute to the open source project, and they also work on their own software based on the open source project. If you look at the git logs of successful permissive open source projects, you'll see that a lot of the development time comes from developers who work for proprietary software companies.
That's true for some projects, but there are others who have a sole maintainer, receiving no other contributions or donations, who are expected to continue maintaining the work for free.
Users lose because the tools with the most functionality, the applications with the most interoperability, the software with the most mindshare,... will be proprietary software.
I think you misunderstood. My software is not permissive. My software is closed source, but based on a big open source project with a permissive license. It's a tool that helps users of the big open source solution.
I can sell licenses to my tool because it is closed source.
If the big open source project was licensed under a copyleft license, I would have to make my tool open source as well. Then I would have a hard time selling licenses.
I couldn't have spent the last 10 years building a tool for the open source software, and users of the open source software would have had fewer tools available.
Your personal financial incentives < humanity. I would rather you had a harder time selling licenses if it meant the ecosystem around whatever big project you're talking about was compelled to contribute to the commons in the same way that the big project does.
Proprietary software is the dead or soon to die branches of the tree of human development, may spawn some useful insights that will ultimately be applied elsewhere, and may be profitable and useful in the short term, but the branch dies, I don't want to subsidize dead branches, and I don't want to write software that enables people to make more of them, even if some of those branches might be nice to hang on for a while.
You can sell GPL software.
You can even sell GPL software under proprietary license as long as you are the copyright owner (dual-licensing).
Really, when you take time to understand real restrictions, as opposed to repeated myths, the only argument in favor of permissive license is that they help big players use free labor.
> You don't even need to be the copyright owner to sell GPL software
You do, if sell means also allowing another vendor to use your software as a part of their proprietary software.
GPL alone, if you're not the copyright owner, would not allow you to sell software in this way. But if you are the copyright owner, you can bypass GPL and sell a proprietary license to the vendor saying "you can use this software in your proprietary product if you pay me $XYZ".
For anyone who thinks I'm talking out of my ass, consider Qt.
If you don't own the Qt copyright, you cannot sell Qt to a company who wants to use it to build a proprietary product - since Qt is GPL licensed, using it in a proprietary product would not be permitted.
But since The Qt Company owns the copyright, they can sell their own software to companies that want to build a proprietary product with it - they just sell it under a proprietary license, specifically tailored for this. The fact that it was previously released under GPL is irrelevant, because the copyright holder can license it under as many licenses as they want.
If you don't own the copyright to it, you can only """sell""" GPL software in a manner in which local DVD shops sold pirated movies back in the early 2000s.
Red Hat, SUSE, Canonical, and myself, among many, "sell" GPL software that are used by others as part of their proprietary products.
Each of those, and the customers, are required to make the source they received (not their own proprietary source), including changes, available to those that receive the software (thus indirectly back to the public and the authors). AGPL is similar but includes the case of providing sources for products used in Internet services.
Permissibly licensed software does not require you to provide sources or the changes to it.
Copyleft licenses require the source, including changes, to be provided.
The original purpose of copyleft is roughly the "right to repair". When you buy a product: you receive the necessary source, build instructions, and tools in some cases, to update and replace the software you received. The use of copyleft software passes those requirements on through proprietary users.
Note that one can still build a proprietary product as long as all the copyleft pass-through software can be modified and replaced on the product.
There are all kinds of nighmarishly horrible ways to sell GPL software. e.g. inscrutable build procedures, undocumented build dependencies, crippleware, indispensable non-open-source tools). A significant part of the reason why I avoid GPL software like the plague is all of the horrible schemes owners of GPL software have come up with in order to monetize their supposedly "free" software.
If it helps any, at two companies I was responsible for ensuring that all FOSS was made available to customers in an easily re-buildable, re-imageable way alongside our proprietary binaries.
> A significant part of the reason why I avoid GPL software like the plague is all of the horrible schemes owners of GPL software have come up with in order to monetize their supposedly "free" software.
These schemes wouldn't be necessary if industry would start paying for the stuff they use. And I wouldn't call it a "scheme" when the deal is clear - you either pay by giving money to the author or by giving source code to your users. Choice is yours, and since you are developing proprietary program you shouldn't really have any qualms with the first approach, unless you are a leech.
I just don't really want to support the sale of any software. I see it as directly harmful compared to subsidizing the development of modifiable, free software, which is the true allure of computation.
I'm not completely familiar with the details and haven't done it myself, but what's stopping you from selling non-restrictive licenses to a GPL project? Can't you just etch out an exception for each customer?
I'm not sure what you are suggesting. My tool is classic shareware: You download a free version, and if you like it you pay to remove feature restrictions. Companies pay per device the tool runs on. I'm not sure how I would implement this when distributing a GPL licensed program? Customers could just remove the license checking code and distribute it for free?
jjcob is taking a permissively licensed project and selling a closed-source product derived from it. That's why they are arguing against copyleft in this thread.
right, so jjcob is saying that it is better for others write permissively licensed software so that jjcob can use it for their restrictively licensed software. That's obviously good for jjcob, but I'm not sure how that benefits anybody else.
I'm building a tool that makes working with the big open source project easier for some users. It makes the open source project accessible for more users.
In theory, it would be nice if my tool was open source / free to use. But nobody makes a tool like mine for free, because it is a lot of work and not essential for using the open source software. The kind of people who need it are not programmers and extremely unlikely to come together to fund an open source version of my tool.
My contribution is that I make the big open source project accessible to more people. (Aside from contributing back to the project, which is in my own interest, because contributing fixes is easier than maintaining a fork. A lot of contributions to this open source project come from proprietary software vendors like myself.)
(Sorry for the generic description, I'm trying not to dox myself on HN...)
> He memorized the textile factory machinery designs as an apprentice to a pioneer in the British industry before migrating to the U.S. at the age of 21.
Doesn't sound like he stole anything? At most it's some infringement of intellectual properties. (Wikipedia says: 'He learned of the American interest in developing similar machines, and he was also aware of British law against exporting the designs.' But again, no theft.)
There is another side to it. Instead of using open solutions, big business then copies the open solution to a degree that is permissable or a copyright vioaltion deniable. It can easily stomach the development costs. And in that case would outcompete you as well, as they offer service level agreements and build a consumer product around you solutions.
On the contrary, you would penalize small business, because they cannot stomach the increased developing costs. Large businesses wouldn't even notice the change.
Sure, for software exclusive businesses there were cases were a too permissive licence was counter productive for economic success.
Instead of using open solutions, big business then copies the open solution
i highly doubt that. sure technically they can copy it, but they could not guarantee 100% compatibility. yes, the model would work, but at least they would compete on somewhat equal footing.
what happened to redis, elastic search, mongodb etc was way worse. people switched to amazon and others because they got the original and did not have to risk issues when switching. further more, they also didn't have to risk lock in. (or at least had the impression of not risking lock-in)
in other words offering a functional copy vs taking the original (without paying for it) are two very different value propositions.
but amazon could have made the same offers using GPL software, or even AGPL. yes, it would have cost them more, but not as much more as developing their own. so the fact that they could afford to develop their own clone is not an argument for the original developer to just allow them to use their software without restrictions.
"you should just give this to me because i could afford to just clone it and put you out of business." what kind of choice is that?
small businesses can use GPL licensed software just as they can use permissible licensed software. if using GPL software hurts your business model more than spending the money to develop your own well then maybe it's the wrong model.
i'd argue that using GPL software can jumpstart your business faster than developing your own, and once you make enough money you can develop your own like everyone else who doesn't want to use GPL software. permissible licensed software doesn't provide that much of a jump-start. it also makes it easier for your competition.
and most importantly, it all happens at the expense of the original developer. why should they accommodate that?
You can use LGPL licensed-software in your proprietary software, as long as you only link to it dynamically - i.e. you provide a clear separation between your software and the LGPL software.
The Free Software Foundation (FSF), as the author of the GPL and LGPL licenses, maintains that any software running in the same process space qualifies as "linked" under the terms of these licenses. This includes dynamic and static linking, plug-ins, loadable modules, and, in some cases, even interpreted code.
However, some developers who use GPL licenses interpret these requirements differently and permit certain uses that the FSF does not. A notable example is the Linux kernel, which allows proprietary kernel modules under specific conditions.
For proprietary software authors, the best-case scenario is having to convince a court that the license they accepted, by redistributing GPL-licensed software, permits linking without the obligation to disclose their own source code, as otherwise required by the license.
I just want anyone to be able to use my code. It isn’t ideological for me, I don’t care what people do with it. No one is harmed by a permissive license and I explicitly want to minimize the obligations of anyone using my code. I’d probably make most of it public domain if it wasn’t for the fact that some jurisdictions don’t recognize that concept.
I’ve written thousands of lines of copyleft code. I don’t see a practical benefit if you don’t want to obligate users of that code.
Some would argue that if you want that, then you need to license it in a way that protects anyone's freedoms and disallows other ones from stepping on them.
No, it won't. It restricts how you modify and distribute, which isn't a user thing to do, it's more of a developer thing.
And whether it works or not is mostly untested in the legal sense but the idea with copyleft licensing is that big money shouldn't be able to ratfuck a free software embrace/extinguish-style, or some other way. I.e. it is supposed to protect the rights of the users to enjoy free software.
It isn't for me either. A permissive license means it's a matter of time before somebody forks and keeps it from me and other users/developers. Copyleft ensures as many users/developers as possible have access to the code.
Maybe, I do think it really depends on the type of software/situation.
In an argument about how much someone is able to use something, I think you could argue that a copyleft license turns more people away than it would force to participate, so the overall amount of work that is freely available is less than it would have been with a permissive license. (That is, if 10% using a permissive license don't contribute back and 90% do, you end up with more contributions than if 100% of users contribute back to copyleft but you only have 50% as many users.)
I do think it's very situational though - clearly Linux being GPL has been a huge boon in forcing contributions, and I think the killer bit here is that forcing those contributions means that people are able to use their hardware that might otherwise only run proprietary operating systems. That is, it's the forcing open of something that hardware companies would otherwise keep closed.
On the other hand, I don't see any huge reason why a good regex library, for instance, shouldn't be permissive - making a regex library copyleft isn't going to force open some proprietary software, they'll just use some other (maybe worse, maybe equivalent) regex library, and you might get fewer overall contributions and end up with a worse overall available library.
Copyleft of course is of limited use since the FSF slept through the desktop to Internet service transition and now sleeps through the "AI" transition.
GPLv4 should be AGPL+no-AI if it is supposed to have any effect.
Usually when I hear "robbery" it brings to mind someone stealing your phone or wallet at knife-point. Certainly not training some model on some code that involves neither violence nor depriving anyone of anything.
The concept of copyright is the fiction that information - something that can be freely modified, copied, or transmitted - is of the commodity-form, that information should be treated as if it were a single real object that were inherently scarce.
It is a fiction that exists to serve of some of the most hateful, mafia-like firms - your Disneys, UMGs, Getty Images, and the like.
So if the AI interests are powerful enough to deal that whole rotten system a serious blow, then I'm all for them.
So far, lawsuits seem to be heading towards a "AI is not subject to copyright" direction, so a no-AI license doesn't seem practical.
As for the internet service industry: AGPL already covers that, I think. However, I'm not sure if AGPL has ever been tested in court (especially outside of the USA). For the EU there's the EUPL v1.2 which should be AGPL compatible.
I don't think there's much you can do to stop AI companies from stealing your code unless you go out of your way to make every method name, variable, and string template contain at least one slur or support for terrorism. That way, someone will need to explicitly rewrite the code for it to be spat back out, which is more effort than most AI bros are willing to put in.
A clause in a license which states something like "You may not use this work for training an artificial neural network" should suffice.
If the AI developer ignores the term and trains his AI with it anyway, they would clearly be in violation of your licensing terms, and then it would be for a court to decide.
Having the term in your license would be preferential than omitting it, since it's hard for a court to rule that they're not in violation of your license when its written there plain as day. If you don't have the clause then you basically have no defence as it can easily be argued that they've followed your licensing terms, which don't prevent its use for such purpose.
> A clause in a license which states something like "You may not use this work for training an artificial neural network" should suffice.
If you needed a license to train AI then that clause would be redundant. It doesn’t matter what clauses you add to a license if the courts let people train AI without a license.
You don't need a license to train AI, but that's not the point.
The license is applying to using your work - a EULA if you like. If you provide a license which states that it cannot be used for training, then users of your work are required to follow its licensing terms, and not following them is a violation of the licensing terms.
If you've made your work available to use, without an explicit clause specifying that it can't be used for training AI, then it is a valid use of your work that doesn't violate its licensing terms.
EULAs themselves are a grey area where their clauses are not necessarily enforcible, but it's ultimately for a court to decide, and it would definitely be better to have it as a clause in your license.
If the user ticks a box to declare they've read the terms and agree, it's hard for them to argue otherwise in court.
A sufficiently advanced LLM should be able to read the term and comprehend that it can't learn from the work and should ignore anything from it. As AI improves it will be more difficult for the providers to argue that term violations were accidental because their AIs would specifically need to be instructed to ignore such clauses - and the developers who provide such instructions would have a much weaker defence in a court.
> Lastly, if you place restrictions on the use of the software then it’s no longer open source.
This particular point is the topic of the post, to be fair. A project with a copyleft license can still be "open source", even if it's not FOSS.
> AI companies don’t need to be users of the software to train on its source code.
I don't think this is true (or at least it's moot given the word "users"); they still need to legally obtain a copy of the source code, which is offered by the creator alongside a license. You can't just download a project from github and disregard the license that comes with it.
I don't really care about "open source" (per OSI definition). I'm specifically proposing a new kind of license which does place a restriction on use - that of using it to train AI - and I'm suggesting that such license apply not only to programs, but to the source code itself, with a prominent notice in each source file - and to non computer programs also - basically any creative works which carry a prominent notice that they're not to be used for AI training.
Disclaimer: IANAL
In any claim you would argue that you made it clear that your code should not be used for such purpose; that the AI developer is made aware of your intention for it to not be used for such purpose, that the AI developer was negligent, and so forth.
There are frameworks for dealing with cases where no contract exists, such as unjust enrichment[1], which (according to the citation), is examined as follows:
Was the defendant enriched?
Was the enrichment at the expense of the claimant?
Was the enrichment unjust?
Does the defendant have a defense?
What remedies are available to the claimant?
The first point would basically be extremely difficult to argue against w.r.t the big AI providers. They're making billions on the backs of other people's creative works.
The third point it can be argued that what AI developers are doing is almost criminally unjust - they are performing mass copyright violations by training their AI on works for which they have no rights to copy, and performing no creative acts themselves, besides specific neural network designs which are essentially useless without training data. Moreover, if you've explicitly given notice that your copyrighted works are not intended to be used for AI training, then it would be easier to argue that the AI company's enrichment is unjust because they've intentionally ignored it.
The fourth point is precisely the reason you want a disclaimer stating that it should not be used for training. An AI developer could argue in their defense, that the work is publicly available and has no restrictions on use of training. However, that defense goes out of the window if you have made it clear that it should not be used for such purpose, and the developer (or AI itself) can be reasonably expected to be aware of this notice. They would need another defense besides "we didn't know".
So your claims in a court would largely come down to the second point: proving that their enrichment has come at your expense. This might be difficult for an individual, though could be proven if your code is something quite unique and that an AI is basically regurgitating it, and that this results in a loss for yourself. More likely, a claim against these AI companies would be a class action suit on behalf of many claimants where they could reasonably demonstrate that the work produced by AI is not original, is infringing on their own creative works, and that the defendant has no defense because they intentionally ignored the disclaimer that the creative works were not to be used for such purpose.
W.r.t the final point, there are two potential avenues for remedy if a claim were successful: One is that the claimants are financially compensated in proportion to damages - the other is that the AI developers are forced to stop using works which prominently display a "NO AI TRAINING" disclaimer, and any successful claim would set a precedent so that AI companies would necessarily need to be more considerate about the works they use for training.
If you went to court with such a claim, which is more likely to result in success: The case where you didn't put a notice to forbid using in AI training, or the case where you clearly put notices everywhere that it should not be used for AI training?
Ideally it shouldn't be necessary to have such a disclaimer, as using whole copyrighted works for training is not "fair use." However, the pressure is currently on courts to permit the use of copyrighted works for AI for several reasons.
But rather than waiting for courts to decide whether AI training is "fair use", why not be proactive and just start asserting that your creative works are not intended for AI slop? Even if courts rule that copyrighted works not carrying any disclaimer against AI training are free to use for training, this doesn't necessarily imply that all copyrighted works are free to use for training if they explicitly state otherwise.
“Fair use” is a notion related to copyright. The authors in Europe who won in cases of free software licensing violations did it on the grounds of breach of contract law.
Right, seems I can prevent AI training by marking a work here in Germany (via UrgH § 44b Abs. 3). That would be perhaps something to add to (a variant of) the GPL.
Doesn't prevent training in the US and use of the trained AI and material it produces here, I guess.
Then I have to start a civil law-suit and prove that the produced work is similar enough to the the original work. That was the step that failed in the Hellwig (Linux) vs. VMWare lawsuit.
Also the whole AI is “fair use” battle is not finished. AI companies like to pretend like it is but big part of “fair use” is that it doesn’t compete with the initial work… it’s hard to make a strong case there.
In either case "Fair use" is a thing very limited to USA and does not apply outside it. There may be similar things in other jurisdictions, but it cannot be assumed the same reasoning and rules apply.
It is pretty much impossible to square the FSF's mission or methods with limiting how people use GPLed software in relation to AI. The FSF should be one of the organisations carrying the banner that people are free to use GPL-ed software as they like.
GPLv4 should be AGPL+release the source code for your deployment/provisioning infrastructure, there is clear market demand for a freedom preserving license that defends against AWS, how hard could it be to craft something up that would be indigestible to them?
> If you make the functionality of the Program or a modified version available to third parties as a service, you must make the Service Source Code available via network download to everyone at no charge, under the terms of this License. ...
> “Service Source Code” means the Corresponding Source for the Program or the modified version, and the Corresponding Source for all programs that you use to make the Program or modified version available as a service, including, without limitation, management software, user interfaces, application program interfaces, automation software, monitoring software, backup software, storage software and hosting software, all such that a user could run an instance of the service using the Service Source Code you make available.
Not really a free software license, at least typically not considered one though.
The SSPL is free in spirit, but it doesn't fit the OSI definition of "Open Source" of the FSF's definition of "Free software" because it places restrictions on purpose of use.
IMO we should have more license like this. Who cares whether you've got the approval and branding of some other foundation?
The SSPL does not place restrictions on purpose of use unless you also believe the AGPL does. You can diff the two licenses and see that the part which some people consider to be a purpose of use restriction is identical.
The OSI declared it nonfree, because the OSI is a consortium of large companies - cloud service providers and the like.
The FSF decided not to issue any opinion.
Debian decided not to issue any opinion, but because it was only affecting a few packages and they have good GPL counterparts, switched to the GPL counterparts anyway.
There is no evidence the SSPL is nonfree. That is big tech disinformation. Please don't spread it. What the SSPL actually is is an extreme point in the WTFPL-to-AGPL spectrum - it's more AGPL than AGPL itself. It's still on that spectrum, not a completely different spectrum like a proprietary license.
It is actually terrifying how successful the cloud provider cartel has been at spreading FUD about SSPL. Denied OSI certification and packages getting dropped from repos?
Debian has an unusually strict policy that will absolutely err on the side of not including packages (famously not packaging things you need, like video drivers) and valkey is simply better than redis anyway (more features, better performance) and is drop-in compatible, so it was an easy choice for them.
My favourite licence has always been the MPLv2. It's basically the best of both worlds, it's copyleft for your code while keeping the ability to statically link (crucial for Rust, Go and other languages that do that by default).
That doesn't work for languages which require all the code at once just for compiling, like C++ (if any of the code involves templates), template Haskell, or Rust (at least, not without heroic efforts).
You could make sure the LGPL code only works through a C-API like interface, then you can sub-divide it off more easily, but that can be quite a bit of work.
Well, no, as someone else already stated forcing people to ship object files is a PITA and it's really not feasible with languages that do codegen, templates, ... because you inevitably end up injecting copylefted code inside of your files.
I like copyleft as a way to protect my code from freeloaders, but I don't think that imposing the same view on other people is fair TBH. I want people to be able to use my code without being forced to "convert" to my creed first
The very name "permissive" licenses is problematic, because it steers the discussion towards the point of view of software middlemen, not of users.
From the point of view of a user, a copylefted software allows the user to do more things: the license grants them permission to see and change the source code of the software, no matter how they got it. With so-called permissive software, the user may be stripped of this permission by third parties.
>Third, Glen Weyl-style economic arguments have convinced me that, in the presence of superlinear returns to scale, the optimal policy is actually NOT Rothbard/Mises-style strict property rights. Rather, the optimal policy does involve some nonzero amount of more actively pushing projects to be more open than they otherwise would be.
(read the rest of his reasoning for this philosophical shift).
Interesting arguments in favour of technology transfer/diffusion that I really needed. However, I wonder if letting greater economies of scale form would accelerate progress overall when those economies of scale are applied to research.
Fundamentally, I think the best companies have a symbiotic (or at least respectful) relationship with their customers and not an adversarial one.
Tech companies that I use and (more importantly) I pay for and I think there is a "respectful" relationship: Google (via Android), Microsoft (via office365), OVH (via servers), Proton (via email). And there are others, but tried to select known ones.
Do I like all that they do? Definitely not. But the services for which I ended up paying are reasonable. I would not like to build my one OS for a mobile, install the 10s applications required in an enterprise or manage my own data center/mail server.
Of course there are some that I find adversarial, but that is more a personal feeling - like Apple which asked X times more in price for minor (in my view) better quality for their devices.
It might help you to consider that pricing in direct proportion to bill of materials means the base model must increase in cost to balance the lost revenue from the higher spec devices. That is, people paying "over the odds" for the larger ssd and so forth subsidise the base model.
There are so many actors out there that collect data (ex: GSM companies, utility companies, CCTV, states, etc.) but nowhere near as much discussions as mobile phones.
Regarding Google, I’m consistently disappointed by the ads they suggests to me. I’d prefer relevant recommendations, but what I get is irrelevant. My favorite example: couple of years ago they kept showing me ads for a household appliance like 6-10 months after I already bought one, even if I had lots of emails regarding the buying process in gmail. I can't think "oh, they are evil masterminds" if they fail at their most basic stuff.
So they can record how many times per day I unlock the phone or open an app or the location if the GPS is on, good luck using that for something useful.
I do have limits to what I think they should collect in bulk (ex: video or audio), but I did not see allegations that they do that so far, and I think sometimes it's healthy anyhow to behave "as if they do it", just in case.
I've taken Rob Landrey to heart on this subject. His argument for use of permissive licences to avoid the power struggle that crops up around bigger copyleft projects.
I primarily care about individual users and developers of my stuff, not "the society" and as such make it easy for them to use my code in any project, even closed source commercial one. The exception is 3D models where I don't anticipate commercial uses and if there are I want them to talk to me and maybe pay me some money, so sharealike it is.
> we are not able to opt out of diffusion of progress. People move between companies and between countries and take their ideas and talents with them.
Never forget that Industry itself came to America by way of direct theft. https://en.wikipedia.org/wiki/Samuel_Slater
=====
If you care about your users, you should always prefer some copyleft license.
Prefer permissive licenses only if you want to subsidize big business. Hopefully you have a plan for when they decide to take it closed-source and outcompete you. You probably don't, though, since that's not mentioned in the aggressive "use permissive licenses" marketing campaign they've been running for the last decade or two.
Permissive licenses can fit in a niche for software that has already been implemented many times. But the downsides remain, and why aren't you just using the existing code directly? The main occasion on which I will consider it is when everybody else is doing it wrong (and content to do so), and I as a user keep getting annoyed at hitting those limitations.
The second-worst reason to use permissive licensing is arguably "I don't understand how to dynamic linking".
> If you care about your users, you should always prefer some copyleft license.
There is no possible way for users to be hurt by a permissive license. Even if some company were to make a closed source fork, who cares? The original is still there, still just as free as it ever was.
> Prefer permissive licenses only if you want to subsidize big business. Hopefully you have a plan for when they decide to take it closed-source and outcompete you.
I prefer permissive licenses because I'm making a gift to the commons. It strikes me as hypocritical to say "I'm giving this to you, but only if you things the way I approve of". Thus, I make it completely free for all users, whatever they want to do with it. You may not see it that way, which is fine. But I'm sick of the false idea of "you're just subsidizing big business" being promulgated. That isn't true.
And why would I care if someone "outcompetes" me? My gift to society is still there to be used if people want, or not if they don't want. It doesn't diminish or harm my efforts in any way.
> Even if some company were to make a closed source fork, who cares?
They can add sufficiently popular functionality to said closed source fork and make the open source original a) obsolete and b) incompatible with the combined ecosystem, and thus deprive the users of a feasible free option.
If the closed fork functionality is superior enough to make the original de facto obsolete then the users have already collectively decided that the tradeoff is worth it.
And if the original can't compete it means the additional functionality was only going to exist because the financial model of the closed fork could pay for it.
> And if the original can't compete it means the additional functionality was only going to exist because the financial model of the closed fork could pay for it.
This completely disregards the fact that the "financial model of the closed fork" explicitly chose to build upon the permissively licensed original.
If the company chooses to build upon free software, they should be obligated to give back to the community from which they leech. Otherwise, they should just build their own thing from scratch with all the money they've hoarded, and keep it closed.
> If the closed fork functionality is superior enough to make the original de facto obsolete then the users have already collectively decided that the tradeoff is worth it.
Users cannot be trusted to make their own decisions about these kinds of things for the same reason that corporations cannot be trusted to be environmental stewards and children cannot be trusted to select a dinner menu or file taxes.
Just because users don't make the same choices you do, doesn't mean those choices are invalid.
users can only choose what is available. the license greatly affects the options that a user can choose from.
That's not the only reason it would become the defacto standard, and it's naive to claim so.
https://en.m.wikipedia.org/wiki/Embrace,_extend,_and_extingu...
Users can get a bundled version that slowly breaks compatibility forcing vendors to align with the closed source version.
at the expense of the original developer whose work was taken advantage of without giving anything back or paying it forward.
You know what would be even better? If the new functionality was open.
The original can't compete because the original author used a permissive license. Do you want to make yourself not compete?
> There is no possible way for users to be hurt by a permissive license. Even if some company were to make a closed source fork, who cares? The original is still there, still just as free as it ever was.
The users of the closed source fork are harmed. They are unable to use the software to the full extent (e.g. inspect, modify, fix bugs). This is what copyleft intends to protect.
Closed source reduces the software into an appliance and the user into a consumer.
The users of the closed fork always had the ability (if not duty) to inspect the program's licence before using it.
I have a different take on permission license. I offer up what little code I produce under the assumption that anyone can use it for whatever, just give me credit. Why? Because it might result in an interesting collaboration on a project I like. If someone uses my existing code to accomplish something neat and sees who I am, they can reach out to me for more, which I'm open to since it would likely result in some growth for me as an amateur. I'm not particularly interested in making money since I recognize that I am nowhere near that level, yet, but creating a future opportunity seems like a good idea since I enjoy coding.
> But I'm sick of the false idea of "you're just subsidizing big business" being promulgated. That isn't true.
All the permissively licensed code that came with my mac disagrees. This is even more true of iphone software: you can technically look at some of the code that comes with it, but you can't modify it.
> you can technically look at some of the code that comes with it, but you can't modify it.
This is equally true of the GPLv2. The attempt to close this loophole - Tivoisation as Stallman called it - only won him a lot of scorn from Linux land and a refusal to adopt the GPLv3!
Fair; I suppose that concern is technically beyond copyleft.
So it sucks to have iPhones?
Yeah, the idea here seems to be, iPhones would be open hackable devices if only more people would use copyleft licences, but in reality Apple would still lockdown their gadgets and just write the software themself if no permissive licence is avaiable.
> in reality Apple would still lockdown their gadgets and just write the software themself if no permissive licence [were] available
This is a common opinion, but seems uninformed to me. The free code helped them get started. They are Goliath now, but the benefitted from not having to do the R&D earlier.
I don’t know if they could have done the R&D. I definitely imagine it would not have been as good as the product they got from using BSD.
Anyway, I believe this kind of thinking is a shallow dismissal of the value that companies receive from starting from Libre software.
"Anyway, I believe this kind of thinking is a shallow dismissal of the value that companies receive from starting from Libre software."
It is a dismissal of the idea, that using a different licence will magically transform the whole industry.
If the code is open, the profit orientated companies don't need to do R&D from scratch. They just take the copyleft source code and make it closed, either by simply not publishing it, or if they care about whistleblowers by rewriting it. Which is easy if you have sources of something working. Even more so today with LLMs.
That is why I am a fan of permissive licences. It is a working compromise for companies who would never consider copyleft code in critical parts of their product.
Compromise implies both parties give something. Can you explain what that is on the company’s side?
Compromise for me is, the concept of open source gets more widewspread. There is no must in giving back. And if you try to force it with copyleft .. you won't get far. I believe history is on my side here.
Android is based on copyleft code and it is far from an open hackable system. On some devices it possible to install an own build but it requires skills and patience so not an option even for most IT professionals. But for most devices there are locked loaders and closed drivers/firmware. I don’t see a copyleft license helping if a business is not willing to participate in open-source.
> but in reality Apple would still lockdown their gadgets and just write the software themself if no permissive licence is avaiable.
That would be preferable to me.
I think if copyleft were more common fewer people would tolerate iPhones.
That would require a far larger fraction of the general population to give a shit about any software license than currently.
If you want to run modified or your own code, yeah, it does.
As a reminder, the GPL (or any other Free Software license, by definition - it's "Freedom 0" in the list) places absolutely no restrictions on users. Its provisions only kick in for those who want to go beyond merely using the software.
> or any other Free Software license
That's not quite true of AGPL (and for this very reason some companies, e.g. in the database provider space, use it for free versions of otherwise paid products)
On the contrary, that's just FUD they spread.
Section 13 of the AGPL (the part that's not in the GPL) explicitly begins with "[notwithstanding,] if you modify the program [and allow users to access it over the network]". Merely using the upstream version is no burden.
> Never forget that Industry itself came to America by way of direct theft.
Two other examples come to mind: John Rolf arrived at the Jamestown settlement in 1612, having purchased tobacco seeds presumed to have originated in Spain (where their export was supposedly punishable by death, though I haven’t found sources confirming that).
William de la Barre adapted the rollers used in Minneapolis wheat mills from a European design. I have seen a few sources claim this was industrial espionage wherein Barre obtained employment in the mill and sketched its design out to copy it for his employers in Minneapolis, though Wikipedia makes no mention of this.
https://en.wikipedia.org/wiki/William_de_la_Barre
The American publishing industry also got its start by refusing to recognize British copyrights; this didn’t change until 1891.
This is to say nothing of the more explicit, above-board technology transfers involved in bringing European experts in to develop the American industrial plant. Similar things are currently going on in Saudi Arabia and China, and to a lesser extent in South America and Africa.
And then you have Operation Paper Clip and the looting of the German industrial plant after World War II.
> Prefer permissive licenses only if you want to subsidize big business.
Another reason to prefer permissive licenses is if you want anyone to use your project.
Most coding out there in the world is done in the service of a commercial enterprise and my experience is that companies won't use copy left projects. Companies are generally happy to use projects with permissive licenses or pay for SaaS or commercial licenses, but anything with the GPL is a non-starter.
There are a few, well known exceptions (Linux, GCC), but I can't think of any exceptions from projects started in the last 20 years.
I don't necessarily think this view is entirely rational or irrational, but I’m confident this attitude exists; companies using GPL projects are pretty rare. Anyone found differently?
Permissive licenses enable small or individual developers (like myself) to sell software based on them.
With copyleft licenses, developers always have to sell stuff adjacent to the software (support, individual development, software-as-a-service).
With permissive licenses, developers can just sell a license to the software itself. Much easier, much more direct.
It's the reason why there is much more tooling around permissive licensed projects vs. projects with a GPL style license.
Permissive licenses are win-win: Proprietory developers win because they get to sell licenses, users win because they have more tools they can use.
There are alternatives. We can dual license software under a strong copyleft and a proprietary license, so that users (developers) who don't want their derived software to be copyleft can pay the proprietary license fee. IMO this is the approach that everyone should be taking - it rewards the developer rather than them subsidizing your derived works with their free time.
Obviously this has constraints. The developer must have full rights or ownership to redistribute the work under other licenses. This can be obtained from contributors via agreement, where every developer retains ownership of code they write, but grants permission to relicense it for the benefit the project. Ideally, contributors would also be rewarded for their time via bounty systems, contracts or direct employment.
It's possible to make non-derived works which utilize copyleft software. The license's conditions apply only to the derived works, and not directly to your code unless it unquestionably a derived work such as a fork. If your application is written in a way that it can function without dependency on a copyleft software, but can be augmented using copyleft software using a plugin system, then you're able to distribute your software under any license you chose, and only the plugin which uses the copyleft software would be required to be licensed under a compatible license. The "aggregate" would need to be distributed under the copyleft license - but there are also ways around that - simply distribute your software without the dependency on the copyleft work, and have it optionally download the copyleft plugins separately so that they're not part of the "aggregate".
Permissive licensing won't sustain itself unless the developers who depend on it start donating to the maintainers of the permissively licensed works they depend on. In some cases this happens, but mostly it's just the case that you can make money using software other people have given away and continue to maintain for free. What happens if they decide to stop maintaining it? Is your product still viable if you have to take over the task, or employ someone new to maintain it? Would it be viable if you, or your users, had to purchase a proprietary license fee for an alternative dependency?
You need to read the GPL licenses more carefully. Your code gets infected by GPL when you link. The principle difference between GPL 2 and GPL 3 being that the definition of "link" is much more expansive GPL 3. So separately downloaded packages don't really cut it unless the packages in question are LGPL-licensd (GPL with a link exception).
GPL 2.0 and GPL 3.0 (to a much more significant extent) disallow use of plugins. Or well, the do allow plugins, but at the cost of infecting the entirety of your own codebase with GPL 3.0 obligations.
And that's my principle objection to GPL. Linking to GPL code infects my code, requiring me to also license my code via a GPL license.
Permissive license stand a better chance of getting funding because they allow commercial users to use your software, and commercial users are more likely to give sustaining grants to projects they depend on than mere-ordinary users.
The GPL doesn't "infect" any code. You can write code which clearly calls GPL licensed code and license it under a MIT license.
The GPL comes into play when you distribute. If you distribute your MIT licensed code alongside the GPL licensed code, then the aggregate must necessarily be GPL licensed. This doesn't mean your code suddenly becomes GPL code and cannot continue to be licensed under MIT when not distributed as an aggregate. It is perfectly reasonable for someone to take the code, strip out the GPL licensed parts and replace them with MIT licensed alternatives, and license the whole thing as MIT.
Obviously for proprietary software this implies the aggregate must have its source code available in a GPL-compatible form.
But the point here is that you're separating the GPL work from your own code by means of a plugin system. If a plugin depends on GPL code, then the plugin itself must be distributed as GPL - but this does not infect the host program which loads the plugin.
That is, unless it can be reasonably argued that the host program is a derived work of the GPL licensed work. If the program basically does not function without dependency on the GPL licensed work then this might be a reasonable argument - but if the program runs fine, and is only augmented by a GPL licensed plugin, then it is not a derived work.
> GPL 2.0 and GPL 3.0 (to a much more significant extent) disallow use of plugins. Or well, the do allow plugins, but at the cost of infecting the entirety of your own codebase with GPL 3.0 obligations.
> And that's my principle objection to GPL. Linking to GPL code infects my code, requiring me to also license my code via a GPL license.
No. You release a host application with no dependence on GPL code.
The end user loads a plugin that is (or uses) GPL code. The end user doesn't distribute anything.
In very practical terms, what is the enforcement here? Who is gonna sue the end user?
> In very practical terms, what is the enforcement here? Who is gonna sue the end user?
There's nothing to enforce because no license violation has occurred.
The end user is allowed to use GPL licensed code alongside proprietary software. This is permitted by Freedom 0.
Your logic seems to assume that there two disjoint groups of people:
1. Developers who work on the open source project
2. Developers who work on proprietary software based on the open source project
In reality these groups overlap. Developers contribute to the open source project, and they also work on their own software based on the open source project. If you look at the git logs of successful permissive open source projects, you'll see that a lot of the development time comes from developers who work for proprietary software companies.
That's true for some projects, but there are others who have a sole maintainer, receiving no other contributions or donations, who are expected to continue maintaining the work for free.
My advice is basically to not be that person.
Users lose because the tools with the most functionality, the applications with the most interoperability, the software with the most mindshare,... will be proprietary software.
That does not follow. You might as well make it "source available" and sell proprietary licenses.
I think you misunderstood. My software is not permissive. My software is closed source, but based on a big open source project with a permissive license. It's a tool that helps users of the big open source solution.
I can sell licenses to my tool because it is closed source.
If the big open source project was licensed under a copyleft license, I would have to make my tool open source as well. Then I would have a hard time selling licenses.
I couldn't have spent the last 10 years building a tool for the open source software, and users of the open source software would have had fewer tools available.
Your personal financial incentives < humanity. I would rather you had a harder time selling licenses if it meant the ecosystem around whatever big project you're talking about was compelled to contribute to the commons in the same way that the big project does.
Proprietary software is the dead or soon to die branches of the tree of human development, may spawn some useful insights that will ultimately be applied elsewhere, and may be profitable and useful in the short term, but the branch dies, I don't want to subsidize dead branches, and I don't want to write software that enables people to make more of them, even if some of those branches might be nice to hang on for a while.
You can sell GPL software. You can even sell GPL software under proprietary license as long as you are the copyright owner (dual-licensing).
Really, when you take time to understand real restrictions, as opposed to repeated myths, the only argument in favor of permissive license is that they help big players use free labor.
You don’t even need to be the copyright owner to sell GPL software, but it’s unlikely many would pay for it.
> You don't even need to be the copyright owner to sell GPL software
You do, if sell means also allowing another vendor to use your software as a part of their proprietary software.
GPL alone, if you're not the copyright owner, would not allow you to sell software in this way. But if you are the copyright owner, you can bypass GPL and sell a proprietary license to the vendor saying "you can use this software in your proprietary product if you pay me $XYZ".
For anyone who thinks I'm talking out of my ass, consider Qt.
If you don't own the Qt copyright, you cannot sell Qt to a company who wants to use it to build a proprietary product - since Qt is GPL licensed, using it in a proprietary product would not be permitted.
But since The Qt Company owns the copyright, they can sell their own software to companies that want to build a proprietary product with it - they just sell it under a proprietary license, specifically tailored for this. The fact that it was previously released under GPL is irrelevant, because the copyright holder can license it under as many licenses as they want.
If you don't own the copyright to it, you can only """sell""" GPL software in a manner in which local DVD shops sold pirated movies back in the early 2000s.
Red Hat, SUSE, Canonical, and myself, among many, "sell" GPL software that are used by others as part of their proprietary products.
Each of those, and the customers, are required to make the source they received (not their own proprietary source), including changes, available to those that receive the software (thus indirectly back to the public and the authors). AGPL is similar but includes the case of providing sources for products used in Internet services.
how would that be any different from selling someone elses permissibly licensed software?
as a user i have the choice between yours and the original in both cases.
Permissibly licensed software does not require you to provide sources or the changes to it.
Copyleft licenses require the source, including changes, to be provided.
The original purpose of copyleft is roughly the "right to repair". When you buy a product: you receive the necessary source, build instructions, and tools in some cases, to update and replace the software you received. The use of copyleft software passes those requirements on through proprietary users.
Note that one can still build a proprietary product as long as all the copyleft pass-through software can be modified and replaced on the product.
There are all kinds of nighmarishly horrible ways to sell GPL software. e.g. inscrutable build procedures, undocumented build dependencies, crippleware, indispensable non-open-source tools). A significant part of the reason why I avoid GPL software like the plague is all of the horrible schemes owners of GPL software have come up with in order to monetize their supposedly "free" software.
If it helps any, at two companies I was responsible for ensuring that all FOSS was made available to customers in an easily re-buildable, re-imageable way alongside our proprietary binaries.
> A significant part of the reason why I avoid GPL software like the plague is all of the horrible schemes owners of GPL software have come up with in order to monetize their supposedly "free" software.
These schemes wouldn't be necessary if industry would start paying for the stuff they use. And I wouldn't call it a "scheme" when the deal is clear - you either pay by giving money to the author or by giving source code to your users. Choice is yours, and since you are developing proprietary program you shouldn't really have any qualms with the first approach, unless you are a leech.
Why do you believe you can't sell copyleft software?
I just don't really want to support the sale of any software. I see it as directly harmful compared to subsidizing the development of modifiable, free software, which is the true allure of computation.
I'm not completely familiar with the details and haven't done it myself, but what's stopping you from selling non-restrictive licenses to a GPL project? Can't you just etch out an exception for each customer?
I'm not sure what you are suggesting. My tool is classic shareware: You download a free version, and if you like it you pay to remove feature restrictions. Companies pay per device the tool runs on. I'm not sure how I would implement this when distributing a GPL licensed program? Customers could just remove the license checking code and distribute it for free?
I don't understand, a shareware software is not permissively licensed.
jjcob is taking a permissively licensed project and selling a closed-source product derived from it. That's why they are arguing against copyleft in this thread.
right, so jjcob is saying that it is better for others write permissively licensed software so that jjcob can use it for their restrictively licensed software. That's obviously good for jjcob, but I'm not sure how that benefits anybody else.
I'm building a tool that makes working with the big open source project easier for some users. It makes the open source project accessible for more users.
In theory, it would be nice if my tool was open source / free to use. But nobody makes a tool like mine for free, because it is a lot of work and not essential for using the open source software. The kind of people who need it are not programmers and extremely unlikely to come together to fund an open source version of my tool.
My contribution is that I make the big open source project accessible to more people. (Aside from contributing back to the project, which is in my own interest, because contributing fixes is easier than maintaining a fork. A lot of contributions to this open source project come from proprietary software vendors like myself.)
(Sorry for the generic description, I'm trying not to dox myself on HN...)
> Never forget that Industry itself came to America by way of direct theft. https://en.wikipedia.org/wiki/Samuel_Slater
Wikipedia says:
> He memorized the textile factory machinery designs as an apprentice to a pioneer in the British industry before migrating to the U.S. at the age of 21.
Doesn't sound like he stole anything? At most it's some infringement of intellectual properties. (Wikipedia says: 'He learned of the American interest in developing similar machines, and he was also aware of British law against exporting the designs.' But again, no theft.)
There is another side to it. Instead of using open solutions, big business then copies the open solution to a degree that is permissable or a copyright vioaltion deniable. It can easily stomach the development costs. And in that case would outcompete you as well, as they offer service level agreements and build a consumer product around you solutions.
On the contrary, you would penalize small business, because they cannot stomach the increased developing costs. Large businesses wouldn't even notice the change.
Sure, for software exclusive businesses there were cases were a too permissive licence was counter productive for economic success.
Instead of using open solutions, big business then copies the open solution
i highly doubt that. sure technically they can copy it, but they could not guarantee 100% compatibility. yes, the model would work, but at least they would compete on somewhat equal footing.
what happened to redis, elastic search, mongodb etc was way worse. people switched to amazon and others because they got the original and did not have to risk issues when switching. further more, they also didn't have to risk lock in. (or at least had the impression of not risking lock-in)
in other words offering a functional copy vs taking the original (without paying for it) are two very different value propositions.
but amazon could have made the same offers using GPL software, or even AGPL. yes, it would have cost them more, but not as much more as developing their own. so the fact that they could afford to develop their own clone is not an argument for the original developer to just allow them to use their software without restrictions.
"you should just give this to me because i could afford to just clone it and put you out of business." what kind of choice is that?
small businesses can use GPL licensed software just as they can use permissible licensed software. if using GPL software hurts your business model more than spending the money to develop your own well then maybe it's the wrong model.
i'd argue that using GPL software can jumpstart your business faster than developing your own, and once you make enough money you can develop your own like everyone else who doesn't want to use GPL software. permissible licensed software doesn't provide that much of a jump-start. it also makes it easier for your competition.
and most importantly, it all happens at the expense of the original developer. why should they accommodate that?
Do you have any examples?
> The second-worst reason to use permissive licensing is arguably "I don't understand how to dynamic linking".
??! Not sure what that has to do with ANYTHING.
You can use LGPL licensed-software in your proprietary software, as long as you only link to it dynamically - i.e. you provide a clear separation between your software and the LGPL software.
The Free Software Foundation (FSF), as the author of the GPL and LGPL licenses, maintains that any software running in the same process space qualifies as "linked" under the terms of these licenses. This includes dynamic and static linking, plug-ins, loadable modules, and, in some cases, even interpreted code.
However, some developers who use GPL licenses interpret these requirements differently and permit certain uses that the FSF does not. A notable example is the Linux kernel, which allows proprietary kernel modules under specific conditions.
For proprietary software authors, the best-case scenario is having to convince a court that the license they accepted, by redistributing GPL-licensed software, permits linking without the obligation to disclose their own source code, as otherwise required by the license.
Sure. But GPL does not permit that. You can also link to LGPL statically.
I just want anyone to be able to use my code. It isn’t ideological for me, I don’t care what people do with it. No one is harmed by a permissive license and I explicitly want to minimize the obligations of anyone using my code. I’d probably make most of it public domain if it wasn’t for the fact that some jurisdictions don’t recognize that concept.
I’ve written thousands of lines of copyleft code. I don’t see a practical benefit if you don’t want to obligate users of that code.
It’s usually a question of “who are your users?”
As long as you consider developers of derived software to be your users — permissive makes most sense.
But if you consider end-users of software it’s definitely copyleft.
"I just want anyone to be able to use my code."
Some would argue that if you want that, then you need to license it in a way that protects anyone's freedoms and disallows other ones from stepping on them.
Why would anyone serious argue that?
This is ridiculous. He said he just wants anyone to be able to use his code. Licensing it copyleft will just restrict the user base.
No, it won't. It restricts how you modify and distribute, which isn't a user thing to do, it's more of a developer thing.
And whether it works or not is mostly untested in the legal sense but the idea with copyleft licensing is that big money shouldn't be able to ratfuck a free software embrace/extinguish-style, or some other way. I.e. it is supposed to protect the rights of the users to enjoy free software.
It isn't for me either. A permissive license means it's a matter of time before somebody forks and keeps it from me and other users/developers. Copyleft ensures as many users/developers as possible have access to the code.
Maybe, I do think it really depends on the type of software/situation.
In an argument about how much someone is able to use something, I think you could argue that a copyleft license turns more people away than it would force to participate, so the overall amount of work that is freely available is less than it would have been with a permissive license. (That is, if 10% using a permissive license don't contribute back and 90% do, you end up with more contributions than if 100% of users contribute back to copyleft but you only have 50% as many users.)
I do think it's very situational though - clearly Linux being GPL has been a huge boon in forcing contributions, and I think the killer bit here is that forcing those contributions means that people are able to use their hardware that might otherwise only run proprietary operating systems. That is, it's the forcing open of something that hardware companies would otherwise keep closed.
On the other hand, I don't see any huge reason why a good regex library, for instance, shouldn't be permissive - making a regex library copyleft isn't going to force open some proprietary software, they'll just use some other (maybe worse, maybe equivalent) regex library, and you might get fewer overall contributions and end up with a worse overall available library.
Copyleft of course is of limited use since the FSF slept through the desktop to Internet service transition and now sleeps through the "AI" transition.
GPLv4 should be AGPL+no-AI if it is supposed to have any effect.
It is copyright as a whole that is sleeping through the "AI" transition/robbery.
Usually when I hear "robbery" it brings to mind someone stealing your phone or wallet at knife-point. Certainly not training some model on some code that involves neither violence nor depriving anyone of anything.
The concept of copyright is the fiction that information - something that can be freely modified, copied, or transmitted - is of the commodity-form, that information should be treated as if it were a single real object that were inherently scarce.
It is a fiction that exists to serve of some of the most hateful, mafia-like firms - your Disneys, UMGs, Getty Images, and the like.
So if the AI interests are powerful enough to deal that whole rotten system a serious blow, then I'm all for them.
So far, lawsuits seem to be heading towards a "AI is not subject to copyright" direction, so a no-AI license doesn't seem practical.
As for the internet service industry: AGPL already covers that, I think. However, I'm not sure if AGPL has ever been tested in court (especially outside of the USA). For the EU there's the EUPL v1.2 which should be AGPL compatible.
I don't think there's much you can do to stop AI companies from stealing your code unless you go out of your way to make every method name, variable, and string template contain at least one slur or support for terrorism. That way, someone will need to explicitly rewrite the code for it to be spat back out, which is more effort than most AI bros are willing to put in.
A clause in a license which states something like "You may not use this work for training an artificial neural network" should suffice.
If the AI developer ignores the term and trains his AI with it anyway, they would clearly be in violation of your licensing terms, and then it would be for a court to decide.
Having the term in your license would be preferential than omitting it, since it's hard for a court to rule that they're not in violation of your license when its written there plain as day. If you don't have the clause then you basically have no defence as it can easily be argued that they've followed your licensing terms, which don't prevent its use for such purpose.
> A clause in a license which states something like "You may not use this work for training an artificial neural network" should suffice.
If you needed a license to train AI then that clause would be redundant. It doesn’t matter what clauses you add to a license if the courts let people train AI without a license.
You don't need a license to train AI, but that's not the point.
The license is applying to using your work - a EULA if you like. If you provide a license which states that it cannot be used for training, then users of your work are required to follow its licensing terms, and not following them is a violation of the licensing terms.
If you've made your work available to use, without an explicit clause specifying that it can't be used for training AI, then it is a valid use of your work that doesn't violate its licensing terms.
EULAs themselves are a grey area where their clauses are not necessarily enforcible, but it's ultimately for a court to decide, and it would definitely be better to have it as a clause in your license.
If the user ticks a box to declare they've read the terms and agree, it's hard for them to argue otherwise in court.
A sufficiently advanced LLM should be able to read the term and comprehend that it can't learn from the work and should ignore anything from it. As AI improves it will be more difficult for the providers to argue that term violations were accidental because their AIs would specifically need to be instructed to ignore such clauses - and the developers who provide such instructions would have a much weaker defence in a court.
AI companies don’t need to be users of the software to train on its source code.
And, again, if they don’t need a license then the license is irrelevant.
Lastly, if you place restrictions on the use of the software then it’s no longer open source. You certainly can’t do it with the GPL.
> Lastly, if you place restrictions on the use of the software then it’s no longer open source.
This particular point is the topic of the post, to be fair. A project with a copyleft license can still be "open source", even if it's not FOSS.
> AI companies don’t need to be users of the software to train on its source code.
I don't think this is true (or at least it's moot given the word "users"); they still need to legally obtain a copy of the source code, which is offered by the creator alongside a license. You can't just download a project from github and disregard the license that comes with it.
I don't really care about "open source" (per OSI definition). I'm specifically proposing a new kind of license which does place a restriction on use - that of using it to train AI - and I'm suggesting that such license apply not only to programs, but to the source code itself, with a prominent notice in each source file - and to non computer programs also - basically any creative works which carry a prominent notice that they're not to be used for AI training.
Disclaimer: IANAL
In any claim you would argue that you made it clear that your code should not be used for such purpose; that the AI developer is made aware of your intention for it to not be used for such purpose, that the AI developer was negligent, and so forth.
There are frameworks for dealing with cases where no contract exists, such as unjust enrichment[1], which (according to the citation), is examined as follows:
The first point would basically be extremely difficult to argue against w.r.t the big AI providers. They're making billions on the backs of other people's creative works.The third point it can be argued that what AI developers are doing is almost criminally unjust - they are performing mass copyright violations by training their AI on works for which they have no rights to copy, and performing no creative acts themselves, besides specific neural network designs which are essentially useless without training data. Moreover, if you've explicitly given notice that your copyrighted works are not intended to be used for AI training, then it would be easier to argue that the AI company's enrichment is unjust because they've intentionally ignored it.
The fourth point is precisely the reason you want a disclaimer stating that it should not be used for training. An AI developer could argue in their defense, that the work is publicly available and has no restrictions on use of training. However, that defense goes out of the window if you have made it clear that it should not be used for such purpose, and the developer (or AI itself) can be reasonably expected to be aware of this notice. They would need another defense besides "we didn't know".
So your claims in a court would largely come down to the second point: proving that their enrichment has come at your expense. This might be difficult for an individual, though could be proven if your code is something quite unique and that an AI is basically regurgitating it, and that this results in a loss for yourself. More likely, a claim against these AI companies would be a class action suit on behalf of many claimants where they could reasonably demonstrate that the work produced by AI is not original, is infringing on their own creative works, and that the defendant has no defense because they intentionally ignored the disclaimer that the creative works were not to be used for such purpose.
W.r.t the final point, there are two potential avenues for remedy if a claim were successful: One is that the claimants are financially compensated in proportion to damages - the other is that the AI developers are forced to stop using works which prominently display a "NO AI TRAINING" disclaimer, and any successful claim would set a precedent so that AI companies would necessarily need to be more considerate about the works they use for training.
If you went to court with such a claim, which is more likely to result in success: The case where you didn't put a notice to forbid using in AI training, or the case where you clearly put notices everywhere that it should not be used for AI training?
Ideally it shouldn't be necessary to have such a disclaimer, as using whole copyrighted works for training is not "fair use." However, the pressure is currently on courts to permit the use of copyrighted works for AI for several reasons.
But rather than waiting for courts to decide whether AI training is "fair use", why not be proactive and just start asserting that your creative works are not intended for AI slop? Even if courts rule that copyrighted works not carrying any disclaimer against AI training are free to use for training, this doesn't necessarily imply that all copyrighted works are free to use for training if they explicitly state otherwise.
[1]:https://en.wikipedia.org/wiki/Restitution_and_unjust_enrichm...
Since AI use is based on "fair-use" how can you add a no-AI clause? Only protection I see is to not publish the code.
“Fair use” is a notion related to copyright. The authors in Europe who won in cases of free software licensing violations did it on the grounds of breach of contract law.
Right, seems I can prevent AI training by marking a work here in Germany (via UrgH § 44b Abs. 3). That would be perhaps something to add to (a variant of) the GPL.
Doesn't prevent training in the US and use of the trained AI and material it produces here, I guess.
Then I have to start a civil law-suit and prove that the produced work is similar enough to the the original work. That was the step that failed in the Hellwig (Linux) vs. VMWare lawsuit.
Also the whole AI is “fair use” battle is not finished. AI companies like to pretend like it is but big part of “fair use” is that it doesn’t compete with the initial work… it’s hard to make a strong case there.
In either case "Fair use" is a thing very limited to USA and does not apply outside it. There may be similar things in other jurisdictions, but it cannot be assumed the same reasoning and rules apply.
It is pretty much impossible to square the FSF's mission or methods with limiting how people use GPLed software in relation to AI. The FSF should be one of the organisations carrying the banner that people are free to use GPL-ed software as they like.
GPLv4 should be AGPL+release the source code for your deployment/provisioning infrastructure, there is clear market demand for a freedom preserving license that defends against AWS, how hard could it be to craft something up that would be indigestible to them?
> ...release the source code for your deployment/provisioning infrastructure...
SSPL works a bit similarly: https://www.mongodb.com/legal/licensing/server-side-public-l...
> If you make the functionality of the Program or a modified version available to third parties as a service, you must make the Service Source Code available via network download to everyone at no charge, under the terms of this License. ...
> “Service Source Code” means the Corresponding Source for the Program or the modified version, and the Corresponding Source for all programs that you use to make the Program or modified version available as a service, including, without limitation, management software, user interfaces, application program interfaces, automation software, monitoring software, backup software, storage software and hosting software, all such that a user could run an instance of the service using the Service Source Code you make available.
Not really a free software license, at least typically not considered one though.
The SSPL is free in spirit, but it doesn't fit the OSI definition of "Open Source" of the FSF's definition of "Free software" because it places restrictions on purpose of use.
IMO we should have more license like this. Who cares whether you've got the approval and branding of some other foundation?
The SSPL does not place restrictions on purpose of use unless you also believe the AGPL does. You can diff the two licenses and see that the part which some people consider to be a purpose of use restriction is identical.
The OSI declared it nonfree, because the OSI is a consortium of large companies - cloud service providers and the like.
The FSF decided not to issue any opinion.
Debian decided not to issue any opinion, but because it was only affecting a few packages and they have good GPL counterparts, switched to the GPL counterparts anyway.
There is no evidence the SSPL is nonfree. That is big tech disinformation. Please don't spread it. What the SSPL actually is is an extreme point in the WTFPL-to-AGPL spectrum - it's more AGPL than AGPL itself. It's still on that spectrum, not a completely different spectrum like a proprietary license.
It is actually terrifying how successful the cloud provider cartel has been at spreading FUD about SSPL. Denied OSI certification and packages getting dropped from repos?
Debian has an unusually strict policy that will absolutely err on the side of not including packages (famously not packaging things you need, like video drivers) and valkey is simply better than redis anyway (more features, better performance) and is drop-in compatible, so it was an easy choice for them.
Very cool, I did not know that the SSPL was like that!
And I'm surprised both by the fake grassroots campaign to attack sspl: https://ssplisbad.com
And by the OSI's interpretation of it as not open source: https://opensource.org/blog/the-sspl-is-not-an-open-source-l....
>g to their most recent 10-Q Barbie sales have been flat YoY so I guess they felt like they had to do something.
No thanks. Seriously please no, I do not want to see it, that stuff has negative value.
My favourite licence has always been the MPLv2. It's basically the best of both worlds, it's copyleft for your code while keeping the ability to statically link (crucial for Rust, Go and other languages that do that by default).
You can link LGPL statically as well, though it puts more obligations (shipping object files) than MPL.
I am always amazed how many myths were created around FOSS licensing.
That doesn't work for languages which require all the code at once just for compiling, like C++ (if any of the code involves templates), template Haskell, or Rust (at least, not without heroic efforts).
You could make sure the LGPL code only works through a C-API like interface, then you can sub-divide it off more easily, but that can be quite a bit of work.
So, LGPL?
Well, no, as someone else already stated forcing people to ship object files is a PITA and it's really not feasible with languages that do codegen, templates, ... because you inevitably end up injecting copylefted code inside of your files.
I like copyleft as a way to protect my code from freeloaders, but I don't think that imposing the same view on other people is fair TBH. I want people to be able to use my code without being forced to "convert" to my creed first
You can obfuscate your code.
The very name "permissive" licenses is problematic, because it steers the discussion towards the point of view of software middlemen, not of users.
From the point of view of a user, a copylefted software allows the user to do more things: the license grants them permission to see and change the source code of the software, no matter how they got it. With so-called permissive software, the user may be stripped of this permission by third parties.
Another name for "permissive license" is "cuck license". Hopefully I don't have to explain the analogy.
>Third, Glen Weyl-style economic arguments have convinced me that, in the presence of superlinear returns to scale, the optimal policy is actually NOT Rothbard/Mises-style strict property rights. Rather, the optimal policy does involve some nonzero amount of more actively pushing projects to be more open than they otherwise would be.
(read the rest of his reasoning for this philosophical shift).
Interesting arguments in favour of technology transfer/diffusion that I really needed. However, I wonder if letting greater economies of scale form would accelerate progress overall when those economies of scale are applied to research.
Fundamentally, I think the best companies have a symbiotic (or at least respectful) relationship with their customers and not an adversarial one.
> I think the best companies have a symbiotic (or at least respectful) relationship with their customers and not an adversarial one.
What are examples of large companies like this?
Tech companies that I use and (more importantly) I pay for and I think there is a "respectful" relationship: Google (via Android), Microsoft (via office365), OVH (via servers), Proton (via email). And there are others, but tried to select known ones.
Do I like all that they do? Definitely not. But the services for which I ended up paying are reasonable. I would not like to build my one OS for a mobile, install the 10s applications required in an enterprise or manage my own data center/mail server.
Of course there are some that I find adversarial, but that is more a personal feeling - like Apple which asked X times more in price for minor (in my view) better quality for their devices.
It might help you to consider that pricing in direct proportion to bill of materials means the base model must increase in cost to balance the lost revenue from the higher spec devices. That is, people paying "over the odds" for the larger ssd and so forth subsidise the base model.
I'm a dev (and also a client) at OVH, just wanted to thank you for using our product <3 <3 <3
> I think there is a "respectful" relationship: Google (via Android)
https://news.ycombinator.com/item?id=26639261
There are so many actors out there that collect data (ex: GSM companies, utility companies, CCTV, states, etc.) but nowhere near as much discussions as mobile phones.
Regarding Google, I’m consistently disappointed by the ads they suggests to me. I’d prefer relevant recommendations, but what I get is irrelevant. My favorite example: couple of years ago they kept showing me ads for a household appliance like 6-10 months after I already bought one, even if I had lots of emails regarding the buying process in gmail. I can't think "oh, they are evil masterminds" if they fail at their most basic stuff.
So they can record how many times per day I unlock the phone or open an app or the location if the GPS is on, good luck using that for something useful.
I do have limits to what I think they should collect in bulk (ex: video or audio), but I did not see allegations that they do that so far, and I think sometimes it's healthy anyhow to behave "as if they do it", just in case.
I've taken Rob Landrey to heart on this subject. His argument for use of permissive licences to avoid the power struggle that crops up around bigger copyleft projects.
I might not always agree with Vitalik but he's always a joy to read.
This is the Ethereum guy. Was something forked and made proprietary that he needs to be open source?
This comment makes no sense. Can you give a single precedent for him to have behaved that way?
I think it depends on what you do/did. Things I wrote is using the ISC
https://en.wikipedia.org/wiki/ISC_license
But anyone and their brother can develop those old utilities I created decades ago and enhanced over the years.
But if you create something very complex and provides a real benefit, then I can agree with this point. You may be better off to using the copyleft.
In anycase, you need to really need to think about the License you choose, not just use one because it is trendy now.
I primarily care about individual users and developers of my stuff, not "the society" and as such make it easy for them to use my code in any project, even closed source commercial one. The exception is 3D models where I don't anticipate commercial uses and if there are I want them to talk to me and maybe pay me some money, so sharealike it is.
Crypto bros rediscovering RMS 50 years later.