Discussion

tptacek: This piece was more interesting than I went in expecting it to be. But it hinges I think on a disputable claim, that the GPL's value is based on a network effect, so that the GPL gets dramatically less useful (and thus attractive to newer projects) as the base of mainstream GPL software shrinks. I'm not sure I understand why that would be the case.The AI rewrite of the Linux kernel also seems farfetched. I don't think it really belongs in the title of this post.

ignoramous: chardet is a Python module ... with 170 million downloads ... licensed under the GPL. A different developer ... to include it in the Python standard distribution ... [reimplemented] it. With [Claude] ... in 5 days ... the reimplementation ... [has] better performance. May be LLMs learning from all the FOSS code there is, and out-performing a team of domain experts, is helping us realise the promise of free as in beer.

iamnothere: AI hopium/copium.Why would you even rewrite it rather than finishing out Redox or Genode OS, which are built on better principles? This is silly.

vips7L: I guess the question is: was Claude trained on chardet? Is that really clean room?

gpcz: Good luck finding every contributor to the Linux kernel and getting them to agree to that.

OneDeuxTriSeiGo: lol that's not even their claim. their claim is that AI will rewrite the entire linux kernel from scratch with an MIT license. which. lmao

number6: AI can't hold copyright (as for now) how will this affect licencing?

friedtofu: Gross. Not sure about y'all but seeing an obviously AI-generated image at the top of the article is an instant nope/close tab for me. Not going to flag but these articles should be DOA.

notepad0x90: I've always wondered how many proponents of open source software also support the concept of copyrights in general. On one hand I see lot of people in these communities support the right to do whatever you want with the things you own, but on the other hand copyright licenses.I've always liked the MIT license because it is closer to copyleft than GPL variants. I get having to use a license so that people who use the software are legally protected, and attribution can be nice too, so while I don't agree with the legitimacy of software licensing as a whole, in practical terms the MIT is a good and safe license that doesn't impose lots of restrictions on its users.WTFPL and unlicense are better in my opinion, but lawyers might not like them. if you don't like the idea of lawyers running the world though, they're great. Public domain is the way. Even then, I despise the idea of even acknowledging "Public Domain" as a concept.But back to my original question, are most people using these licenses because they actually believe in their legitimacy? I always assumed it was to facilitate nonsensical copyright laws.From wikipedia:""" Copyleft is the legal technique of granting certain freedoms over copies of copyrighted works with the requirement that the same rights be preserved in derivative works. In this sense, freedoms refers to the use of the work for any purpose, and the ability to modify, copy, share, and redistribute the work, with or without a fee. Licenses which implement copyleft can be used to maintain copyright conditions for works ranging from computer software, to documents, art, and scientific discoveries. Similar approaches have been applied to certain patents. """Ugh..yeah. Then I don't get it OP, I hope copyleft does indeed die. Either you have a commercial agreement or a contract with a person or you don't. The idea of publishing some work (software, book,etc..) and then by default and without any contractual agreement, dictating what a person does with that work that you published publicly is ridiculous. I know it is the law, but that doesn't make it right. What you put out there to the public, you have no right to control. Either close source your software and require a proper contractual agreement to use it, or make it actually free and actually open. It isn't free if you're going to tell me what to do with it later on, that's a deceptively hidden cost, it isn't really free.Wouldn't it be in the true spirit of open source and the Linux kernel if absolutley anyone can do whatever they want with the kernel's source code?The thing that irritates me is that if they want money, let them charge for it. No problem with that. You can charge for your work. But to make it "free" and then demand that the law enforces control over the software after someone has adapted it feels like such a crooked way of doing things. I would prefer to pay for it, so that I too can get some guarantees of the software being viable in return as well.It sounds to me like 1) You don't truly own copyleft software 2) You don't get any warranty, or expectation of viability 3) certainly, there is no expectation of support of any kind 4) It is marketed as 'free' (deceptively)if people make modifications to the software and keep that private later on, let them. it's free, it's theirs now, they can do what they want. If they value it though, it is in their interest to keep the upstream project viable, so sensible users will contribute back. Especially considering how globalized software dev has been, you can't even practically enforce something as weak as copyleft outside of the EU, Canada, Japan, and maybe (long shot) in the US.

spwa4: What do you mean? It's pretty damn clear how copyright and courts work, but let's review:1) when a large copyright conglomerate sues a teenager: 30.000 USD per violation + enforcement of the license by an armed police officer coming to their house.2) when a teenager has copyright violated by a large copyright conglomerate, or any large company: change the rules and force a compromise, resulting in $0 in damages.In fact it has just been made clear that copyright DOES NOT protect against AI training. That's "fair use".But let's pretend the courts are fair. Let's say I take all disney movies and tell AI to make a new one. Will courts suddenly decide Disney owns the copyright on both the model AND the movie? 100%.

dTal: That's funny, because the "network effect" of the GPL was always extremely obvious to me as part of the point. The idea being that the more that GPL software forms a cohesive ecosystem, the less financially viable it becomes to operate outside of that ecosystem. It's viral, right? And software builds on software, we're about 15 layers deep at this point. The hope is that eventually, so much software is GPL that you effectively have two choices when writing something new: 1) join the GPL borg, or 2) boil the ocean reimplementing the entire stack from scratch.

tptacek: Why does the GPL network effect mean that I, as a library developer, am incentivized to choose GPL over MIT?

akerl_: > The hope is that eventually, so much software is GPL that you effectively have two choices when writing somethingI’m really glad then that it didn’t work out this way, because I wasn’t really keen on all the individual freedom of joining the borg.

dTal: Which "individual freedom" do you feel the GPL denies you? As far as I can tell, it only prevents you from piggybacking on other people's work, and adding unfair stipulations to the resulting product. It is a very symmetrical, "do-unto-others" type license.

akerl_: In the scenario you’re describing, when I write my own code, I am limited in what license I can pick for that code because of licensing choices other people made.

47282847: > It isn't free if you're going to tell me what to do with it later on, that's a deceptively hidden cost, it isn't really free.You are confusing two different definitions of the word free. And overall you seem to operate from a place of very little understanding of why protection of authorship, copyright, licensing of works etc exists. Please go and research some more. I hope you learn something. You will not be taken seriously by anyone if you continue like this.

delotrag: You seem to misunderstand the GPL.> if people make modifications to the software and keep that private later on, let them.This is perfectly legal under the GPL. What's not legal is redistributing that software you modified but not giving _your_ users the same rights to modification that you yourself got.Nothing in the GPL requires you to release or distribute personal modified versions of GPL software.If by "keep that private later on" you mean "plagiarize GPL code to add to a proprietary program and distribute it," yeah that's not allowed- but unless you're a fan of pilfering the commons for personal profit, this is an unmitigated good feature of the license.As an aside, copyleft is on the exact same legal foundation as the EULAs you seem to respect. It is extremely confusing that you think copyleft is bad but EULAs which provide significantly more restrictions are good.

dTal: Libraries or executables, it makes no difference. You are incentivized to use GPL because you wish to build on top of work that is GPL.Obviously in large part that didn't happen, because of a cultural tendency to use more permissive licence variants (such as AGPL) for libraries, in the pragmatic hope that this would encourage their use even in proprietary programs, and therefore incentivize back-contributions from a wider audience. But this indeed halts the "virality" of the GPL, and so one is once again forced to conclude - incredibly - that Stallman was not radical enough...

up2isomorphism: Even GPL never seems to be great idea to me. But at least it considers a core problem of any knowledge sharing seriously. Can you just steal other people’s work? I would rather trust GPL folks than this guy.

dTal: But you're not actually restricted from doing anything, are you? What is it exactly you want to do that other people's choice of GPL prevents? Steal their work and sell it? Oh how unfair!

akerl_: Say more about how licensing my code as MIT would be unfair.

dTal: My pleasure!We are talking about a hypothetical universe in which nearly all software is GPL, such that it is almost impossible to write useful software without building upon other GPL code. In such a universe, licensing "your" code as MIT would indeed be unfair, because you would be taking the work of others, illegally stripping the label, and making it available to profitable interests to use without compensation to the original developers against their express wishes - said compensation merely being the extremely reasonable request to share back, as you were shared to.You still haven't really explained why you're so keen on doing that sort of thing.

notepad0x90: > This is perfectly legal under the GPL. What's not legal is redistributing that software you modified but not giving _your_ users the same rights to modification that you yourself got.That's still controlling what users do, and worse, it's what they do with modifications, not even your own original work!! and without even agreeing to a contract of such terms.It's plagarism if the activity was done under some system that required attribution, or if they misrepresented the change's provenance. neither is the case.I agree with you on EULA's, they've been rendered useless in some cases and jurisdictions. But at least EULA's have a software prompt that enters their users into an agreement. There is a reason they require you to scroll all the way down and read it before agreeing, it is because how inherently weak they are.I don't necessarily think EULA's are good, but at least as a user, i actually agreed to them. No one is trying to force me to agree to an EULA simply because EULA.txt exists somewhere in the directory tree of the software. And even if they did and had legal grounds, I still would disagree with any terms I didn't explicitly agree to, or any terms of agreement that are lopsided to the advantage of one side.copyleft is an attempt at corporate greedy litigious manipulative behavior, except the greedy party doesn't actually want money, they just want control.

delotrag: It is explicitly _not_ controlling what _users_ do. It is a restriction on _distributors_ of software. As a user of the software, you may use it for any purpose, modify it to your heart's desire, and even redistribute it. You just can't redistribute it and then refuse to give downstream users the same rights.That is not meaningfully a restriction unless you're trying to unjustly profit off the work of others. The copyright holder doesn't exercise control over users here.

akerl_: > licensing "your" code as MIT would indeed be unfair, because you would be taking the work of others, illegally stripping the label, and making it available to profitable interests to use without compensation to the original developers against their express wishesI'm not sure why there are quotes around "your".If I write code and license it MIT, but it includes code that has a different non-GPL license (lets say Apache), my code is MIT-licensed, and the included code is still Apache-licensed.I haven't illegally (or legally) stripped any licenses, or changed how it's available to others. I've picked a license for code I wrote, and the developers of code I took a dependency on picked a license for their code. People who want to use my code have to consider the license of my code and also the dependencies I used.The GPL is largely unique in its desire to control what license I can pick for my own code.I'm keen on picking my own license for my own code because I personally don't want to block my code from being used by anybody, commercially or otherwise. I've got no issue with developers who do want to prevent closed-source, commercial, or any other kind of downstream usage. And I'm happy to comply with the licenses of code that I leverage as part of my code. I do take issue with developers who want to impose their licensing preferences on my code.

OneDeuxTriSeiGo: lmao good luck with using an AI to try to rewrite the entire linux kernel. Doubly so good luck doing so without accidentally recycling a bunch of GPL code in the process.

7e: The Linux kernel isn’t that complex. It’s mainly device drivers. Linus wrote the first version himself as a hobby, for crying out loud. AIs will be creating custom kernels within five years, from scratch, as customized as you wish. Probably based on BSD, or other liberally licensed code, for legal reasons.

ctoth: There is no test suite that captures kernel correctness. This is simply wrong. and will not happen.

7e: Well the UNIX certification and POSIX comes close. But man pages might be all an AI needs.

7e: GPL licenses are already only a small minority of the total OSS world. But the GPL will die because AI can write equivalent software easily, and it’s more customized to boot. And for that reason OSS will also wither away; there’s little reason to upstream AI changes. Code is cheap and the FSF is a dinosaur.

notepad0x90: I think I am failing at communicating my view. What I am saying is those rights you speak of are invalid. I could use it,redistribute it, profit from it, backdoor it, or print it and make a door stop from it. It is mine. The original author has no say in it.Perhaps it would help if I mentioned that my objection stems from an overall revulsion at copyright and anti-piracy laws. It is hard for me to object to those and at the same time support the application of those same laws. I either accept the law being used this way by everyone (open source, bigcorp, bigmedia,etc...) or I don't.Set it free, and if it was meant to be yours, it will come back!In my opinion, the proper way to solve this is by requiring users agree to a non-distribution, and/or non-commercial use, prior to being allowed to download the open source software. And that agreement is strictly between the person publising it and the person downloading it. If I obtain a copy of it from someone else, I am not bound by the terms of that agreement. Another approach is to actually charge for the open source software for commercial use, yet allow downloading of the software (with a confirmation prompt for non-commercial use prior to download) free of charge. That way, the publisher has a commercial claim, loss of profit, something under tort law against whoever is using their code and profiting from it.But even then, I don't get in what world a modification, which by definition is new original work that was added to the software, could be a thing the original author have any say over. If publishing software is speech, then that is compelled speech. and you're being coerced into speech, not because you agreed to any terms, but simply because someone put a license term in a file and presumed agreement to those terms, not by the person that obtained a copy from the publisher, but by absolutely anyone who happened to obtain a copy of that software.My problem if it isn't clear, is that those same laws are used to control what people do with their software and devices in many other contexts. What's good for the goose and all..

delotrag: In a world where everyone by default had unlimited rights to use, modify, and redistribute software, sure, copyleft would be unnecessary. But we don't live in such a world. You believe users ought to have unlimited rights to do what they want with their software- copyleft hijacks the copyright system to ensure these rights can't be removed downstream. "What's good for the goose" is exactly the rationale behind copyleft- if large corporations can use it to restrict user rights, the community can use it to protect them.By not supporting copyleft, what you seem to think you're doing is consistently opposing copyright encumberances. But practically speaking, you're just giving up the fight- large corporations can enforce copyright and restrict users, and you don't support fighting back because you believe it would be philisophically inconsistent.My contention here is that you're wrong, in the sense that we share a goal (software freedom) and your strategy will less effectively accomplish that goal than the one you oppose. Opposing copyleft will not end copyright, but it _will_ give all the benefits of copyright to those looking to restrict user freedom.

notepad0x90: We just have different approaches. In my view, I will use the law when I don't have a choice, but in this case there is a choice. I think the copyleft approach is in fact giving up, because as you put it, it is using copyright laws, justifying means by ends.The fact is, there are public domain licenses, and they work well, they only exist because of legalistic reasons. My view is that the world is what we make of it. Oppressing others because you're oppressed isn't right. I make no distinction between "users" and "redistributors" like you're making. this isn't class warfare from my perspective.Think of it differently, with public domain, everyone gets the same access. In my view, if the software is modified and redistributed, I don't care, because as a publisher I never claimed any rights over the software I published to begin with. The license is "do whatever you want with it". to be more prescient, property rights, and the ability to freely share what I own is more important to me than the free accessibility of software. I also believe that good free software shouldn't rely on the contributions of commercial entities. From redhat to Google, I've seen good contribution from them, but they also wield an unfair influence over open source projects. From the kernel to systemd, there are endless complaints about them.copyleft is attempting to entice corporate beneficiaries of open source, whereas I think I'd like to see the opposite of that. Maybe that made sense at the founding of the FSF, but these days the power dynamics are wildly different. I prefer for governments and individuals alike to fund good open source software, and for that software to be truly free. I don't want some corporation supporting the project so that they can wield undue influence over it, and corrupt it to serve their own self interests.There is no shame in asking for funds or monetary support. Or with asking users to pay for the software directly. the open source community is very large, it isn't a small band of devs writing code on their free time anymore.

bloppe: This article ignores the other side of the coin. You can argue that AI makes "clean room" re-implementation so cheap that copyleft is doomed. But then it follows that re-implementing proprietary software as free software would also become "free". So mb AI will kill copyleft, but then it would also kill the need for copyleft in the first place.

OneDeuxTriSeiGo: v0.01 took Linus months of work and even then it barely ran anything. It wasn't POSIX compliant. It couldn't run GNU userspace. It only ran on one arch and had no concept of multi-arch support and limited to no real driver compartmentalization.It wouldn't be until 2 years later that almost a hundred devs joined the Linux project and they'd gradually have achieved compatibility with GNU userspace. Full POSIX compliance wouldn't come until later.And it wouldn't be until another year or two later that they'd finally reach a "production ready" 1.0 by which point hundreds of devs were contributing to the project.You are vastly underestimating the insane amount of work that went into the Linux kernel in the early days to get it to where it is today that it can be developed at such scale and with support for so many intermixed, overlapping, or mutually incompatible features, devices, and platforms. To call that effort (just for core linux, not the drivers) anything less than herculean is frankly an insult to all the people who have dedicated so much time and energy towards that project.