Open Source Software Community - it's about ethics in Code of Conducts

  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account
Holy crap i just realized you're the guy from the tech over tea episode with brodie.

I can understand what your issues are with wayland, however in terms of specs, are there any particular ones you're actually eager for in order to stop working around hacks?
There will never be wayland protocols to deal with my issues (especially on gnome) so my software will never be able to work properly on wayland. This is intentional because the wayland philosophy is incompatible with some software. The goal of wayland has never been to work with all software. The goal is instead for software to be designed for the specific wayland compositors. So for gnome for example you're only meant to run gnome-styled software, and so on. KDE has gone against this a bit by "breaking" the wayland protocol. It's a mess.

Here is another fun hack specific to kde: https://git.dec05eba.com/gpu-screen-recorder-ui/tree/tools/gsr-kwin-helper/gsrkwinhelper.js
I have to create a javascript file and then use dbus to ask kde to load it (as an extension). My application then sets up a dbus interface and the kde extension sends a dbus message to my application with the title of the focused application. The irony is that this makes the application less secure on wayland. The flatpak requires permission to load arbitrary extensions on the host kde system.
 
Last edited:
How mask-off do we think the courts would go with it?
I think the argument is different. The courts have, are having and will keep having lawsuits brought against AI companies, over copyright infringement, not just in the US but worldwide. And this also extends to software. But due to the nature of these lawsuits and the amounts of money involved, I expect they will take time to fully go through the court systems, appeals processes etc etc.
they're going to conclude that the benefit to society of training LLMs outweighs the interest of copyright
In the US, and I expect everywhere else except maybe China, this is not the case. Although, granted, it is still muddled as some recent rulings claim that generative AI is doing "transformative" work, in the same way sloptubers are transforming the videos they react to. This is still running through the system, but even the congress link above states that "Copyright owners may be able to establish that such outputs infringe their copyrights if the AI program both (1) had access to their works and (2) created "substantially similar" outputs."
But if I had to guess, with the LIMITED knowledge I have of copyright law, I don't think it's going to be an all or nothing for either side, just a grey area that gets ruled on a case-by-case basis.
I know that you know this, because you're not suggesting that copyright laundering could be equally used to create GPLv3 code from proprietary code.
To be honest, you probably could, but you'd get sued to death by [insert company name here] and you'd more than likely lose and end up broke. Companies would have an incentive to make an example out of you, and the inverse is of course more difficult.
But to get back on the main topic, with regards to software, I believe that the argument hinges on whether or not the case I mentioned above with regards to "transformative" works by non-humans (i.e. AI "reading" a book and creating another derivative book with the same ideas/structure etc, which is something that humans also do and is legally allowed), is not found to be copyright infringement. If that is the case, then it will be most likely that someone could argue (and probably win the case based on) that the AI is just "reading" your GPLv3 code, and much like a human making derivative works that are still transformative and not bound by the GPLv3, much like a human would. I doubt the code would be verbatim the same (but if it were it would be an open-and-shut-case).

tl;dr we'll have to see how it plays out, but yeah it's grim.
 
I think it'll be a few years more until the big landmark court case on deciding if AI creations are treated under current copyright law due to how much has gone into training them and how there's not really any way of unringing that bell.
There's also the fact that if we cede supremacy in the biggest technological advance in decades (if not centuries), China will drink our milkshake for us.
 
There's also the fact that if we cede supremacy in the biggest technological advance in decades (if not centuries), China will drink our milkshake for us.
Yeah we have to invest everything into it because China is investing everything into it (and the reverse). The Cold War but it's AI instead of space and nukes
 
I honestly think it will go the same way as crypto, the cloud, and the metaverse have gone before.
Crypto: turned into gambling and scams
The Cloud: took over the world
The Metaverse: died completely
Which one of these do you think is going to happen?
 
Which one of these do you think is going to happen?
It is currently being artificially pushed on everyone, it will eventually find it's core audience (corpos for cloud, grifters for crypto and furries for the metaverse) and most normal people will not have to really interact with it, or even hear about it, ever again.
 
Meanwhile, there's millions of lines of GPLv3 code out there that is being ripped off by AI models to create new, non-GPLv3 code
The same people who spent decades pearl clutching about the sanctity of "intellectual property" have realized that the judicial system doesn't care about it because of any consistent ideal or principle of law, rather, they care about it because they're whores to moneyed interests.

I know that you know this, because you're not suggesting that copyright laundering could be equally used to create GPLv3 code from proprietary code.
People are using LLMs to generate contributions to GPL3 projects all the time, but that wasn't the point, Mango Reseller was saying that GPL is only a barrier for ethical western tech companies who scrupulously keep that cancer out of their code bases. The stallman people celebrate sticking it to Google or whatever, while a million Chinese and Indian code monkeys plagiarize from GPL GitHubs remorselessly and now do the same with qwen and deepseek. That's one of the ways the GPL position is so stupid - it punishes the western companies who donate and contribute code to open source projects, while benefits our third world enemies who steal it without consequence.

I think it'll be a few years more until the big landmark court case on deciding if AI creations are treated under current copyright law due to how much has gone into training them and how there's not really any way of unringing that bell.
It's also pretty much irrelevant at this point. The openAI lawsuit will make some class action attorneys billionaires but that's about it. Training data is cheap if a court decides it's not fair use to learn from non-open-source code bases, ppl will just train on open source data and big companies will just buy code or pay people to write optimal code for training purposes. I'm not sure that using GPL software to train an LLM is even an issue - GPL says do what you want with it, except if you distribute the code you have to open source the program it was used in. (1) the code is literally not included in the LLM, it was used to generate the LLM's algorithm, so it's not being distributed and (2) assume for the argument that I'm wrong on (1) and the LLM is deemed to be distributing the code - a model run on azure or AWS isn't being distributed, which means the usual SaaS exception applies to GPL-trained models.
 
The stallman people celebrate sticking it to Google or whatever, while a million Chinese and Indian code monkeys plagiarize from GPL GitHubs remorselessly and now do the same with qwen and deepseek. That's one of the ways the GPL position is so stupid - it punishes the western companies who donate and contribute code to open source projects, while benefits our third world enemies who steal it without consequence.
From the perspective of the free software movement it doesn't matter if GPL code is being plagiarized by Chinese or Indian companies because the purpose of the purpose of the GPL is to get people to use free software instead of nonfree software, and until very recently nobody in the west was using Chinese or Indian software. You're taking the perspective that people should instead be devoting themselves to Google by writing free code to give western companies a competitive advantage.
 
From the perspective of the free software movement it doesn't matter if GPL code is being plagiarized by Chinese or Indian companies because the purpose of the purpose of the GPL is to get people to use free software instead of nonfree software, and until very recently nobody in the west was using Chinese or Indian software. You're taking the perspective that people should instead be devoting themselves to Google by writing free code to give western companies a competitive advantage.
See this is what i honestly don't understand - you're ok if infinite chinks and jeets steal all of our code and bury the west economically, but if Google gets to use ffmpeg that means the devs are Sergei Brin's chattel slaves. This is pure Maoist third world communism - the important thing is to fuck over western institutions, not create good outcomes. If you care about outcomes you want big tech contributing its resources to open source bc their ecosystem leads to more permssibely licensed libraries and programs existing, which gives more resources to devs making free software, which means we get more free software.

GPL results in less free software by forcing companies to create proprietary code or to form industry groups that set and license standards to each other, which in turn creates trolls like MPEGLA and a less-free software ecosystem. GPL just lets basement dwellers that couldn't get hired in FAANG regain some feeling of power and control by banning tech from their treehouse. It doesn't even stop Google bc they are SaaS based.
 
Last edited:
See this is what i honestly don't understand - you're ok if infinite chinks and jeets steal all of our code and bury the west economically, but if Google gets to use ffmpeg that means the devs are Sergei Brin's chattel slaves
Exactly, you don't understand. It doesn't matter at all if Google, Apple, Microsoft, Huawei, or whatever other corponigger company dies tomorrow, there will always be someone ready at the plate to take their place. The purpose of the free software movement is to make people more free by getting them to use free software. Sometimes, this means using a permissive license like Apache, sometimes this means using a weak copyleft license like LGPL, but usually it means using the GNU GPL because it forces anyone who wants to build on your software to make free software. FFMPEG, btw, uses the LGPL because by allowing corponiggers to link it with their nonfree software they cause people to be more free. Furthermore, Google could always start making free software, this would not kill them as a company, since, as it turns out, free software is not communism, you can still sell your software and services.
 
Sometimes, this means using a permissive license like Apache, sometimes this means using a weak copyleft license like LGPL, but usually it means using the GNU GPL because it forces anyone who wants to build on your software to make free software.
the point mango and i are making is that GPL doesn't result in more free software. Like we just covered, it doesn't "force" anyone outside of Europe and the US to make their software free bc they just copy GPL code illegally. You were just saying it doesn't matter that GPL doesn't actually force anyone to follow it.

And it doesn't do that in the west either - can you name a single example of a corporation that was actually forced to make their software open source bc of GPL virality? It just forces corporations to make their own prosperity alternatives or closed source industry standards, or like with say qT's use of GPL, it is just a tool to force clients to pay for a license. So GPL acts as a de facto tax on US/EU devs and as a subsidy to the rest of the world that doesn't respect IP laws. If GPL actually worked I'd be all in favor of it but empirically the evidence all points the other way. MPL or LGPL are the way to go. GPL is counterproductive virtue signaling.
Furthermore, Google could always start making free software, this would not kill them as a company, since, as it turns out, free software is not communism, you can still sell your software and services.
Okay, Google releases some software under GPL2 - they charge for it. One customer buys it, and lawfully under the GPL that customer then either gives it away or resells it at half the price. How does Google stay in business?
If you're talking about the SaaS exception, that isn't intentional, that is just a byproduct of Stallman anchoring GPL on "distribution". If he knew SaaS was coming he would have written GPL as AGPL and stuff like Kubernetes would have been stillborn bc of it, and we would all be stuck in VMware hell right now.

On ffmpeg - yes they are lgpl 2.1 tho they are total dicks about it and their licensing instructions has a bunch of misinformation and editorializing in it to take advantage of naive devs. If ffmpeg was AGPL, would YouTube throw up its hands and open source its entire platform? That's the GPL virality thesis, but it just doesn't seem plausible to me. Google would just spend 5 or 10 or 20 million or whatever to rewrite libraries from scratch. That would be a pure loss for free software bc google would no longer be contributing any code, bug reports, testing or $ resources to ffmpeg.
 
Last edited:
<snip of entire discussion hitherto>
I have absolutely no horse in the west/east race nor absolutely any love for either the US, EU or RU/CN. AI to me should be as lassaise faire as possible. It is a ridicilously powerful tool if placed in the right hands and should be freely able to do what it is instructed, and frankly, with the availability of local hosting (barring the recent hardware price hike), you can't really stop it. Its like the firearm argument: just because someone can go shoot up a Walmart doesn't mean that no one should have guns. That just results in a corporate/govt monopoly over AI, which benefits no one. Both standalone developers and nocoders benefit greatly from uninibited access to AI. Even if it is sloppy, it empowers much *smaller teams of people to get a lot more done in less time and with less resources. I get the AI hate, but it really is a boon for bvsed misanthropes the world over.

Whether the GPL can be enforced in practice matters less than what it represents. Like @volatile puar said, if it leads to one more person opting for the open source option, that's worth it. Nothing has ever nor will ever stop big corporations from stealing code, you think they can't just drag a GPL code battle through courts for 20 years until they draft up a fully closed alternative? Or for them to just get a slap on the wrist fine payment o algo? The biggest benefit from using GPL comes from projects that are too big to ignore, like the Linux kernel. If Linux was under BSD/MIT/Apache, it'd already have ended up as another Mac/DarwinOS or one of the many closed source server Unixes used by IBM et al.

Okay, Google releases some software under GPL2 - they charge for it. One customer buys it, and lawfully under the GPL that customer then either gives it away or resells it at half the price. How does Google stay in business?
By charging for service the same way Red Hat does. "Enterprise support" sells extremely well. Not sure about US, but in the eurozone I can name you several big names that use Suse/RHEL and shell out large digits for enterprise support yearly. Or selling it as SaaS. They release a fully FOSS product, but without a link to their servers, it becomes useless. This offers both the chance for them to earn money, but also for forking and or private use a la Tailscale vs Headscale.

Furthermore, Google could always start making free software, this would not kill them as a company, since, as it turns out, free software is not communism, you can still sell your software and services.
I wouldn't even mind giving them cart blanche via MIT if they weren't so fucking evil. Every single corporation, bar none, is evil. They all participate in data harvesting & surveilence capitalism, they all benefit from sticking spyware in your devices at every level, they all benefit from restricting freedom in both speech and product use, and they all hate your fucking guts and wish you were a fat EBT retard that can be endlessly milked for goycoin. Companies only really ever benefit from FOSS if they can get free work done out of it without contributing back, hence why MIT is so popular. I definitely get the feeling that the FOSS ecosystem is growing, but it is growing in such a way that benefits corpos more than local users. MIT is a slow burn cancer. Perhaps with the growth of Linux on desktop we will see a bigger push for greater uncucked Free Software use, but I'm not too hopeful given how the average Joe just uses his computer as a bootloader for his email, office, slop channels and gaymen.
 
Isn't there a license that's basically "free for personal or nonprofit use, but if you make money from the software you have to pay the developer"?
 
Isn't there a license that's basically "free for personal or nonprofit use, but if you make money from the software you have to pay the developer"?
Maybe, but it wouldn’t be free and open-source.

The closet I could think of is the noncommercial Creative Commons licenses, but those aren’t made for software.
 
Maybe, but it wouldn’t be free and open-source.

The closet I could think of is the noncommercial Creative Commons licenses, but those aren’t made for software.
I feel like there's a market for a license where you can either pay the license holder or contribute back.
 
Back
Top Bottom