@thatbrickster @SuperDicq i believe there certainly are enough organizations in this area who try to define "freedom".

only they tend to define it as "free as in free beer" like the whole permissive license crowd. or "freedom until you wrongthink", like FSFE who'd seem to rather have some kind of "ethical license".

FSF and RMS are a clear cut case of exceptionalism because most people (including myself) couldn't come up with something which ONLY protects freedom without additional limitations.

@bonifartius @thatbrickster @SuperDicq abolishing copyright / IP would do this. some of the FSF 'freedoms' are not freedoms at all but positive rights (i.e., obligations on third parties) that only get enforced via copyright law.

@lain@lain.com @bonifartius@qoto.org @thatbrickster@shitposter.world Abolishing copyright does not fix all of the world's software freedom issues. If we abolished copyright proprietary software will not automatically become free software, because we would often still lack the source code to make any meaningful changes.

some of the FSF 'freedoms' are not freedoms at all but positive rights (i.e., obligations on third parties) that only get enforced via copyright law.

This is not technically true either. You are confused with concepts like copyleft, which is enforced through a copyright license.

Copyleft is not part of the free software definition. It is just a way to ensure free software stays free, but it does is not a requirement for software to be considered free.

Outside of copyright there also exists free software. This is for example software released into the public domain (if the source code is available).

@SuperDicq @bonifartius @thatbrickster mandating access to the source code is the point that is an obligation on a third party

@lain @SuperDicq @bonifartius @thatbrickster
>mandating access to the source code
*if the people ask for it.
It's only mandatory when people are asking for it.
And it should be like that for any overly complex engineering systems. It's important for maintenance, having future, or even present people being able to work on them instead of being stuck and creating issues.

@lain @SuperDicq @thatbrickster
> that only get enforced via copyright law.

i'd say that a license can exist without copyright law as it's a private contract. the way GPL is written it is creating an exemption of copyright, because copyright law exists - it would be useless otherwise :)

if these laws wouldn't exist, you still could either do a license as private contract just for using the software (the ususal EULA) or something like GPL allowing for more freedom.

@bonifartius @SuperDicq @thatbrickster

the difference would be that private contract law only binds the two parties that actually signed the contract. if microsoft makes a contract with that i can not give away their software for free, that won't bind you if i give it to you for free anyway. ms might sue me for damages but you can keep giving away the software for free once its out. same for sources.

@bonifartius @SuperDicq @thatbrickster we can see that playing out with the model data of LLMs, which are currently not thought to fall under copyright, but often you have to sign some sort of agreement to get access to them (for llama, for example). but then somebody inevitably 'leaks' it (i.e, puts out a torrent) and everyone else can work with it with no strings attached.

@lain@lain.com @bonifartius@qoto.org @SuperDicq@minidisc.tokyo @thatbrickster@shitposter.world It would be hilarious if someone leaks (read opens) openais models before their 'open source' model release. so much for their fair use argument which should apply to them lol

@lain @SuperDicq @bonifartius @thatbrickster
Who in the right mind would think they don't fall under copyright? If LLM uses some work as input then the result and the models are clearly derivative works. Any transformation of works results in derivative work with LLM being only one of possible means to. "Fair use" and other special cases can apply too but LLM doesn't automatically imply fair use.

@loonycyborg@fediverse.wesnoth.org @lain@lain.com @bonifartius@qoto.org @thatbrickster@shitposter.world

Who in the right mind would think they don't fall under copyright?

This is the product of current "AI" marketing machine hype boom.

They have successfully managed to deceive all of the non-tech people in power that "AI" is something different than just a turing machine which takes input and produces output.

Many people, due to lack of critical thinking skills, unfortunately believe machine learning models have some sort of actual intelligence or sentience comparable to humans.

So they make fair use exceptions for them as if machine learning model output is somehow comparable to a human making an original work while simply being inspired by another work.

@loonycyborg @SuperDicq @bonifartius @thatbrickster this is not about if the creation of the models violated copyright, that might very well be. the question is whether the models themselves fall under copyright, and the current legal opinion is that they do not. this is not 'marketing hype' because AI companies would love it if those models were copyrightable, but they aren't created by humans, so there's no copyright. it's the same case as this one: https://en.wikipedia.org/wiki/Monkey_selfie_copyright_dispute

@lain @SuperDicq @bonifartius @thatbrickster
In this case it's really hard to find someone who can be defined as an author, yes.

@lain @SuperDicq @thatbrickster i'm not a lawer but shouldn't including a term that giving away the software applies the same license be enough?

you get the software. you give it to me for free, without informing me about the license. still, as the license stated distribution _must always_ happen under the same terms the license still applies and i get in trouble, only that i could sue you for the damages the original producer sues me
for, as i received the software under the false assumption it wasn't licensed this way.

i've been spending most my life, living in a lolberts paradise 🎶

@bonifartius @SuperDicq @thatbrickster the license is irrelevant to you as a third party, you never went into any contractual obligations with either me or the second party.

@lain @SuperDicq @thatbrickster it's moot to discuss this as it's a fictional world :) i believe usually one enters a contract by _doing something_ as well, not only signing something. i accept the software you offere to me. thus a contract is entered.

@lain @SuperDicq @bonifartius @thatbrickster
>abolishing copyright / IP would do this.
Everyone would love this. Sadly it's easier to make licenses than to cancel a law made by a conglomerate of sociopaths.

>some of the FSF 'freedoms' are not freedoms at all but positive rights (i.e., obligations on third parties) that only get enforced via copyright law.
That's the same perspective that proprietary copyrighters have. But where are the positive rights when you deprive people from their natural abilities ?

@bonifartius@qoto.org @thatbrickster@shitposter.world

only they tend to define it as "free as in free beer" like the whole permissive license crowd.

That's not true. Permissive software doesn't have to be free as in price.

like FSFE who'd seem to rather have some kind of "ethical license".

Also not true, the FSFE follows the same free software definition that the FSF does.

@SuperDicq @thatbrickster
bit of a rant here:

not in the mood of digging around on the FSFE page, but i think i have seen something along these lines from them.

iirc they support things like this weird EU license which glows like thousand suns. on top they act like amoral assholes in general. FSFE wanted to kick out RMS on made up accusations, they will make up reasons to take away freedoms.

@bonifartius @SuperDicq @thatbrickster Correct.

@bonifartius@qoto.org @thatbrickster@shitposter.world The EUPL is not a glowie license. It is also approved by the FSF.

I am aware that the FSFE has fallen victim to the lies spread against rms and do not want rms to have any significant role within the FSF.

This bad act of diplomacy aside, I still think the FSFE is a good organization that supports software freedom in Europe and does mostly good things.

@thatbrickster @bonifartius @SuperDicq It's hard to not have a monopoly on the definition of something when everyone else so far are just maliciously attempting to insert malware be it consciously or unconsciously.