Skip Navigation

InitialsDiceBearhttps://github.com/dicebear/dicebearhttps://creativecommons.org/publicdomain/zero/1.0/„Initials” (https://github.com/dicebear/dicebear) by „DiceBear”, licensed under „CC0 1.0” (https://creativecommons.org/publicdomain/zero/1.0/)B
Posts
12
Comments
1593
Joined
2 yr. ago

  • someone who paints in the style of Michelangelo or Studio Ghibli does not pay royalties to either of those entities

    I think it's actually not entirely true. There is a thing called "trade dress" (part of the trademark law), which can protect certain design or "look&feel" elements of any products a company produces; the requirements for this are a bit strict, but if the "style" of your painting is sufficiently similar to a Ghibli animation to cause confusion for customers (e.g. someone may reasonably think that the painting is by Studio Ghibli), there is a possibility that it's a trademark violation. But this is also beside the point.

    if someone reads ten thousand books and then writes their own story, they are not committing copyright infringement.

    There is a legal distinction between human learning and LLM training.

    The neural connections in someone's brain formed by reading a book are not considered to be a derivative work, because they are not a "work of authorship" as they are not “fixed in any tangible medium ... from which it can be perceived, reproduced, or otherwise communicated", and they are not “sufficiently permanent or stable to permit it to be perceived, reproduced, or otherwise communicated for a period of more than transitory duration".

    LLM weights meanwhile totally fit the definition of a "work", stored in the medium of a digital file (fixation) and produced by humans through a computational process (human authorship), making it a derivative work of the training material by definition.

    Once again, I agree that this is actually unfair, but it is the consequence of the law as written, because the law itself is unfair.

  • Ah, I see. Yeah, that makes more sense and I don't have an answer for your question then. I suspect it wouldn't really work due to long reproductive cycles and extremely limited total food supply.

  • If a company unlawfully ... reproduced protected works in a way that violates copyright law, then it has broken the law and should be prosecuted.

    Merely training an LLM on copyrighted material without holder's permission (and then distributing the weights or selling access to inference on those weights) is a violation of copyright law if it were to be applied consistently (ignoring the fair use argument, which I'll get back to). That is, if you applied any other computational process in this way, the result would be a derivative work and subject to approval by copyright holder(s). The reason it's "different" this time is that the people violating the law are richer than those who wrote the law in the first place, not because of any legal argument.

    As of right now, there are multiple lawsuits against major LLM developers. In some of those cases, the courts have ruled that training on publicly available data can qualify as fair use.

    If a court ruled that it's "fair use", that actually lends more credence to the idea that LLM weights are a derivative work - "fair use" is a defense for copyright infringement that only makes sense in this case if the new work is an unauthorized derivative of the original.

    Whether it's actually fair use or not is another question. I can see the fair use argument for open-weight non-commercial models, not so much for commercial offerings - commercial LLMs seem to fail all 4 elements of fair use.

    BTW, I'm not even necessarily anti-AI (at least the open-weight, local models). I use a local model in my job almost daily, and also I think it's mostly good that the entirety of FOSS corpus is available for download in a compressed and easily remixable form. I'm just pointing out the hypocrisy of the legal system which applies its already unjust copyright law (and most other laws) only against poor people.

  • And, since we’re all cannibals now, what portion of the hucows is going to be reserved as hucow food?

    I think this is where your farming idea falls apart (or becomes identical to the problem discussed in the video). You can't sustainably farm animals without an external energy source. The reason we farm herbivores and some omnivores is that it's possible to feed them with grass (or whatever processed awfulness passes as grass nowadays). When we do farm carnivores for some reason, we feed them with meat harvested from herbivores.

    A fully cannibal-based farm is simply impossible because even if humans were perfect at turning food into muscles on their body, there wouldn't be anything to (sustainably) extract from the overall farm system, since there wouldn't be any extra energy/muscle mass/whatever going into it. Also, humans are very very not perfect at turning food into muscle mass, and waste most of the energy that came into their system on stuff like breathing and heartbeat and other metabolism. So, if the requirement is that all humans are cannibals, the only thing left to do is to farm non-sustainably, i.e. extract meat from the system as it runs out, feeding hucows with other hucows to keep at least some of them alive, but eventually you will run out of hucows (which is, once again, pretty much what's discussed in the video).

    Alternatively you could claim that "hucows" are not "humans" per se and feed them with non-hucow-based food, in which case it could probably work in a fairly similar manner to modern-day pig farming but on an even larger and more horrifying scale - pigs are quite similar to humans in many ways. Sans all the pandemics that would eventually happen due to cannibalism, of course.

  • I don't think the farmed humans would even be able to consent in this scenario, so it would not be vegan. However to me it also doesn't seem much worse than the world we're living in now 🤷

  • I don't know what to tell you, legally there's a difference between a person learning something and training an LLM.

  • Copyright protects the expression itself, not the ideas, facts, patterns, grammar, writing styles, or knowledge learned from that expression.

    Copyright absolutely does protect "ideas, patterns, grammar, writing styles". It does not cover material facts, but this is beside the point here.

    Nobody claims that someone who read 10,000 copyrighted novels is committing copyright infringement every time they sit down and write a new story.

    Actually, if they take certain elements from other works it often can be considered copyright infringement, but this is also beside the point.

    You don’t get to argue that learning is legal when humans do it and suddenly becomes theft when a machine does it. Either learning from publicly available information is allowed, or it isn’t. The standard cannot magically change because you dislike the technology.

    This is the crux of the issue. The LLM is not a person from a legal perspective, therefore it cannot "learn". What is happening is that a legal person - a company - is consuming a bunch of copyrighted material and transforming it into a bunch of data to be interpreted by a computer program. This makes that data definitionally a derivative work made by the company. Sure, the transformation process is probably "creative" enough for the company to be able to claim copyright on the resulting weights, as long as that company gains explicit or implicit approval from all the copyright holders of the works they used to create the weights. Legally speaking, this is no different from you pirating a bunch of movies and making a compilation of the funniest moments - see how the legal system would react to you doing that.

    Of course, I find the entire concept of modern copyright system to be an awful idea whose entire purpose is impeding the human creativity for the purposes of monetization.

    But if we read the law as it is - the AI companies are absolutely doing copyright infringement by training their models on GPL code and not releasing the weights under a GPL-compatible license.

  • GPL absolutely should cover shit like this. Training an LLM on your code makes it definitionally a derivative work, therefore it must be licensed under GPL too (with limited fair use exceptions which shouldn't apply here). The problem is that the US government is not willing to enforce this at all, because it is owned by the same billionaires as the AI companies.

  • LLMs are absolutely trained on FOSS software, including GPL'd stuff. Accelerating software development is also a large part of how they are making money. I believe training on GPL'd software and then charging for access is copyright infringement, but it doesn't really matter because entities supposed to be enforcing copyright are paid for by the same billionaires who run the AI companies, so literally nothing will happen.

  • As intended by the founding fathers, of course

    1. Try ping mediacenter.local to see if your system DNS resolver can resolve the address
    2. Try curl -v mediacenter.local to see if the web server actually responds with HTTP data on port 80
  • Is there any research on how effective of a prevention technique this is? Jail alone is an ineffective crime prevention for humans, I wonder if it's different for bears.

  • Playing stupid games during an energy crisis to benefit a couple local oligarchs. Whatever potential cybersecurity threat exists from Chinese inverters is outweighed a thousandfold by the US actively extorting the EU by leveraging fossil fuel dependence.

  • A Prius can tow like 700 kg if you put a hitch on it, almost the limit for a B class license.

  • I was like 90% sure it's a joke. It's not

    Thankfully it was published 4 years ago and didn't seem to go anywhere in that time. Good.

  • The art is in how the science is applied.

    No, the art is how you express yourself. The theory is just "the way most people in a culture do it because other people in this culture like it". There is certainly some biological reasoning behind it that is common across cultures, but it's a fairly small part of music theory. Consider a simple question like "why are dissonant chords considered dissonant?" and it can lead you down a path of discovery.

    It isn’t eurocentric to believe harmonious and rhythmic music sounds better than music that isn’t.

    It absolutely is eurocentric to enforce western music theory as some holy book that separates good music from bad music. It's a tool that helps you write music that sounds good to a western ear. But a lot of good music can disagree or break the norms of western music theory.

  • Control over the straight is one of the main reasons they're still geopolitically relevant. I don't see them giving it up for good, they will at least say that the blockade is still on. Given they have a lot of naval drones and missiles which are not 100% countered by US hardware, I don't foresee any insurance companies co-signing on commercial ships crossing the straight.

  • It's gonna be about warming soil releasing even more CO₂ into the atmosphere, isn't it.....

  • Au contraire, a truly "mutualist" society would never fall for the bullshit that is happening.

    If everyone had trust and a good working relationship with their neighbors and solidarity with their working class comrades, the US capitalist class and therefore the government in its current form would quickly cease to exist.

    The problem is specifically that the US capitalists inserted themselves as unavoidable middlemen into every aspect of US life. Notice how you're not getting food from your community kitchen, you're getting food from Uber Eats. You're not usually calling a repairman from your local community, you're calling a company that sends one out to you. You're not getting help from your community militia, you're getting it from an armed wing of the bourgeois government.

    (ironically, this was possible to do through years or "rugged individualism" propaganda, that you are repeating here; other things like car dependency also made things worse)

    This lack of local community is a big part of why there is so little organized protest. It's very difficult to rile up your neighbors to take up arms and meaningfully protest when you barely even know them. This would apply even more in your imagined scenario of everyone becoming individualists who raise their own cows - people wouldn't even have time to protest because they would spend all of it on their own survival, and those who stick their head up are easily arrested and thrown into jail, with noone to protect them.

  • Offroad @lemmy.ml

    It's not sexy but it is (was) mine

  • Inventing Reality @lemmy.ml

    Just Bezos Things

  • ThinkPad @lemmy.ml

    While we're sharing setups...

  • TrailViews @piefed.social

    Desert Sunrise

  • pics @lemmy.world

    Desert Sunrise

  • Photography @lemmy.ml

    Desert Sunrise

  • pics @lemmy.world

    Moonrise at sunset

  • Photography @lemmy.ml

    Moonrise at sunset

  • pics @lemmy.world

    Winter Dawn

  • Photography @lemmy.ml

    Winter Dawn

  • pics @lemmy.world

    Late spring in Georgia

  • Photography @lemmy.ml

    Late spring in Georgia