- cross-posted to:
- [email protected]
- cross-posted to:
- [email protected]
It’s hard for me how he recognizes all the (social) issues with LLMs and still comes to the conclusion that he isn’t in part responsible for them.
It’s hard for me how he recognizes all the (social) issues with LLMs and still comes to the conclusion that he isn’t in part responsible for them.
I’m constantly baffled by the copyright argument.
Intellectual property isn’t even real property (exclusive power over a thing), it’s a licensed monopoly enforced by the state, which gives the copyright owner power over many many things, present and future, not actually owned by them. Books that are in my house, software that is on your phone.
I don’t want to argue that intellectual property should be abolished, I’m not convinced that that would be beneficial. But I think we can all agree that the currently landscape of intellectual property legislation is utter nonsense, that it’s gotten completely out of hand, and that it mainly encourages concentration and gatekeeping.
So yes, the model producers have blatantly disregarded copyright, and have gotten away with it in ways which are not available to us peasants.
But we’re the copyleft people, we should be pointing at this and shouting that the emperor is finally naked, and that it’s time to curtail all that nonsense legislation once and for all, so that we can all be a little freer from the shackles of corporate monopoly.
And instead, we (the copyleft guys!) are just sitting here clutching our copyright pearls while muttering “but… But… Copyright! It violates the precious copyright!”.
You need to read about why RMS wrote the GPL in the first place. Yes obviously Imaginary Property is imaginary, but we have to exist within the legal system. RMS used the copyright system as a way to legally ensure the freedoms that copyright traditionally restricts, and that is why he calls it “copyleft”.
Once you’ve reformed or repealed copyright law in some imaginary revolutionary future, you can use whatever code or material you want in your work and publish it freely as a derivative work, but until then you have to acknowledge the existence of the legal frameworks that we live in.
If we start filling our own FLOSS-licensed codebases with code of unknown provenance we open ourselves to legal challenges, it makes the entire license unenforceable, and it destroys the share-alike ecosystem that we’ve built over the past 35 years.
To be fair to the original commenter: They did bring up copyright in the context of “what makes this a different kind of tool”, which is an empirical claim and not a normative one.
True, and I have to admit I went off on a tangent, because I hear the other argument, the normative one, all too often.
Unknowingly using potentially copyrighted code from an LLM in a FLOSS project opens us up to being attacked via being sued for infringement, which most small devs can’t afford to fight, especially in the US’s current extremely pro-corporate courts.
It’s putting a target on your back for down the road when it becomes profitable for patent trolls to use AI to try to scan for copyrighted code on public code bases.
Big tech companies could do the same to squash an open source competitor.
We know for a fact that LLMs have been trained on all available open source codebases.
Do we know that there is a relevant amount of closed source code in the training data?
Additionally to the other answer: Not all published source code is open souce, and I’m pretty sure, the scrapers didn’t stop to read the licence.md files on github
Corporate devs who use a non-self hosted LLM for developing their closed source apps are sending that code to the LLM’s servers, where it can be used as further training data (and considering how many of the corporate AI’s have purposefully trained their AI on pirated copyrighted content (like… A lot of them are doing that), it follows that they would train their AI on user prompts as well).
I completely disagree with you. Copyright law is generally a good thing. For example, I made a video recently that went super viral (like, on international news), and I was able to make money off of it because I own the copyright. If copyright weren’t a thing, all of those news orgs could have just run the video without me seeing any benefit from it, even though I created it.
Copyright law is good for most people, and especially good for creators. It can be abused, just like most laws, but overall, copyright law is good. That’s the main reason I don’t use AI to write code. If the AI wrote the code, I don’t own the copyright (in fact, no one does).
Copyleft also very much depends on copyright law. Without copyright law, the GPL would be completely unenforceable. Anyone could take any open source code and do anything with it without giving anything back to the community.
I exactly said that I don’t want the abolition of IP. I quote from my own comment:
What is your problem again?
Copyright law isn’t just a two states exists/doesn’t exist kind of thing.
While it exists it can be good or written like ass in about 50 different dimensions.
I have a problem with these statements:
I do own my copyrighted material. You own a copy. You can’t redistribute it (without my permission), because I own the copyright.
We cannot agree. It’s not gotten out of hand. It needs to be enforced and strengthened, in order to protect creatives from AI companies stealing our IP.
Exactly. That’s why we need to strengthen copyright protection.
That would literally give corporations more power over our creations.
Because it does violate our copyrights. I didn’t release my code as open source so that it could be copied and stolen without regard to the license I specifically chose to protect it.
So you have a right over my copy, and you can decide what I do with it. That’s what I said. You are the sole person to have the right to produce copies. There’s a word for that, it’s a monopoly.
And since your material can, materially, be copied, copyright only exists as enforced by the state. It’s a state enforced private monopoly. And so are patents.
And these didn’t exist until the states introduced them. This is pretty factual.
So you think that at automatic, untraceable term of 90 years if perfectly reasonable?
You think that the ability by large corporations of amassing and gatekeeping (no derivative work without their permission) 100 years of human culture is just fine? The ability to keep under control works that they are not even distributing, which then means that these can’t even be read, listened or watched… This is all fine?
You like how the US patent systems grants a patent without any real examination for basically anything that written in English, no matter how ridiculous?
You love the endless cultural slop milking the same damn IPs for nostalgia?
It pleases you that every artist who records a version of “my favourite things” (or any other jazz standards) ends up giving 90% of the proceeds to the heirs of the same two dead white dudes? Or how sampling basically destroys the revenue potential of a new song because the author of the sample is going to eat up the work of new artists?
You think your viral video wouldn’t have been adequately protected with, say, a 30 year term? You think 5 years in jail for a private copy of a blue ray is good and proportional?
Apparently it often doesn’t. Courts have already ruled on this. When the billionaires do it, it’s fair use.
To the extent that I can prevent you from distributing copies of it to other people, yes. I can’t stop you from doing anything else with it. I can’t stop you from burning it, eating it, fucking it, shooting it, or using it as a door stop. I only have the power to prevent you from redistributing it.
Sure. You say that like it’s a bad thing, but that’s why inventors can do what they do and survive. I’m an inventor. If I didn’t have my patent, Google could just come and take my idea, build their own version, and undercut me. Isn’t that a worse monopoly? As it stands, if Google wants to sell my product, they have to buy my patent. That’s a great system that protects small inventors like me.
Correct. The states invented them because of a problem. They fix the problem. Copyrights and patents give power to the little guy to prevent big companies from stealing their intellectual property. I don’t really care what you do with my code because you probably wouldn’t know what to do with it, but if Google starts selling my code and my patented product, you can bet I’m going to sue them.
I’m not sure what “untraceable” in this context means, but yes, I should be able to profit off of my works for my entire life. Arguably, my heirs should be able to benefit from my works too. If I don’t have a fortune to give them, because I’m not a capitalist pig, at least they can benefit from my hard work. And yes, copyright being automatic is necessary to prevent the system from being abused even more. Imagine if I lost out on the revenue from my work because I couldn’t afford to register it, or a thief broke in and stole my draft before I registered it.
Like I said, the law can be abused. If the alternative is that the creators of these pieces of culture don’t see a dime, then yes, this system is absolutely better.
You’ve clearly never been through the patent process. They went over my patent with a fine toothed comb. I went through three rounds of revision and had to take an entire section of claims out (because it might have been too close to an existing patent, which I thought was ridiculous, but whatever). I spoke with the patent examiner who absolutely had read my entire application and understood it very well.
Just because you don’t like Disney doesn’t mean that copyright laws are broken. Yeah, Disney kinda sucks lately. So what? Come up with your own ideas and make something worth watching, like tons of independent creators have recently. I just finished The Amazing Digital Circus. What an awesome example of creativity. I’m very happy that Disney can’t just copy it.
If you don’t like giving to the creators whose work you are building on, just come up with your own ideas. Tons of artists use their own original content.
No. I would like to own that video for my entire life. I created it. It wouldn’t exist without me. If someone wants to use it, they can license it from me. Has anyone ever gotten 5 years in jail for privately copying a blu ray?
And that’s the problem. Our IP laws are not strong enough.
You have drank so much koolaid that you think you’re not a working class creator, you’re a temporarily embarrassed Taylor Swift.
The accumulation, gatekeeping, and rent extraction is the point of the system, and if you think it’s a system that’s made to protect the little guy, just look around yourself.
But you can stop me. Because copyright law legally protects any DRM you want to place over that, and breaking it is copyright infringement regardless of the final purpose of it. And “noone will know if I don’t distribute it” is not an answer.
No, like I said, I believe it should exist, I just believe it should be reduced significantly.
Good for you. You’ll offer me a beer when you make your first billion. But you won’t.
The IP system is part of the reason why the tech sector is now so consolidated. Your patent would have been a lot more valuable in a more competitive world.
So, intellectual property is not like “normal” property of limited physical goods. We finally agree on something.
My father worked in a factory his entire life. How come I don’t get continuous revenue from his work, but you expect your heirs to?
You want multi-generational profit from work you are completely done with. You actually are a capitalist pig, you’re just a less successful one.
A 90 year automatic copyright means that the overwhelming majority of works produced in the last 100 years are orphan works. Stuff which you can’t use, but you can’t even tell who, if anyone, still has the copyright. The combination of a shorter automatic term with a registration-based extension would solve that.
Yeah, I am not in the US, but I do read US patents on a regular basis. While I can only speak for my field, I see the kind if crap that the USPO passes as if it’s valid, and it’s quite frankly embarrassing. Yes, an existing patent can be a hurdle. But the fact that the invention is totally trivial to a person “trained in the art” or even already published, does not seem to be a problem.
Very often US patents can’t be converted into patents into other jurisdictions, or half the claims gets struck down.
And I have seen the exact same (trivial) invention patented twice by different companies, because the stated purpose of it was slightly different. A fictional difference as in practice both products are exactly interchangeable and do the same work.
And let’s not even get into the whole nonsense of gene patenting.
But no problem here, let’s patent physical laws next.
It’s not just Disney. It’s everything and everyone. Nintendo? Sega? Are Alien sequels being made because they are such great ideas or because it’s profitable to milk the IPs?
It’s just too easy for them to extract value from existing IPs, to be bothered with putting the same money into new stuff.
Lol, just say you have no idea how jazz works, or that you don’t like this kind of degenerate art. You’ll be in good company.
And jazz is just an example of the kind of “remix art” that have been fundamental to cultural development of the 20th century.
I have zero interest in being a billionaire. You make a lot of assumptions.