Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Meta torrented & seeded 81.7 TB dataset containing copyrighted data

arstechnica.com

Post reply on HN
Print view View original

951–960 of 981 posts

Re: Meta torrented & seeded 81.7 TB dataset containing copyrighted data

#951

Post by Terr_ »

Workaccount2 wrote: post #309
JW_00000 wrote: post #289

I don't understand why it's even a question that Meta trained their LLM on copyrighted material. They say so in their paper! Quoting from their LLaMMa paper [Touvron et al., 2023]: > We include two book corpora in our training dataset: the Gutenberg Project, [...], and the Books3 section of ThePile (Gao et al., 2020), a publicly available dataset for training large language models. Following that reference: > Books3…

There are two different things when it comes to discussing training LLM's on "copyright" protected data, and I almost never see people differentiate. 1.) Training on copyright that is publicly available. You write a poem and publish it online for the world to read. That is your IP, no one else can take it an sell it, but they are free to read and be inspired by it. The legalitly of training on this is in the courts,…

3.) The company making an unauthorized copy of your work and storing it permanently in a giant corporate library of their own making which they refer to over and over.

This is distinct from (1) where the content is streamed or only ephemeral/incidental copies are made.

Re: Meta torrented & seeded 81.7 TB dataset containing copyrighted data

#952

Post by Terr_ »

crazygringo wrote: post #564
Workaccount2 wrote: post #309

Earlier quoted context omitted.

There are two different things when it comes to discussing training LLM's on "copyright" protected data, and I almost never see people differentiate. 1.) Training on copyright that is publicly available. You write a poem and publish it online for the world to read. That is your IP, no one else can take it an sell it, but they are free to read and be inspired by it. The legalitly of training on this is in the courts,…

I'm not sure there's any legal distinction though. Is a book publicly available? No, you have to purchase it. But once you do, you're legally allowed to let your friends and family and so forth read it too. As long as you don't sell copies of it (the "copy" part of "copyright"), or meaningfully take away the ability for the publisher to make money from sales (so you can't post it for the whole world to see on the int…

But they are making unauthorized copies: Their training data set is analogous to private collection of duplicates.

What do you think copyright law(suits) would do if a regular person made copies of every book and movie and song they saw, placing the duplicate media in a room of their house?

Re: Meta torrented & seeded 81.7 TB dataset containing copyrighted data

#953

Post by lrvick »

palata wrote: post #940
lrvick wrote: post #935

Earlier quoted context omitted.

> This is not public domain. MIT, ISC licenses etc offer minimum anti-liabiliy CYA here and are what I typically use. They are a hack to get around the fact you cannot actually public domain code without legal risk. > So... you benefit from copyright laws. I sometimes exploit laws that I feel should not exist at all specifically to target corporations who abuse them the most. In my preferred world where copyright and…

Right. So you use copyright laws because you need them. I wish we did not need money, but I don't go around telling people that they should stop using money because in my preferred world, money would not exist and we would all live happily.

I do not need copyright law. Just anti-liability licenses so i do not get sued. My interest is protecting myself, not the work.

I believe all knowledge should be freely available and live by that.

While I do not believe in IP law but I am happy to take money from those that do.

I sell my time so long as the buyer allows my work to be freely licensed for all. They can do with their own copy whatever they want and not credit me though. That is fine.

Re: Meta torrented & seeded 81.7 TB dataset containing copyrighted data

#954

Post by palata »

lrvick wrote: post #953
palata wrote: post #940

Earlier quoted context omitted.

Right. So you use copyright laws because you need them. I wish we did not need money, but I don't go around telling people that they should stop using money because in my preferred world, money would not exist and we would all live happily.

I do not need copyright law. Just anti-liability licenses so i do not get sued. My interest is protecting myself, not the work. I believe all knowledge should be freely available and live by that. While I do not believe in IP law but I am happy to take money from those that do. I sell my time so long as the buyer allows my work to be freely licensed for all. They can do with their own copy whatever they want and not…

It honestly feels like your view on the topic is entirely defined around yourself. Like if a billionaire could genuinely not understand why people need medical insurance, because "you can just pay cash, like I do".

You are in a special situation where customers pay for your time. Of course it's easy to say "as long as I am paid for my time, I want my work to be public". But say you create a video game. You invest two years working on it, and at the end you sell it for 50$. The first customer buys it, and publishes online for everybody to download legally for free. Does that sound okay for you?

Because you don't need copyright laws doesn't mean that nobody does. And again, you admitted yourself that you do need them and you do use them. You just wish you didn't. And I can agree with you on that. I wish there was no war, and no poverty, too.

I am not saying that everything about IP laws or copyright is perfect. Just that it's more complicated than "it's just a ball of worms, throw it away". Some people abuse copyright, some people need it. Whoever is fine screwing those who do because it doesn't affect them are jerks.

Re: Meta torrented & seeded 81.7 TB dataset containing copyrighted data

#955

Post by greenie_beans »

sanderjd wrote: post #945
greenie_beans wrote: post #887

Earlier quoted context omitted.

i wrote a book and copyright was not once on my mind. having created something is the incentive to create for most artists

I don't think we can infer the motives of most artists from your personal motives.

fine, because that means your claim that artists create for the copyright incentive is also false

Re: Meta torrented & seeded 81.7 TB dataset containing copyrighted data

#956

Post by BriggyDwiggs42 »

ericyd wrote: post #928
BriggyDwiggs42 wrote: post #922

Earlier quoted context omitted.

I just don’t think streaming is really working for anybody. Artists are being given less and less creative control as time goes on as each streaming company attempts to optimize itself in a largely fixed market. It’s working the same way cable used to work. I think a disruption is in order, but last time disruption started with people moving to nascent torrent sites. I guess the way I’d put it is that if you can only…

But is torrenting better than buying music e.g. through bandcamp or directly from the artist's label? I do a combination of streaming for exploratory listening and buying to support artists. And live shows when I can. I'm not trying to suggest I'm holier than thou, but I really don't think torrents are the answer to this equation. However I strongly agree that the current system isn't working for artists and needs di…

I know music can be bought from the artist, but is that true for shows or movies?

Re: Meta torrented & seeded 81.7 TB dataset containing copyrighted data

#957

Post by wing-_-nuts »

tiahura wrote: post #703
wing-_-nuts wrote: post #645

Earlier quoted context omitted.

Why would I?

As a general rule, I'm not a fan of prosecutors. However, blaming a prosecutor for the suicide of a defendant, because they filed charges in a clear cut case, is just as unfair as blaming friends and family.

I mean yes, the prosecutor had discretion to offer a light plea deal, and instead chose to try and ruin someone's life over papers. I absolutely blame the prosecutor.

Re: Meta torrented & seeded 81.7 TB dataset containing copyrighted data

#958

Post by ericyd »

BriggyDwiggs42 wrote: post #956
ericyd wrote: post #928

Earlier quoted context omitted.

But is torrenting better than buying music e.g. through bandcamp or directly from the artist's label? I do a combination of streaming for exploratory listening and buying to support artists. And live shows when I can. I'm not trying to suggest I'm holier than thou, but I really don't think torrents are the answer to this equation. However I strongly agree that the current system isn't working for artists and needs di…

I know music can be bought from the artist, but is that true for shows or movies?

Fair point, I tend to think more about music artists than film/TV.

Re: Meta torrented & seeded 81.7 TB dataset containing copyrighted data

#959

Post by palata »

roguecoder wrote: post #932
palata wrote: post #911

Earlier quoted context omitted.

> I publish 100% of code I write as FOSS. This is not public domain. > and so corporations bound by such silly rules have a difficult time using my work in private codebases without paying me for an alternate license. So... you benefit from copyright laws. > to ensure I do not get sued Why would you get sued by releasing your code as public domain?

Liability laws often don't care if something is free.

Which is beside the point. The person I am answering to says that they don't believe in copyright laws, but still they use licenses like MIT. Why not using CC0 then?

Re: Meta torrented & seeded 81.7 TB dataset containing copyrighted data

#960

Post by BriggyDwiggs42 »

ericyd wrote: post #958
BriggyDwiggs42 wrote: post #956

Earlier quoted context omitted.

I know music can be bought from the artist, but is that true for shows or movies?

Fair point, I tend to think more about music artists than film/TV.

Yeah for torrenting I tend to think more about video media.
Post reply on HN