That's the point of publishing deals, though. Authors don't need to pay upfront costs, in return the publishers get to keep the revenues. Not really related to AI.
Fifty-four known class members and over 500,000 potential class members will each receive an estimated $3,100 per work, from approximately 482,000 works.
Still, paying even just the $3000 for a single book from this lawsuit is going to be quite hard when you are making zero profit whatsoever. LLM AI was unprofitable when the training data was free, they are even more so if they have to pay even a cent per item.
And I really doubt it's going to stop with this one batch and payment and they can just go back to pirating the stuff for free.
Well in this case its lump sum or percentage of negative profit, so lump sum might be the better path here... Still not remotely enough compensation to the actual craters of the content they've stolen though.
That's why fines should always be out of revenue, like the GDPR fines, and never profit. Just because the business you did illegally was bad business shouldn't mean you don't have to pay fines for it.
No, that's entirely fair, and I agree should be the SOP. I was just thinking of the recent articles about how AI companies in general are hiding massive amounts of debt a-la Enron style, thus likely even their revenue figures are made up.
I wonder how the venture capital firms that fund Anthropic's money-burning operation feel about this. "Oh look, now I know how that particular funding round's money was actually spent. Fantastic."
Just forwarded the search link to a friend. He coauthored a long out-of-date book on podcasting, and it's on their list. A lot of these authors will get a nice bonus.
Thia is to settle their infringement from when they literally torrented giant chunks of their training corpus. It has nothing to do with any sort of infringement theory based on encoding learning from material but just old fashioned liability for flagrant piracy.
27 Comments
gravitas_deficiency@sh.itjust.works · 115 pts · 30d
Wow, so uh, that’s gonna come out to fractions of a cent on the dollar, in terms of actual revenue generated from flagrant copyright violations.
It is deeply infuriating how fundamentally our society is designed to protect of people and organizations that are already blindingly wealthy.
username_1@discuss.tchncs.de · 79 pts · 30d
Authors will receive next to nothing too :) Only publishers.
state_electrician@discuss.tchncs.de · 29 pts · 30d
Just like God intended.
woelkchen@lemmy.world · -2 pts · 29d
That's the point of publishing deals, though. Authors don't need to pay upfront costs, in return the publishers get to keep the revenues. Not really related to AI.
XLE@piefed.social · 11 pts · 30d
Still too low IMO but a big upgrade
JohnEdwa@sopuli.xyz · 10 pts · 30d
Still, paying even just the $3000 for a single book from this lawsuit is going to be quite hard when you are making zero profit whatsoever. LLM AI was unprofitable when the training data was free, they are even more so if they have to pay even a cent per item. And I really doubt it's going to stop with this one batch and payment and they can just go back to pirating the stuff for free.
Trilogy3452@lemmy.world · 2 pts · 29d
Will this perhaps set a precedent for future models where authors will be able to price their work differently?
richardisaguy@lemmy.world · 0 pts · 30d
https://youtu.be/cBHouBDjqsc
piskertariot@lemmy.world · 87 pts · 30d
$1.5B, and they get to keep the knowledge, which will be distilled by future models for generations.
You will own nothing. Not even your own art.
OrganicMustard@lemmy.world · 42 pts · 30d
Why is a single payment for an arbitrary amount of money still enforced to the authors?
Anthropic should have all their models deleted so they start from the beginning following copyright laws.
frongt@lemmy.zip · 4 pts · 30d
The claimants have deemed it acceptable.
UltraBlack@lemmy.world · 2 pts · 29d
And it's probably a fine line between something and the entire deal being deemed excessive
itrealgood@mander.xyz · 22 pts · 30d
Pyotr@lemmy.world · 4 pts · 30d
Well in this case its lump sum or percentage of negative profit, so lump sum might be the better path here... Still not remotely enough compensation to the actual craters of the content they've stolen though.
JohnEdwa@sopuli.xyz · 15 pts · 30d
That's why fines should always be out of revenue, like the GDPR fines, and never profit. Just because the business you did illegally was bad business shouldn't mean you don't have to pay fines for it.
Pyotr@lemmy.world · 5 pts · 30d
No, that's entirely fair, and I agree should be the SOP. I was just thinking of the recent articles about how AI companies in general are hiding massive amounts of debt a-la Enron style, thus likely even their revenue figures are made up.
errer@lemmy.world · 15 pts · 30d
Check if Anthropic owes you $3100 here: https://secure.anthropiccopyrightsettlement.com/lookup
Scotty_Trees@lemmy.world · 4 pts · 29d
Good link, but I don't think that's going to apply to any of us here 🤷♀️️
kescusay@lemmy.world · 8 pts · 30d
I wonder how the venture capital firms that fund Anthropic's money-burning operation feel about this. "Oh look, now I know how that particular funding round's money was actually spent. Fantastic."
ranzispa@mander.xyz · 6 pts · 29d
I know this has nothing to do with that, but the word publishers always reminds me of a thing.
FUCK YOU ELSEVIER
fubarx@lemmy.world · 4 pts · 30d
Just forwarded the search link to a friend. He coauthored a long out-of-date book on podcasting, and it's on their list. A lot of these authors will get a nice bonus.
sanitation@lemmy.today · 2 pts · 30d
my company gonna sue them too
scottmeme@sh.itjust.works · 4 pts · 30d
How much does this affect their bankruptcy estimate?
ReluctantZen@feddit.nl · 2 pts · 30d
But this means there's still no precedence, no?
FatCrab@slrpnk.net · 3 pts · 30d
Thia is to settle their infringement from when they literally torrented giant chunks of their training corpus. It has nothing to do with any sort of infringement theory based on encoding learning from material but just old fashioned liability for flagrant piracy.
frongt@lemmy.zip · 0 pts · 30d
Yes. A settlement means there was no finding, and no admission of guilt.
melfie@lemmy.zip · 1 pts · 30d
ryannathans@aussie.zone · 0 pts · 29d
Interesting to see Lemmy users who are normally "information should be free" supporters suddenly support the copyright system
tb_@lemmy.world · 6 pts · 29d
Information between you and me? Yes.
Information a trillion dollar company really wants to get their greedy hands on? Fuck you, play by the rules of the society you have shaped.