Anthropic agrees to pay $1.5B to settle lawsuit with book authors

This is sad for open source AI, piracy for the purpose of model training should also be fair use because otherwise only the big companies who can afford to pay off publishers like Anthropic will be able to do so. There is no way to buy billions of books just for model training, it simply can't happen.

This implies training models is some sort of right.

No, it implies that having the power to train AI models exclusively consolidated into a handful of extremely powerful companies is bad.

No. It means model training is transformative enough to be fair use. They should just be asked to pay them back plus reimbursement/punishment, say pay 10x the price of the pirated books

I wonder how much it would cost to buy every book that you'd want to train a model.

Fair use isn't about how you access the material, its about what you can do with it after you legally access it. If you don't legally access it, the question of fair use is moot.

That's true. Those handful of companies shouldn't get to do it either.

500,000 x $20 = $10 million

Obviously there would be handling costs + scanning costs, so that’s the floor.

Maybe $20 million total? Plus, of course, the time it would take to execute.

It implies that people want everyone to do this when it's clear no one should do it. I'm not exactly a fan of "this isn't profitable for small businesses to steal from so we should make it so everyone should steal".

Hence, "should"

Piracy is not stealing. I don't know why everyone on HN suddenly turned into a copyright hawk, only big companies benefit from our current copyright regime, like Disney and their lobbying for increasing its length.

I wish the hn rules were more flexible because I would write the best comment to you right now.

This is a settlement. It does not set a precedent nor even admit to wrongdoing.

> otherwise only the big companies who can afford to pay off publishers like Anthropic will be able to do so

Only well funded companies can afford to hire a lot of expensive engineers and train AI models on hundreds of thousands of expensive GPUs, too.

Something tells me many the grassroots LLM training people are less concerned about legality of their source training set than the big companies anyway.

> only big companies benefit from our current copyright regime

You’ve never authored, created, or published something? Never worked for a company that sells something protected by copyright?

All my works are open source or in the public domain. I don't like copyright for a reason.

> Never worked for a company that sells something protected by copyright?

I.e. never created software in exchange of money.

It’s the sign of a health economy when we respect the creation of content.

Isn’t that already the case, with the capacity required to train these models?

It’s shifts like that, going from a copyleft to a copyright crowd, that make me increasingly suspicious the HN is authentic as it was years ago. Another weird one is the socialism lean instead of more libertarian ideals by many commentators. I think it might be generational issues and being 50+ years old makes me an old timer!

It's a sign of rent seeking economy in decline. Rising economies never respect IPs.

I don't know if I agree with it, but you could argue that if a model was built for purely academic purposes, and then used for purely academic purposes, it could meet requirements for fair use.

It isn't black and white, you can be against some aspects of copyright and be for some others.

It’s very funny when people declare “piracy isn’t stealing”, as if the metaphor of piracy is all about singing and drinking.

Copying and distributing works isn’t identical to theft (deliberately depriving someone of their property), but you’re enjoying someone’s work without compensating them, so it isn’t totally unlike depriving them of something.

I guess it depends how you feel about refusing to pay a window washer. Or indeed you not being paid by your employer. It isn’t theft, but someone is clearly stiffing someone else.

As for only big companies benefitting from the copyright regime… seems like an ideological assumption. I know plenty of authors and they are quite happy having legal protections around their work which means they can earn from their labour.

Curious: are you trolling? Or do you really think doing math shouldn’t be a priori allowed to humans?

(Half joking but) I wonder if musicians need to worry if they learned to play by listening to cassette mixtapes.

> you’re enjoying someone’s work without compensating them

Which is foreseen in societal decision: libraries (again and again).

> refusing to pay a window washer

The window washer is providing a service for a price, that service is not equivalent to knowledge production, and nobody has decided that that service (cleaning windows) should be done for free.

Setting aside whether or not I think it should be fair use, you’re only going to be training a new foundation model these days if you have billions of dollars to spend on the endeavor anyway. Nobody is training Llama 5 in their garage.

I'm guessing it's because every year, younger people join HN. I've been on this site for at least a decade and a half with various accounts and it's interesting to see the shift.

IP protections are in the US Constitution. Has the US been in decline since the late 1700s?

Yes. Great recession was its death throes. It would have fallen by now if it wasn't the only economy not damaged by Second World War and as a result economy that put its currency as currency of global trade. Since then it's mostly sustained by selling freshly printed paper and bullying everyone around with IP laws.

Do you seriously think it's a matter of doing math in a vacuum? The issue is pirating books, using them as inputs to math which is then resold as an LLM without compensation to the authors who have clearly copyrighted material. I can't tell if _you're_ trolling.

Libraries are great and authors get paid when you borrow the book. But we’re talking about piracy.

As for window washing vs knowledge production, not sure what you mean. Books have a price. Nobody’s decided they should be free either.

> authors get paid

In the standard system for libraries, the book is paid once.

That is "libraries" as in "we have societally decided to make published knowledge freely available".

> window washing vs knowledge production

Societies have not decided that window washing should be freely available - on the other side, they have decided that published knowledge be freely available (that is the meaning of the establishment of libraries).

It must differ by geography. In my country, authors are paid library revenue according to the number of times their books are borrowed. Perhaps as well as an upfront purchase.

Re: what “society has decided”, are you arguing that because libraries exist, no one may sell books for a price?

Seems extreme, not widely agreed by the population or the relevant parties, and likely to cause immense problems with the economics of knowledge production, but it’s certainly one point of view!

Similarly I argue that because open source code exists, software engineers must all work for free, and that because public parks exist, everyone’s home gardens are open to all.

The real expense is in the data centers/hardware.

The cost of the books is negligible in comparison.

Somewhere a gritty warehouse in a developing country is receiving shipping containers of old books, massive teams manually flipping each page as a 2nd hand Canon digicam takes a pic of each page, to be OCR’d by the same AI being trained.

Once the book is done, 99% of them go into the furnace at the district heating boiler next door. The other 1% back to a developed country for resale.

> The issue is pirating books, using them as inputs to math which is then resold as an LLM without compensation to the authors who have clearly copyrighted material.

No one seems to be able to explain what exactly the issue is here. How are authors harmed by LLMs (where such harm is often used to understand the range of copyright in lawsuits)? I don't see anyone replacing authors and their works, such as JK Rowling being harmed just because people can output Harry Potter esque texts. And if that's the case, well, fan fiction has been around for a long time, with no LLMs in its writing.

Millions, not billions, as DeepSeek and others have shown.

> are you arguing that because libraries exist

No, no, no. In general, contextually to the topic of which the submission in part, I am showing that libraries were established through a societal decision that published knowledge shall be freely available. That means, if you want to consult a text, you are enabled to just go to a library and do it - the cost will be societal and contained.

Contextually to your post, and the expression «enjoying someone’s work without compensating them», I showed that the Principle establishing libraries implies you are not required to directly compensate authors to access their work.

And I told you that there is no similar principle regulating access to all other goods or services, such as "mowing your lawn" - society has not decided to bear the cost of the operation. It has, for the realm of accessing knowledge.

Do not misread.

Edit: maybe this will further help you to understand:

some societies have decided that health care services shall be easy to access, with collectively borne costs. Most societies have decided that published knowledge shall be easy to access, with collectively borne costs. Few societies have decided that having private lawns mowed shall be easy to access, with collectively borne costs.

Are you implying some sort of threat?

With respect to the IP protected under the Constitution of the United States, unquestionably and absolutely. The past 50 years of macroeconomic activity is basically a warning sign that says "your IP is basically worthless" in big red flashing letters. China mass-manufactures ARM chips and turbojet engines using unlicensed stolen schematics - Russia endlessly pirates foreign entertainment with impunity. These are entire nations of people who benefit from American financial assets without paying us a dime for them.

In case you missed it, the other day China showed off their F-35 clone for the PRC military parade. Your constitution can say pigs fly for all George Washington cares, trusting absolutely in IP protection is a game for butthurt chumps.

Look into how the US industrial economy started. It was "stealing IP" from the english and not respecting copyright. Would be interested if there are any cases in history of rising economies respecting IP and succeeding because in every case I know about they flagrantly disregard it while developing

And the British "stole" from Dutch before that.

"open source" and "in the public domain" aren't like separate things from "copyright", they describe specific sub-sets of stuff underneath "copyright", which is a top-level category that establishes a meaningful definition of stuff like "your work(s)" in the first place

i.e. "copyright" describes a legal concept, "copyleft" describes a licensing concept

The point is the legal concept is a fiction, useful perhaps in the past, but increasingly not today, and like all fictions, it can be dispelled if people agree to do so.

i mean sure if you're gonna take some epistemological position then every legal concept is a fiction, but that's not a particularly interesting nor useful angle on the discussion

or, to put it more succinctly, and quoting one of the great contemporary aphorisms: sure, and if my grandmother had wheels, she would have been a bicycle