4 ms·

Why is it fair use (hashed out in court already) for google to copy every book they can get a hold of and store the full pages and use them to create their n-gr

Why is it fair use (hashed out in court already) for google to copy every book they can get a hold of and store the full pages and use them to create their n-grams data and presumably to train their ai, but not for this company? If they had bought each book themselves would it be fair use? So this is only about the piracy?

sidewndr46 1y ago
The way I understand the case, yes. If Anthropic had just 1 copy of each book that would be permitted. Apparently they at some point bought copies of most books then shredded them. I'm not entirely sure how that makes any difference, but it was done after the piracy.
- qgin 1y ago
  At some point this starts to feel like witchcraft spells.
  - ACCount36 1y ago
    Machine learning is 20% programming, 30% math and 50% demon summoning.
    - danaris 1y ago
      That may be so, but this is really more about the law. So, y'know, probably more like 80% demon summoning.
- s1mplicissimus 1y ago
  > Apparently they at some point bought copies of most books then shredded them. An awfully convenient explanation of what went down. Gives some good "dog ate my homework" vibes. The double standard vs google etc. is of course despicable
  - tonyhart7 1y ago
    they would target google too but google money is so big that they could just buys the entire thing + publisher Google selling books as a publisher on playstore so it can get away with it
justonceokay 1y ago
N grams doesn’t threaten anyone’s business. To have a lawsuit you need damages.
ctkhn 1y ago
With n-gram it isn't just repeating the exact content of the book back to you the way the book itself does, which is what AI does, it's a statistical overview of the words from many books. Knowing what year use of the word "slouch" peaked in print isn't any kind of substitute for reading a book that uses the word.
- bigmadshoe 1y ago
  “which is what AI does” what? The term “statistical overview of the words from many books” is a great way to describe an LLM. It’s not like the weights encode every book verbatim.
- Workaccount2 1y ago
  >repeating the exact content of the book back to you the way the book itself does, which is what AI does Are you sure about that? Maybe on a good day you'll get a paragraph, but getting a few pages equivalent to a "book preview"? No shot.
kristofferR 1y ago
Yes, only about the piracy.
dfedbeef 1y ago
There's a 'first sale doctrine' in the US. Once you purchase a book, you can make a copy of it for yourself. Same with an album or game. "Copy" right. Get it? Right to copy.
- JohnFen 1y ago
  True. But you have to treat the thing you bought and all copies as an indivisible unit. If you sell or give away the thing or any of its copies, you have to include (or destroy) all other copies of that thing.
  - dfedbeef 1y ago
    Yep. Google shouldn't have won that suit, just like Marvin Gaye should not have won his suit against Bruno Mars. The courts don't always make rational decisions. They're dumb and corrupt.
ijk 1y ago
> If they had bought each book themselves would it be fair use? So this is only about the piracy? The earlier ruling covered exactly that question: - Anthropic downloaded many books (from LibGen and elsewhere). This piracy is what the current case is about, and is unrelated to the training. - Separately, Anthropic bought and scanned a million used books. They trained the AI on this data. This was ruled as fair use, and is not involved in the current case.
- Ajedi32 1y ago
  That's very interesting, because it totally makes sense legally, but the practical effect is ludicrously stupid. The law is effectively forcing companies to spend millions re-scanning the same books over and over for no reason. It'd be like if we had a law which stated "Before you can train an AI, you must light 1 million dollars on fire. After that you can do whatever you want.". It serves no purpose but to waste societal resources on nothing.
  - JohnFen 1y ago
    The cost might reduce the number of entities who can afford to do it, though, which would reduce the amount of abuse.
  - aswegs8 1y ago
    Why would they need to "re-scan the same books over and over"? It's as simple as they can use the books to train their AI if they bought them.
    - Ajedi32 1y ago
      Because company A needs to scan the books, then company B wants to train their AI so they need to scan the same books, then company C wants to train their AI so they need to scan the same books... etc. It would be one thing if they were buying "used" digital copies of the books, but the fact that this is only legal with scanned physical copies makes it extremely wasteful.

sidewndr46 1y ago

The way I understand the case, yes. If Anthropic had just 1 copy of each book that would be permitted. Apparently they at some point bought copies of most books then shredded them. I'm not entirely sure how that makes any difference, but it was done after the piracy.

qgin 1y ago
At some point this starts to feel like witchcraft spells.
- ACCount36 1y ago
  Machine learning is 20% programming, 30% math and 50% demon summoning.
  - danaris 1y ago
    That may be so, but this is really more about the law. So, y'know, probably more like 80% demon summoning.
s1mplicissimus 1y ago
> Apparently they at some point bought copies of most books then shredded them. An awfully convenient explanation of what went down. Gives some good "dog ate my homework" vibes. The double standard vs google etc. is of course despicable
- tonyhart7 1y ago
  they would target google too but google money is so big that they could just buys the entire thing + publisher Google selling books as a publisher on playstore so it can get away with it

justonceokay 1y ago

N grams doesn’t threaten anyone’s business. To have a lawsuit you need damages.

ctkhn 1y ago

With n-gram it isn't just repeating the exact content of the book back to you the way the book itself does, which is what AI does, it's a statistical overview of the words from many books. Knowing what year use of the word "slouch" peaked in print isn't any kind of substitute for reading a book that uses the word.

bigmadshoe 1y ago
“which is what AI does” what? The term “statistical overview of the words from many books” is a great way to describe an LLM. It’s not like the weights encode every book verbatim.
Workaccount2 1y ago
>repeating the exact content of the book back to you the way the book itself does, which is what AI does Are you sure about that? Maybe on a good day you'll get a paragraph, but getting a few pages equivalent to a "book preview"? No shot.

kristofferR 1y ago

Yes, only about the piracy.

dfedbeef 1y ago

There's a 'first sale doctrine' in the US. Once you purchase a book, you can make a copy of it for yourself. Same with an album or game. "Copy" right. Get it? Right to copy.

JohnFen 1y ago
True. But you have to treat the thing you bought and all copies as an indivisible unit. If you sell or give away the thing or any of its copies, you have to include (or destroy) all other copies of that thing.
- dfedbeef 1y ago
  Yep. Google shouldn't have won that suit, just like Marvin Gaye should not have won his suit against Bruno Mars. The courts don't always make rational decisions. They're dumb and corrupt.

ijk 1y ago

> If they had bought each book themselves would it be fair use? So this is only about the piracy? The earlier ruling covered exactly that question: - Anthropic downloaded many books (from LibGen and elsewhere). This piracy is what the current case is about, and is unrelated to the training. - Separately, Anthropic bought and scanned a million used books. They trained the AI on this data. This was ruled as fair use, and is not involved in the current case.

Ajedi32 1y ago
That's very interesting, because it totally makes sense legally, but the practical effect is ludicrously stupid. The law is effectively forcing companies to spend millions re-scanning the same books over and over for no reason. It'd be like if we had a law which stated "Before you can train an AI, you must light 1 million dollars on fire. After that you can do whatever you want.". It serves no purpose but to waste societal resources on nothing.
- JohnFen 1y ago
  The cost might reduce the number of entities who can afford to do it, though, which would reduce the amount of abuse.
- aswegs8 1y ago
  Why would they need to "re-scan the same books over and over"? It's as simple as they can use the books to train their AI if they bought them.
  - Ajedi32 1y ago
    Because company A needs to scan the books, then company B wants to train their AI so they need to scan the same books, then company C wants to train their AI so they need to scan the same books... etc. It would be one thing if they were buying "used" digital copies of the books, but the fact that this is only legal with scanned physical copies makes it extremely wasteful.