285
AI companies destroy physical books — let’s scan rare books before it’s too late (annas-archive.gl)
submitted 2 months ago* by to c/technology@lemmy.world
29 comments fedilink hide all child comments
you are viewing a single comment's thread
view the rest of the comments
[–] 6 points 2 months ago* (last edited 2 months ago) (9 children)

The destruction of the book is done to mitigate some copyright concerns. There will be a fairly small number of large-scale projects like this, and each only needs one copy of a given book, so outside of maybe very rare books, the impact is going to be pretty negligible on availability

like someone buying one copy of that book.

  • permalink
  • fedilink
  • source
  • hideshow 9 child comments
    •
  • [–] 1 point 2 months ago (7 children)

    The destruction of the book is done to mitigate some copyright concerns.

    Can you explain? The First Sale doctrine means its OK to sell books after you're done with them. I would think it would be the scanning itself that could be a copyright problem, not the disposition of the physical book after.

    I thought the real reason, the main reason, was, as the article says, "Destroying books is cheaper."

    I agree that the concern over the destruction of physical books, and what impact that will have on their availability, is probably exaggerated. Heck, the actual importance of old books to the AI companies may be exaggerated, that seems to be their hype style.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 7 child comments
    •
  • [–] 5 points 2 months ago (6 children)

    404 media reported this recently. Excerpt from the article :

    It’s possible to scan books without destroying them, but cutting the spine makes it cheaper and faster. Additionally, the judge in the lawsuit ruled it was fair use and not a copyright violation for Anthropic to scan a book for training data in part because it destroyed the original, printed copy. Essentially, it’s the customer’s right to take physical media and store it digitally, and destroying the original copy means that copy isn’t duplicated and resold, and isn’t cutting into the publisher’s business.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 6 child comments
    •
  • [–] 2 points 2 months ago (5 children)

    Then they plagarize that data though, using it without attribution or payment, which is aganst copyright law. If a person did this they would be stopped. Because it's big money, the law doesn't apply.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 5 child comments
    •
  • [–] -1 points 2 months ago (4 children)

    Now you've entered the land of fairy tales and make believe. A person who memorizes a book and then burns it and then recites to others would in fact not be put in front of a firing squad like you want to.

    Not saying corporations are people (they are though) but they are literally not doing anything else than that. Now they only have to teach the AI to learn double think, to learn from the book, but then forget the book, only to be able to remember what that book was about, while being unable to faithfully repeat it.

    That's literally going to solve any and all problems with AI. It's not like we should demand meaningful regulations about the actual problems like authoritarian China does. You know, because, think of the book publishers. They are people too!

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 4 child comments
    •
  • [–] 3 points 2 months ago (3 children)

    If a person reproduced copyrighted works and sold them, they could be sued under patent law. In the US, and in all of europe that I'm aware, at least in the west. In Japan, in South Korea, and a host of other countries.

    Reciting a book is one thing, reproducing parts of it for commercial gain is another and is absolutely a violation of patent law. Just as a musician that takes a song and changes the lyrics and beat a bit and releases it would be sued.

    Do you understand the difference?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 1 point 2 months ago (2 children)

    Yeah, memorization is a bug they have to fix. Like I said, then all problems will be solved.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 1 point 2 months ago (1 child)

    No problems will be solved in our favour. Us being those that aren't the oligarchy and their well to do tools.

    I'm surprised to hear someone on .ml defending llm's. I thought you guys knew which side you were on on this issue?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 1 point 2 months ago

    I was sarcastic. Maybe I'm being a bit unfair today.

    But while fuckai is busy beating the drum for intellectual property expansion, a non-capitalist state like China has regulated their AI preventing workers being replaced by AI or regulating emotionally addictive companion bots. Meanwhile I have not heard a single actionable demand for regulation from fuckai.

    So there is a problem with capitalists abusing AI. There is no problem fundamentally with LLMs themselves, they are just tools. They could be amazing to help with economic planning because it could allow to parse democratic input from millions of people.

    IP laws are very much a tool of imperialist oppression. Patents are outright genocide, preventing rapid adoption of new technologies to reduce greenhouse gas emissions. And ultimately the IP argument will lead to open weight models becoming "illegal", monopolizing the "means of generation" and advanced robotics in the hands of the plutocrats. That is going to be "it" for the future of humanity then. They are going to pay a pittance in "general license costs", replace more and more labor, while you can try to pirate an open source model to use this emergent technological power yourself.

  • permalink
  • fedilink
  • source
  • parent