6 ms·

If using copyrighted material to train an LLM is theft, so is reading a book.

bakugo 2y ago
How is a human reading a book in any way related or comparable to a machine ingesting millions of books per day with the goal of stealing their content and replacing them?
- ysofunny 2y ago
  it's comparable exactly in the way 0.001% can be compared to 10^100 humans learning is the old-school digital copying. computers simply do it much faster, but it's the same basic phenomenon consider one teacher and one student. first there is one idea in one head but then the idea is in two heads. now add book technology1 the teacher writes the book once, a thousand students read it. the idea has gone from being in one head (book author) onto most of the book readers!
  - somenameforme 2y ago
    > humans learning is the old-school digital copying. computers simply do it much faster, but it's the same basic phenomenon Train an LLM on the state of human knowledge 100,000 years ago - language had yet to be invented and bleeding edge technology was 'poke them with the pointy side.' It's not going to be able to do or output much of anything, and it's going to be stuck in that state for perpetuity until somebody gives it something new to parrot. Yet somehow humans went from that exact starting to state to putting a man on the Moon. Human intelligence, and elaborate auto-complete systems, are not the same thing, or even remotely close to the same thing.
    - Terr_ 2y ago
      > bleeding edge technology was 'poke them with the pointy side.' Relevant: https://www.smbc-comics.com/comic/rise-of-the-machines https://www.smbc-comics.com/comic/rise-of-the-machines
  - dasil003 2y ago
    > humans learning is the old-school digital copying. computers simply do it much faster, but it's the same basic phenomenon This is dangerous framing because it papers over the significant material differences between AI training and human learning and the outcomes they lead to. We all have a collective interest in the well-being of humanity, and human learning is the engine of our prosperity. Each individual has agency, and learning allows them to conceive of new possibilities and form new connections with other humans. While primarily motivated by self interest, there is natural collective benefit that emerges since our individual power is limited, and cooperation is necessary to achieve our greatest works. AI on the other hand, is not a human with interests, it's an enormously powerful slave that serves those with the deep pockets to train them. It can siphon up and generate massive profits from remixing the entire history of human creativity and knowledge creation without giving anything back to society. It's novelty and scale makes it hard for our legal and societal structures to grapple with—hence all the half-baked analogies—but the impact that it is having will change the social fabric as we know it. Mechanistic arguments about very narrow logical equivalence between human and AI training does nothing but support the development of an AI oligarchy that will surely emerge if human value is not factored in to how we think about AI regulation.
    - ysofunny 2y ago
      you're reading what I say in the worst possible light if anything, the parallel I draw between AI learning and humans learning is all the opposite of narrow and logical... in my intent, the analogy is loose and poetic, not mechanistic and exact. AI are tools, if AI are enslaving is because there are human actors (I hope....) deciding to enslave other humans, not because of anything inherent to training (if AI; learning if humans) but what I really think is that there are collections of rules (people "just doing their jobs") all collectively but disjointedly deciding that it makes the most sense to utilize AI technology to ensalve other humans because the data models indicate greater profit that way.
      - dasil003 2y ago
        Your response is fair and I hope you didn't take my message personally. I agree with you, AI is just a tool same as countless others that can be used for good or evil.
- supriyo-biswas 2y ago
  Because humans cannot reasonably memorize and recall thousands of articles and books in the same way, and because humans are entitled to certain rights and privileges that computer systems are not. (If we are to argue the latter point then it would also raise interesting implications; are we denying freedom of expression to a LLM when we fine-tune it or stop its generation?)
- JoosToopit 2y ago
  Directly. What if while reading you make notes - are you strealing content? If yes - should then people be forbidden from taking notes? How does writing down a note onto a piece of paper differ from writing it into your memory?
  - wonnage 2y ago
    The nice thing about law as opposed to programming is that legal scholars have long realized it's impossible to cover every possible edge case in writing so judges exist to interpret the law So they could easily decide logically unsound things that make pedants go nuts, like taking notes, or even an AI system that automatically takes notes, could be obvious fair use, while recording the exact same strings for training AI are not.
    - ysofunny 2y ago
      > The nice thing about law as opposed to programming in programming that is called "Undefined behavior"
surfingdino 2y ago
Reading a book is not theft. Building a business on processing other people's copyrighted material to produce content is.
- bmicraft 2y ago
  I think that's called a school
  - insane_dreamer 2y ago
    Schools pay for books, or use public domain materials
  - dylan604 2y ago
    If you think going to school to get an education is the same thing as training an LLM then you are just so misguided. Normal people read books to gain an understanding of a concept, but do not retain the text verbatim in memory in perpetuity. This is not what training an LLM does.
    - brookst 2y ago
      Some people memorize verbatim. Most LLM knowledge is not memorized. Easy proof: source material is in one language, and you can query LLMs in tens to a hundred plus. How can it be verbatim in a different language?
      - dylan604 2y ago
        These "some people" would not fall under the "normal people" that I specifically said. but you go right ahead and keep thinking they are normal so you can make caveats on an internet forum.
      - mitthrowaway2 2y ago
        If you buy a copy of Harry Potter from the bookstore, does that come with the right to sell machine-translated versions of it for personal profit? If so, how come even fanfiction authors who write every word themselves can't sell their work?
snyp 2y ago
But you pay money to buy a book and read it.
- dylan604 2y ago
  Not if you check it out from the library
  - hiatus 2y ago
    The library paid. Similarly, you can't go to a public library, photocopy entire books, then offer them for sale behind a subscription based chatbot.
  - MikeTheGreat 2y ago
    >Not if you check it out from the library ...who paid money for the book on your behalf
Angostura 2y ago
Is reading a book the same as photocopying it for sale? Which of the scenarios above is more similar to using it to train a LLM?
- ben_w 2y ago
  If I was forced to pick, LLMs are closer to reading than to photocopying. But, and these are important, 1) quantity has a quality all of its own, and 2) if a human was employed to answer questions on the web, then someone asked them to quote all of e.g. Harry Potter, and this person did so, that's still copyright infringement.
hiatus 2y ago
Computers are not people. Laws differ and consequences can be different based on the actor (like how minors are treated differently in courts). Just because a person can do it does not automatically mean those same rights transfer to arbitrary machines.
- dr_dshiv 2y ago
  Corporations are people. Not saying that’s right. But is that not the law?
  - mftrhu 2y ago
    Corporations are legal persons, which are not the same as natural persons (AKA plain old human beings). The law endows natural persons with many rights which cannot and do not apply to legal persons - corporations, governments, cooperatives and the like can enter into contracts (but not marriage contracts), own property (which will not be protected by things like homestead laws and the such), sue, and be sued. They cannot vote, claim disability exemptions, or have any rights to healthcare and the like, while natural persons do. Legal persons are not treated and do not have to be treated like natural persons.
mrweasel 2y ago
So if I get access to the Perplexity AI source code (I borrow it from a friend), read all of it, and reproduce it at some level, then Perplexity will be:" sure, that's fine no harm, no IP theft, no copyright violation, because you read it so we're good"? No, they would sue me for everything I got, and then some. That's the weird thing about these companies, they are never afraid to use IP law to go after others, but those same laws don't apply to them... because? Just pay the stupid license and if that makes your business unsustainable then it's not much a business is it?
- bshacklett 2y ago
  If Perplexity’s source code is downloaded from a public web site or other repository, and you take the time to understand the code and produce your own novel implementation, then yes. Now, if you “get it from a friend”, illegally, _or_ you just redeploy the code, without creating a transformative work, then there’s a problem. > Just pay the stupid license and if that makes your business unsustainable then it's not much a business is it? In the persona of a business owner, why pay for something that you don’t legally, need to pay for? The question of how copyright applies to LLMs and other AI is still open. They’d be fools to buy licenses before it’s been decided. More importantly, we’re potentially talking about the entire knowledge of humanity being used in training. There’s no-one on earth with that kind of money. Sure, you can just say that the business model doesn’t work, but we’re discussing new technologies that have real benefit to humanity, and it’s not just businesses that are training models this way. Any decision which hinders businesses from developing models with this data will hinder independent researchers 10 fold, so it’s important that we’re careful about what precedent is set in the name of punishing greedy businessmen.
  - flutas 2y ago
    > If Perplexity’s source code is downloaded from a public web site or other repository, and you take the time to understand the code and produce your own novel implementation, then yes. Even that can be considered infringement and get you taken to court. It's one of the reasons reading leaked code is considered bad and you hear terms like cleanroom[0] when discussing reproductions of products. [0]: https://en.wikipedia.org/wiki/Clean_room_design https://en.wikipedia.org/wiki/Clean_room_design
Terr_ 2y ago
If the LLM is automatically equivalent to a human doing the same task, that means it's even worse: The companies are guilty of slavery. With children. It also means reworking patent law, which holds that you can't just throw "with a computer" onto something otherwise un-patentable. Clearly, there are other factors to consider, such as scope, intended purpose, outcome...
timeon 2y ago
Is it same as human reading a book? We are not even giving same rights to other mammals. So why should we give it to software.

How is a human reading a book in any way related or comparable to a machine ingesting millions of books per day with the goal of stealing their content and replacing them?

ysofunny 2y ago
it's comparable exactly in the way 0.001% can be compared to 10^100 humans learning is the old-school digital copying. computers simply do it much faster, but it's the same basic phenomenon consider one teacher and one student. first there is one idea in one head but then the idea is in two heads. now add book technology1 the teacher writes the book once, a thousand students read it. the idea has gone from being in one head (book author) onto most of the book readers!
- somenameforme 2y ago
  > humans learning is the old-school digital copying. computers simply do it much faster, but it's the same basic phenomenon Train an LLM on the state of human knowledge 100,000 years ago - language had yet to be invented and bleeding edge technology was 'poke them with the pointy side.' It's not going to be able to do or output much of anything, and it's going to be stuck in that state for perpetuity until somebody gives it something new to parrot. Yet somehow humans went from that exact starting to state to putting a man on the Moon. Human intelligence, and elaborate auto-complete systems, are not the same thing, or even remotely close to the same thing.
  - Terr_ 2y ago
    > bleeding edge technology was 'poke them with the pointy side.' Relevant: https://www.smbc-comics.com/comic/rise-of-the-machines https://www.smbc-comics.com/comic/rise-of-the-machines
- dasil003 2y ago
  > humans learning is the old-school digital copying. computers simply do it much faster, but it's the same basic phenomenon This is dangerous framing because it papers over the significant material differences between AI training and human learning and the outcomes they lead to. We all have a collective interest in the well-being of humanity, and human learning is the engine of our prosperity. Each individual has agency, and learning allows them to conceive of new possibilities and form new connections with other humans. While primarily motivated by self interest, there is natural collective benefit that emerges since our individual power is limited, and cooperation is necessary to achieve our greatest works. AI on the other hand, is not a human with interests, it's an enormously powerful slave that serves those with the deep pockets to train them. It can siphon up and generate massive profits from remixing the entire history of human creativity and knowledge creation without giving anything back to society. It's novelty and scale makes it hard for our legal and societal structures to grapple with—hence all the half-baked analogies—but the impact that it is having will change the social fabric as we know it. Mechanistic arguments about very narrow logical equivalence between human and AI training does nothing but support the development of an AI oligarchy that will surely emerge if human value is not factored in to how we think about AI regulation.
  - ysofunny 2y ago
    you're reading what I say in the worst possible light if anything, the parallel I draw between AI learning and humans learning is all the opposite of narrow and logical... in my intent, the analogy is loose and poetic, not mechanistic and exact. AI are tools, if AI are enslaving is because there are human actors (I hope....) deciding to enslave other humans, not because of anything inherent to training (if AI; learning if humans) but what I really think is that there are collections of rules (people "just doing their jobs") all collectively but disjointedly deciding that it makes the most sense to utilize AI technology to ensalve other humans because the data models indicate greater profit that way.
    - dasil003 2y ago
      Your response is fair and I hope you didn't take my message personally. I agree with you, AI is just a tool same as countless others that can be used for good or evil.
supriyo-biswas 2y ago
Because humans cannot reasonably memorize and recall thousands of articles and books in the same way, and because humans are entitled to certain rights and privileges that computer systems are not. (If we are to argue the latter point then it would also raise interesting implications; are we denying freedom of expression to a LLM when we fine-tune it or stop its generation?)
JoosToopit 2y ago
Directly. What if while reading you make notes - are you strealing content? If yes - should then people be forbidden from taking notes? How does writing down a note onto a piece of paper differ from writing it into your memory?
- wonnage 2y ago
  The nice thing about law as opposed to programming is that legal scholars have long realized it's impossible to cover every possible edge case in writing so judges exist to interpret the law So they could easily decide logically unsound things that make pedants go nuts, like taking notes, or even an AI system that automatically takes notes, could be obvious fair use, while recording the exact same strings for training AI are not.
  - ysofunny 2y ago
    > The nice thing about law as opposed to programming in programming that is called "Undefined behavior"

surfingdino 2y ago

Reading a book is not theft. Building a business on processing other people's copyrighted material to produce content is.

bmicraft 2y ago
I think that's called a school
- insane_dreamer 2y ago
  Schools pay for books, or use public domain materials
- dylan604 2y ago
  If you think going to school to get an education is the same thing as training an LLM then you are just so misguided. Normal people read books to gain an understanding of a concept, but do not retain the text verbatim in memory in perpetuity. This is not what training an LLM does.
  - brookst 2y ago
    Some people memorize verbatim. Most LLM knowledge is not memorized. Easy proof: source material is in one language, and you can query LLMs in tens to a hundred plus. How can it be verbatim in a different language?
    - dylan604 2y ago
      These "some people" would not fall under the "normal people" that I specifically said. but you go right ahead and keep thinking they are normal so you can make caveats on an internet forum.
    - mitthrowaway2 2y ago
      If you buy a copy of Harry Potter from the bookstore, does that come with the right to sell machine-translated versions of it for personal profit? If so, how come even fanfiction authors who write every word themselves can't sell their work?

snyp 2y ago

But you pay money to buy a book and read it.

dylan604 2y ago
Not if you check it out from the library
- hiatus 2y ago
  The library paid. Similarly, you can't go to a public library, photocopy entire books, then offer them for sale behind a subscription based chatbot.
- MikeTheGreat 2y ago
  >Not if you check it out from the library ...who paid money for the book on your behalf

Angostura 2y ago

Is reading a book the same as photocopying it for sale? Which of the scenarios above is more similar to using it to train a LLM?

ben_w 2y ago
If I was forced to pick, LLMs are closer to reading than to photocopying. But, and these are important, 1) quantity has a quality all of its own, and 2) if a human was employed to answer questions on the web, then someone asked them to quote all of e.g. Harry Potter, and this person did so, that's still copyright infringement.

hiatus 2y ago

Computers are not people. Laws differ and consequences can be different based on the actor (like how minors are treated differently in courts). Just because a person can do it does not automatically mean those same rights transfer to arbitrary machines.

dr_dshiv 2y ago
Corporations are people. Not saying that’s right. But is that not the law?
- mftrhu 2y ago
  Corporations are legal persons, which are not the same as natural persons (AKA plain old human beings). The law endows natural persons with many rights which cannot and do not apply to legal persons - corporations, governments, cooperatives and the like can enter into contracts (but not marriage contracts), own property (which will not be protected by things like homestead laws and the such), sue, and be sued. They cannot vote, claim disability exemptions, or have any rights to healthcare and the like, while natural persons do. Legal persons are not treated and do not have to be treated like natural persons.

mrweasel 2y ago

So if I get access to the Perplexity AI source code (I borrow it from a friend), read all of it, and reproduce it at some level, then Perplexity will be:" sure, that's fine no harm, no IP theft, no copyright violation, because you read it so we're good"? No, they would sue me for everything I got, and then some. That's the weird thing about these companies, they are never afraid to use IP law to go after others, but those same laws don't apply to them... because? Just pay the stupid license and if that makes your business unsustainable then it's not much a business is it?

bshacklett 2y ago
If Perplexity’s source code is downloaded from a public web site or other repository, and you take the time to understand the code and produce your own novel implementation, then yes. Now, if you “get it from a friend”, illegally, _or_ you just redeploy the code, without creating a transformative work, then there’s a problem. > Just pay the stupid license and if that makes your business unsustainable then it's not much a business is it? In the persona of a business owner, why pay for something that you don’t legally, need to pay for? The question of how copyright applies to LLMs and other AI is still open. They’d be fools to buy licenses before it’s been decided. More importantly, we’re potentially talking about the entire knowledge of humanity being used in training. There’s no-one on earth with that kind of money. Sure, you can just say that the business model doesn’t work, but we’re discussing new technologies that have real benefit to humanity, and it’s not just businesses that are training models this way. Any decision which hinders businesses from developing models with this data will hinder independent researchers 10 fold, so it’s important that we’re careful about what precedent is set in the name of punishing greedy businessmen.
- flutas 2y ago
  > If Perplexity’s source code is downloaded from a public web site or other repository, and you take the time to understand the code and produce your own novel implementation, then yes. Even that can be considered infringement and get you taken to court. It's one of the reasons reading leaked code is considered bad and you hear terms like cleanroom[0] when discussing reproductions of products. [0]: https://en.wikipedia.org/wiki/Clean_room_design https://en.wikipedia.org/wiki/Clean_room_design

Terr_ 2y ago

If the LLM is automatically equivalent to a human doing the same task, that means it's even worse: The companies are guilty of slavery. With children. It also means reworking patent law, which holds that you can't just throw "with a computer" onto something otherwise un-patentable. Clearly, there are other factors to consider, such as scope, intended purpose, outcome...

timeon 2y ago

Is it same as human reading a book? We are not even giving same rights to other mammals. So why should we give it to software.