4 ms·

There are many ways an artist can compensate their influences. Some of them are monetary. When discussing our work, we can name them. When one of our influenc

There are many ways an artist can compensate their influences. Some of them are monetary. When discussing our work, we can name them. When one of our influences comes out with a new body of work, we can gush about it to our own fans. When we find ourselves in a position of authority, we can offer work to our influences. No animation studio is really complete without someone old enough to be a grandfather hanging out helping to teach the new kids the ropes in between doing an amazing job on their own scenes, and maybe putting together a few pitches, for instance. We can draw fan art and send it to them. None of these are mandatory, but artists tend to do this, because we are humans, and we recognize that we exist in a community of other artists, and these all just feel like normal human things to do for your community. And if an artist suddenly starts wholesale swiping another artist's style without crediting them, their peers get angry. [1] 1: https://en.wikipedia.org/wiki/Keith_Giffen#Controversy https://en.wikipedia.org/wiki/Keith_Giffen#Controversy OpenAI isn't gonna tell you that it was going for a Cat & Girl kind of feel in this drawing. OpenAI isn't gonna offer Dorothy Gambrell a job. OpenAI isn't going to tell you that she just came out with a new collection and she's still at the top of her game, and that you should buy it. OpenAI's not going to send her a painting of Cat & Girl that it did for fun. OpenAI isn't going to do anything for her unless the courts force it to, because OpenAI is a corporation who has found a way to make money by strip-mining the stuff people post publicly on the Internet because they want other humans to be able to see it.

csallen 3y ago
Most people know 20,000-40,000 words. Let's call it 30,000. You've learned 99.999% of those 30,000 people from other people. And don't get me started on phrases, cliches, sentence structures, etc. How many of those words do you remember learning? How many can you confidently say you remember the person or the book that taught you the word? 5? 10? Maybe 100? That's how brains work. We ingest vast amounts of information that other people put out into the world. We consume and it incorporate it and start using it on our own work. And we forget where we even got it. My brain works this way. Your brain works this way. Artists' brains work this way. GPT-4 works this way. The idea that a visual artist can somehow recall where they first saw many of the billions of images stored in their brain -- the photos, movies, architecture, paintings, and real-life scenes that play out every second of every day -- is laughable. Almost all of that goes uncredited, and always will. This is what it is to learn.
- Dylan16807 3y ago
  Individual words aren't comparable to the things people are worried about getting copied. People are much more able to tell you where they learned about more sophisticated concepts and styles.
  - csallen 3y ago
    The same principle applies, though. They can tell you maybe a dozen, maybe a few dozen, concepts they've learned and use in their work. But what about the thousands of concepts they use in their work they can't tell you about? The patterns they've noticed, the concepts that don't even have names, but that came from seeing things in the world world that were all created by other people?
    - WereAllMadHere 3y ago
      For example, how many artists drawing street scenes credit the designer at Ford Motors for teaching them what a generic car looks like? How many even know which designers created their mental model of a car?
      - Dylan16807 3y ago
        That is again a single word. There is a strong correlation between how copyrightable a concept is and how well you can point to where you learned it.
- tga_d 3y ago
  I tend to fall more on the "training should be fair use" side than most, but your comment seems to be missing the point. Nobody is arguing that models are violating copyright or social norms around credit simply because they consume this information. Nobody ever argued/argues that the traditional text generation in markov models on your phone's keyboard runs afoul of these issues. The argument being made is that these particular models are now producing content that very clearly does run into these norms in a qualitatively different way. You cannot convincingly make the argument that the countless generated "X, but in the style of Y" images, text, and video going around the internet are exclusively the product of some unknowable mishmash of influences -- there is clearly some internalized structure of "this work has this name" and "these works are associated with this creator". To take it to an extreme, you obviously can't just use one of the available neural net lossless compression algorithms to circumvent copyright law or citation rules (e.g., distributing a local LLM that helpfully displays the entirety of some particular book when you ask it to), you can't just tweak it to make it a little lossy by changing one letter, or a little more lossy than that, etc., while on the other hand, any LLM that performs exactly the same as a markov model would presumably be fine, so there is a line somewhere.

csallen 3y ago

Most people know 20,000-40,000 words. Let's call it 30,000. You've learned 99.999% of those 30,000 people from other people. And don't get me started on phrases, cliches, sentence structures, etc. How many of those words do you remember learning? How many can you confidently say you remember the person or the book that taught you the word? 5? 10? Maybe 100? That's how brains work. We ingest vast amounts of information that other people put out into the world. We consume and it incorporate it and start using it on our own work. And we forget where we even got it. My brain works this way. Your brain works this way. Artists' brains work this way. GPT-4 works this way. The idea that a visual artist can somehow recall where they first saw many of the billions of images stored in their brain -- the photos, movies, architecture, paintings, and real-life scenes that play out every second of every day -- is laughable. Almost all of that goes uncredited, and always will. This is what it is to learn.

Dylan16807 3y ago
Individual words aren't comparable to the things people are worried about getting copied. People are much more able to tell you where they learned about more sophisticated concepts and styles.
- csallen 3y ago
  The same principle applies, though. They can tell you maybe a dozen, maybe a few dozen, concepts they've learned and use in their work. But what about the thousands of concepts they use in their work they can't tell you about? The patterns they've noticed, the concepts that don't even have names, but that came from seeing things in the world world that were all created by other people?
  - WereAllMadHere 3y ago
    For example, how many artists drawing street scenes credit the designer at Ford Motors for teaching them what a generic car looks like? How many even know which designers created their mental model of a car?
    - Dylan16807 3y ago
      That is again a single word. There is a strong correlation between how copyrightable a concept is and how well you can point to where you learned it.
tga_d 3y ago
I tend to fall more on the "training should be fair use" side than most, but your comment seems to be missing the point. Nobody is arguing that models are violating copyright or social norms around credit simply because they consume this information. Nobody ever argued/argues that the traditional text generation in markov models on your phone's keyboard runs afoul of these issues. The argument being made is that these particular models are now producing content that very clearly does run into these norms in a qualitatively different way. You cannot convincingly make the argument that the countless generated "X, but in the style of Y" images, text, and video going around the internet are exclusively the product of some unknowable mishmash of influences -- there is clearly some internalized structure of "this work has this name" and "these works are associated with this creator". To take it to an extreme, you obviously can't just use one of the available neural net lossless compression algorithms to circumvent copyright law or citation rules (e.g., distributing a local LLM that helpfully displays the entirety of some particular book when you ask it to), you can't just tweak it to make it a little lossy by changing one letter, or a little more lossy than that, etc., while on the other hand, any LLM that performs exactly the same as a markov model would presumably be fine, so there is a line somewhere.