4 ms·

I've been saying this since it came out... Stable Diffusion is equivalent to hip-hop sampling in the 80s and 90s. The outcome is obvious.

haswell 4y ago
I’ve heard this argument on numerous occasions but I have never heard someone justify it or why they believe it. Are there specific similarities that make you believe these are equivalent scenarios? Not just “it feels thematically similar”.
- shanebellone 4y ago
  It's the closest thing to a precedent. Hip-hop originally recorded and transformed vocals, instruments, and beats to create something new from pieces of something old. The practice occurred without permission and obviously ended up in court. Now sampling requires a licensing agreement. The additional cost has fundamentally changed the genre (over the last 40 years). Hip-hop and tech both ignored IP rights because neither started with a legal framework and both would have found the additional cost prohibitive.
  - haswell 4y ago
    That's helpful. I wasn't aware of the eventual licensing enforcement. If I'm understanding you correctly, you see the similarity more in how the initial side-stepping of copyright eventually gave way to new licensing rules (or adherence to existing rules). I've heard similar sentiment trying to make another point entirely - something closer to arguing that the AI is creatively inspired the way humans are, and therefore is by definition not infringing. I suspect this might be where the flurry of downvotes came from.
    - shanebellone 4y ago
      "you see the similarity more in how the initial side-stepping of copyright eventually gave way to new licensing rules" The systems are similar too despite having completely different internal processes. Both transformed existing IP without permission, to produce sufficiently remixed art as an output. A sufficiently generic abstraction would look very similar despite the disparity of domain. The primary difference between hip-hop and Stable Diffusion is that AI cannot rationalize, explain, or attribute inspiration to a final product. There was no aha moment and thus no creativity. It has no vested interest in its work.
sebzim4500 4y ago
I think there is a pretty big difference though. In the case of sampling you can play the original against the new media and show that they are 'the same'. I have no idea how you would go about doing something similar for Stable Diffusion. "Those three pixels look different when you remove image X from the training set" is probably not a convincing argument to anyone.
- shanebellone 4y ago
  So the litmus test for IP theft is "efficacy of obfuscation"? Remember, input is fundamentally required. Without that dataset, Stable Diffusion delivers exactly nothing.
  - sebzim4500 4y ago
    >Remember, input is fundamentally required. Without that dataset, Stable Diffusion delivers exactly nothing. That's true of lots of things that are legally considered transformative though. E.g. without all those books that Google scanned, Google Books would be completely useless.

haswell 4y ago

I’ve heard this argument on numerous occasions but I have never heard someone justify it or why they believe it. Are there specific similarities that make you believe these are equivalent scenarios? Not just “it feels thematically similar”.

shanebellone 4y ago
It's the closest thing to a precedent. Hip-hop originally recorded and transformed vocals, instruments, and beats to create something new from pieces of something old. The practice occurred without permission and obviously ended up in court. Now sampling requires a licensing agreement. The additional cost has fundamentally changed the genre (over the last 40 years). Hip-hop and tech both ignored IP rights because neither started with a legal framework and both would have found the additional cost prohibitive.
- haswell 4y ago
  That's helpful. I wasn't aware of the eventual licensing enforcement. If I'm understanding you correctly, you see the similarity more in how the initial side-stepping of copyright eventually gave way to new licensing rules (or adherence to existing rules). I've heard similar sentiment trying to make another point entirely - something closer to arguing that the AI is creatively inspired the way humans are, and therefore is by definition not infringing. I suspect this might be where the flurry of downvotes came from.
  - shanebellone 4y ago
    "you see the similarity more in how the initial side-stepping of copyright eventually gave way to new licensing rules" The systems are similar too despite having completely different internal processes. Both transformed existing IP without permission, to produce sufficiently remixed art as an output. A sufficiently generic abstraction would look very similar despite the disparity of domain. The primary difference between hip-hop and Stable Diffusion is that AI cannot rationalize, explain, or attribute inspiration to a final product. There was no aha moment and thus no creativity. It has no vested interest in its work.

sebzim4500 4y ago

I think there is a pretty big difference though. In the case of sampling you can play the original against the new media and show that they are 'the same'. I have no idea how you would go about doing something similar for Stable Diffusion. "Those three pixels look different when you remove image X from the training set" is probably not a convincing argument to anyone.

shanebellone 4y ago
So the litmus test for IP theft is "efficacy of obfuscation"? Remember, input is fundamentally required. Without that dataset, Stable Diffusion delivers exactly nothing.
- sebzim4500 4y ago
  >Remember, input is fundamentally required. Without that dataset, Stable Diffusion delivers exactly nothing. That's true of lots of things that are legally considered transformative though. E.g. without all those books that Google scanned, Google Books would be completely useless.