3 ms·

the way a program works isn't actually how a human would work, and those (false) equivalences, don't make the way a program works to be excusable. anyway, it do

by pxoe 3y ago

the way a program works isn't actually how a human would work, and those (false) equivalences, don't make the way a program works to be excusable. anyway, it doesn't matter, it's just distracting bullshit ultimately. meanwhile, a program operates in a series of very concrete tangible operations over bytes. was data of original works downloaded? yep. was that data processed? yep. was some kind of output based on that data created? yep. so what is that data, if not a derivative? and then, if some other data was created based on that derivative data - damn, that's a derivative of a derivative. is it not?

mym1990 3y ago
Are you pretending to understand how human creativity works? Your last 1-3 sentences is literally what a human does, in some situations, or in many situations early in the development process of a style. If this was not the case, art school would not be the study of previous artists and styles, it would purely be the study of the physical world.
- pxoe 3y ago
  there's "literally what human does (and that's why it's fine for AI to peruse all of your data, don't even question it lol)" (which is bullshit. cute for a sales pitch, but it will always fundamentally be bullshit), and then there are actual ways program work. and these two are pretty much disconnected, no matter how many 'analogies and parallels' are attempted to be drawn. we're not computers, and computers are not humans. if a person were to "do these steps" and "do creativity" in that way - as in, "downloading something" and doing a bunch of quantifiable, traceable operations, as part of their process - well, shit, that's still gonna be just the computer part of it. not 'human creativity' part. still. if there was that 'human creativity' in the loop of 'downloading and perusing a bunch of IP in a traceable way', it's the latter that would be questionable - and useable, for questioning of, for investigation, for lawyers. and unfortunately, computers and their "creativity" work pretty much entirely in that "quantifiable and traceable" way. the 'human creativity' could be chalked up to 'well i don't know'. 'program "creativity"' - very definite 'an executable did this and this, and spit this out'. and if it reproduces well, 'here's a memory snapshot, a complete step-by-step of the process'. can't do that with brains. very doable for software. it could be just a bunch of incomprehensible stuff, but it'd still be a complete byte dump. "it would purely be the study of the physical world" - well, it's still a very much existing option. would it make a difference if that was 'the only option available' or just 'one of the ways'? and AI studies the real world as well, through photographs, and 3d scans, and so on. art is influenced, and those aspects can be pointed out across artists, with some artists readily admitting, 'oh, I was inspired by so-and-so'. people do all kinds of shit with art, and get told off or even sued. art can be forged, with people making something in style and trying to pass it as a work of different artist and sell it. which can be a crime. is that surprising? and AI services try to sell their outputs as well.
  - mym1990 3y ago
    Your posts are very weird, almost all of them are long form, interchanging quote types, quoting at the wrong times, a ramble of ideas...I would not be surprised if this is like a 'snarky AI' model that is producing these.
    - pxoe 3y ago
      not the ad hominem lol
stale2002 3y ago
> was data of original works downloaded? yep. was that data processed? yep. was some kind of output based on that data created? yep. so what is that data, if not a derivative? If I have a program that downloads an image, process the image, and simply returns a yes or no, single byte information, of if the image is blue or not, are you really going to call that derivative work? Because this "is it blue" program does ever single thing that you brought up here. And clearly my "is it blue" program is not derivative work.
- pxoe 3y ago
  program may not be, results could very much be. they are based on something. "but what does the word "derived" mean anyway" lol. i don't know, if you really don't want for those 'results' to be 'derived' from something and be a 'derivative', do a coin flip and just pick that random result.
  - stale2002 3y ago
    > results could very much be. So you are actually going to argue that the one single bit of information, that says "yes or no, is this image blue" is derivative work? Really? Clearly is not. It is clearly the case that if a program outputs if an image is blue or not, with a yes or no answer, that this yes/no answer is not derivative work. You will not lose a lawsuit for outputting if an image is blue.
    - pxoe 3y ago
      do that enough times and it becomes a sizeable database. is it gonna be a derivative work then? at which size? some databases, like ones that point to and describe 'training data', are pretty much just 'saying if an image is blue'. describing what's on that image. is 'one annotation' a derivative work? is a million of them? either way, whatever's the number, it's apparently enough for those databases to put up a license on them, and regulate whether and on what terms are derivatives gonna be made of them.
      - 3y ago

by pxoe 3y ago

mym1990 3y ago

Are you pretending to understand how human creativity works? Your last 1-3 sentences is literally what a human does, in some situations, or in many situations early in the development process of a style. If this was not the case, art school would not be the study of previous artists and styles, it would purely be the study of the physical world.

pxoe 3y ago
there's "literally what human does (and that's why it's fine for AI to peruse all of your data, don't even question it lol)" (which is bullshit. cute for a sales pitch, but it will always fundamentally be bullshit), and then there are actual ways program work. and these two are pretty much disconnected, no matter how many 'analogies and parallels' are attempted to be drawn. we're not computers, and computers are not humans. if a person were to "do these steps" and "do creativity" in that way - as in, "downloading something" and doing a bunch of quantifiable, traceable operations, as part of their process - well, shit, that's still gonna be just the computer part of it. not 'human creativity' part. still. if there was that 'human creativity' in the loop of 'downloading and perusing a bunch of IP in a traceable way', it's the latter that would be questionable - and useable, for questioning of, for investigation, for lawyers. and unfortunately, computers and their "creativity" work pretty much entirely in that "quantifiable and traceable" way. the 'human creativity' could be chalked up to 'well i don't know'. 'program "creativity"' - very definite 'an executable did this and this, and spit this out'. and if it reproduces well, 'here's a memory snapshot, a complete step-by-step of the process'. can't do that with brains. very doable for software. it could be just a bunch of incomprehensible stuff, but it'd still be a complete byte dump. "it would purely be the study of the physical world" - well, it's still a very much existing option. would it make a difference if that was 'the only option available' or just 'one of the ways'? and AI studies the real world as well, through photographs, and 3d scans, and so on. art is influenced, and those aspects can be pointed out across artists, with some artists readily admitting, 'oh, I was inspired by so-and-so'. people do all kinds of shit with art, and get told off or even sued. art can be forged, with people making something in style and trying to pass it as a work of different artist and sell it. which can be a crime. is that surprising? and AI services try to sell their outputs as well.
- mym1990 3y ago
  Your posts are very weird, almost all of them are long form, interchanging quote types, quoting at the wrong times, a ramble of ideas...I would not be surprised if this is like a 'snarky AI' model that is producing these.
  - pxoe 3y ago
    not the ad hominem lol

stale2002 3y ago

> was data of original works downloaded? yep. was that data processed? yep. was some kind of output based on that data created? yep. so what is that data, if not a derivative? If I have a program that downloads an image, process the image, and simply returns a yes or no, single byte information, of if the image is blue or not, are you really going to call that derivative work? Because this "is it blue" program does ever single thing that you brought up here. And clearly my "is it blue" program is not derivative work.

pxoe 3y ago
program may not be, results could very much be. they are based on something. "but what does the word "derived" mean anyway" lol. i don't know, if you really don't want for those 'results' to be 'derived' from something and be a 'derivative', do a coin flip and just pick that random result.
- stale2002 3y ago
  > results could very much be. So you are actually going to argue that the one single bit of information, that says "yes or no, is this image blue" is derivative work? Really? Clearly is not. It is clearly the case that if a program outputs if an image is blue or not, with a yes or no answer, that this yes/no answer is not derivative work. You will not lose a lawsuit for outputting if an image is blue.
  - pxoe 3y ago
    do that enough times and it becomes a sizeable database. is it gonna be a derivative work then? at which size? some databases, like ones that point to and describe 'training data', are pretty much just 'saying if an image is blue'. describing what's on that image. is 'one annotation' a derivative work? is a million of them? either way, whatever's the number, it's apparently enough for those databases to put up a license on them, and regulate whether and on what terms are derivatives gonna be made of them.
    - 3y ago