2023-01-21                         from the editor of ~insom
  ------------------------------------------------------------

  Hi. No one needs to read this but I want to get the thoughts
  out of my head on AI generated art. Feel free to skip.

  I'm specifically talking about Stable Diffusion, which is
  the only thing I've really used, and that I have a local
  copy of, so I'm not limited by cloud computing credits or
  however people are limited on the public / proprietary
  models.

  I'm not like, a _booster_ of this tech -- it's going to take
  another already underpaid field and make it worse, and in
  many cases used a bunch of artists' work uncompensated to
  help to kneecap their field. Pretty dark.

  That said, the technology itself is amazing that it works at
  all. And it's probably not going away now. I feel like we
  can wait out blockchain until it dies off, but as AI art
  generation can be used to make labour cheaper for capital
  it's almost guaranteed to stick around.

  I get that it's fun on social media to joke about how it
  can't count fingers, but the real answer to that is just ...
  scale. I have an RTX3060 video card and I can generate ~200
  images on a text prompt in a couple of hours, and I can
  train a model on around a 1000 images to a reasonable degree
  in about 12. Running things locally changes the dynamics
  entirely:

  Yes, many images are unusably bad, ranging from "a little
  off" to "many mouthed eldritch horror" but when all you've
  done is sink a dollar's worth of electricity into something,
  only one of those 200 images needs to be usable. It turns
  the process into curating and selecting from a large body of
  possible images. Also: retouching and cropping still exist,
  so even if you end up with one of those 6+ fingered hands,
  it doesn't render the whole thing unusable.

  IMO, when the marginal cost of creating things through
  automation drops so close to zero, humans end up getting
  fucked. Self checkouts are bad, but they let one person
  supervise 6 lanes, so companies put up with their badness.
  Google Translate is highly flawed but it's free for most
  uses and real translators are (justifiably) expensive. AI
  images look weird but even cheap illustration is orders of
  magnitude more expensive. People will lower their standards
  -- very far -- for something that's free or cheap. I'm
  guilty of this too -- I've literally ordered two or three of
  the same part from AliExpress in the hope that at least one
  of them will actually be real (and not a fake or a clone) --
  it's still often cheaper than buying one of that same part
  from Digikey.

  On a personal note: as someone who has tried on and off to
  get better at drawing, knowing that this technology exists
  is extremely demotivating. Even if I put in regular practice
  for the rest of my life, it's very likely that someone with
  a dollar worth of GPU time will always be able to create
  something which _looks_ better than I can produce. I used to
  just be comparing myself to artists (which is, itself, a
  negative habit) -- now I'd be comparing myself to _anyone
  with a tool_.