• 0 Posts
  • 12 Comments
Joined 1 year ago
cake
Cake day: June 22nd, 2023

help-circle







  • AEsheron@lemmy.worldto196@lemmy.blahaj.zoneAI rule
    link
    fedilink
    arrow-up
    2
    ·
    9 months ago

    Because one is a black box that very well may just be a much more advanced version of what current AI does. We don’t know yet. It’s possible that with the training of trillions of trillions of moments of experience a person has that an AI may be comparable.

    I mean, the likelihood is basically zero, but it’s impossible to prove the negative. At the end of the day, our brains are just messy turing machines with a lot of built in shortcuts. The only things that set us apart is how much more complicated they are, and how much more training data we provide it. Unless we can crack consciousness, it’s very possible some day in the future we will build an incredibly rudimentary AGI without even realizing that it works the same way we do, scaled down. But without truly knowing how our own brain works fully, how can we begin to claim it does or doesn’t work like something else?


  • AEsheron@lemmy.worldto196@lemmy.blahaj.zoneAI rule
    link
    fedilink
    arrow-up
    3
    ·
    9 months ago

    That is a fundamental misunderstanding of how AI works. It does not shred the art and recreate things with the pieces. It doesn’t even store the art in the algorithm. One of the biggest methods right now is basically taking an image of purely random pixels. You show it a piece of art with a whole lot of tags attached. It then semi-randomly changes pixel colors until it matches the training image. That set of instructions is associated with the tags, and the two are combined into a series of tiny weights that the randomizer uses. Then the next image modifies the weights. Then the next, then the next. It’s all just teeny tiny modifications to random number generation. Even if you trained an AI on only a single image, it would be almost impossible for it to produce it again perfectly because each generation starts with a truly (as truly as a computer can get, an unweighted) random image of pixels. Even if you force fed it the same starting image of noise that it trained on, it is still only weighting random numbers and still probably won’t create the original art, though it may be more or less undistinguishable at a glance.

    AI is just another tool. Like many digital art tools before it, it has been maligned from the start. But the truth is what it produces is the issue, not how. Stealing others’ art by manually reproducing it or using AI is just as bad. Using art you’re familiar with to inspire your own creation, or using an AI trained on known art to make your own creation, should be fine.


  • AEsheron@lemmy.worldto196@lemmy.blahaj.zoneAI rule
    link
    fedilink
    arrow-up
    7
    ·
    9 months ago

    Do you actually think artists using AI tools just type shit into the input and output decent art? It’s still just a new, stronger digital tool. Many previous tools have been demonized, claiming they trivialize the work and people who used them were called hacks and lazy. Over time they get normalized.

    And as far as training data being considered stealing IP, I don’t buy it. I don’t think anyone who’s actually looked into what the training process is and understands it properly would either. For IP concerns, the output should be the only meaningful measure. It’s just as shitty to copy art manually as it is to copy it with AI. Just because an AI used an art piece in training doesn’t mean it infringed until someone tries to use it to copy it. Which, agreed, is a super shitty thing to do. But again, it’s a tool, how it’s used is more important than how it’s made.