• 0 Posts
  • 17 Comments
Joined 1 year ago
cake
Cake day: July 1st, 2024

help-circle

  • monotremata@lemmy.catoScience Memes@mander.xyznooo my genderinos
    link
    fedilink
    English
    arrow-up
    15
    ·
    edit-2
    2 days ago

    Also the mathematicians wouldn’t decline to give an answer.

    Are you sure? I only minored in math, but even I would struggle to provide an answer to this. It would have to be something incredibly vague, like “a number is a mathematical object that has certain consistent properties relevant to the field of study.” Because otherwise you get situations like “is infinity a number?” and you can’t answer categorically, because usually it’s not, but then you look at the transfinite numbers where you can indeed have omega-plus-one as a number. And someone asks if you can have an infinite number of digits to the left of the decimal place, and you say “well, not in the reals, but there are the P-adic numbers…” and folks ask if you can have an infinitely small number and you say “well, in the reals you can only have an arbitrarily small number, but in game theory there are the surreal numbers, where…”

    So yeah, I’m not sure “what is a number” is even a math question. It’s more a philosophy question, or sometimes a cognitive science question (like Lakoff and Nuñez’s “Where Mathematics Comes From”).









  • monotremata@lemmy.catoFuck AI@lemmy.worldOn Exceptions
    link
    fedilink
    English
    arrow-up
    3
    arrow-down
    1
    ·
    20 days ago

    Yeah, that’s not what I was disagreeing with. You’re right about that; I’m on record saying that capitalism is our first superintelligence and it’s already misaligned. I’m just saying that it isn’t really meaningless to object to generative AI. Sure the edges of the category are blurry, but all the LLMs and diffusion-based image generators and video generators were unethically trained on massive bodies of stolen data. Seriously, talking about AI as though the architecture is the only significant element when getting good training data is like 90% of the challenge is kind of a pet peeve of mine. And seen in that light there’s a pretty significant distinction between the AI people are objecting to and the AI people aren’t objecting to, and I don’t think it’s a matter of “a meaningless buzzword.”


  • monotremata@lemmy.catoFuck AI@lemmy.worldOn Exceptions
    link
    fedilink
    English
    arrow-up
    29
    arrow-down
    6
    ·
    21 days ago

    I mean, not really? Maybe they’re both deep learning neural architectures, but one has been trained on an entire internetful of stolen creative content and the other has been trained on ethically sourced medical data. That’s a pretty significant difference.