• 0 Posts
  • 90 Comments
Joined 6 months ago
cake
Cake day: April 30th, 2024

help-circle

  • daniskarma@lemmy.dbzer0.comtoScience Memes@mander.xyzNobel Prize 2024
    link
    fedilink
    English
    arrow-up
    1
    ·
    edit-2
    10 minutes ago

    Generative AI is part to AI. And it has its own merits. Very big merits. Like or not it is a milestone on the field. That it is mostly hated not because it doesn’t work but because it does.

    If generative AI could not create images the way it does I assure you we wouldn’t have the legion of etsy and patreon painters complaining about it.

    The nobel prize is not to generative AI, of course, it’s about the fathers of the fields and their complex neural networks that made most advanced since then possible.



  • daniskarma@lemmy.dbzer0.comtoScience Memes@mander.xyzNobel Prize 2024
    link
    fedilink
    English
    arrow-up
    65
    arrow-down
    6
    ·
    edit-2
    16 hours ago

    The whole “all AI bad” is disconnected and primitivism.

    John J. Hopfield work is SCIENCE with caps. A decade of investigations during the 80s when computational power couldn’t really do much with their models. And now it has been shown that those models work really good given proper computational power.

    Also not all AI is generative AI that takes money out of fanfic drawers pockets or an useless hallucinating chatbot. Neural networks are commonly used in science as a very useful tool for many tasks. Also image recognition is nowadays practically a solved issue thanks to their research. Proteins folding. Dataset reduction. Fluent text to speech. Speech recognition… AI may be getting more track nowadays because the generative AIs (that also have their own merit, like or not) but there is much more to it.

    As any technological advance there are shitty use cases and good use cases. You cannot condemn a whole tech just for the shitty uses of some greedy capitalists. Well… you can condemn it. But then I will classify you as a primitivist.

    Scientific theory that resulted in practical applications useful to people is why the nobel prize was created to begin with. So it is a well given prize. More so than many others.







  • There’s no way internet traffic hasn’t drastically increased. Also, domain price wasn’t what bubbled…

    And don’t miss my main point. That is that if there’s an “AI bubble” it has nothing to do with AI disappearing, consolidating or even stop increasing if the bubble burst. Same as happened with internet and dotcom bubble.

    This kind of bubbles mean that there’s a bunch of companies overvalued and will disappear once they cannot keep getting investor money without any real return. But it means nothing to the core technology that will continue and keep being developed.









  • Last time I checked they admitted just being a metasearch engine. That their own search index was unusable. Which would be reasonable as a reliable proper search index is incredibly expensive.

    I just checked their website, I don’t find references to their only using their own search engine. In fact they seems to be switched development to AI tools (that I would bet that they are just chatgpt API wrappers).

    They also stated having like 20.000 customers. With that money and based on California there’s no way they have the ability to do any core development besides API wrapping other commercial services.

    Some people asked for source: https://help.kagi.com/kagi/search-details/search-sources.html

    Fron their own web:

    Our search results also include anonymized API calls to all major search result providers worldwide,

    They try to downplay it as their whole bussiness model is “we are not google search” though it’s pretty obvious that they don’t have the resources to create a functional search index of the whole web. Most of their results are filtered metasearch and that can be self host without paying a subscription.


  • This is an incredibly itchy and complicated theme. So I will try not go go really further into it.

    But prosecute what is essentially a work of fiction seems bad.

    This it not even a topic new to the AI. CP has been wildly represented in both written and graphical media. And the consensus in most free countries is not to prosecute those as they are a work of fiction.

    I cannot think why an AI written CP fiction is different from human written CP fiction.

    I suppose “AI big bad” justify it for some. But for me there should be a logical explanation behind if we would began to prosecute works of fiction why some will be prosecuted and why other will not. Specially when the one that’s being prosecuted is just regurgitating the human written stories about CP that are not being prosecuted nowadays.

    I essentially think that a work of fiction should never be prosecuted to begin with, no matter the topic. And I also think that an AI writing about CP is no worse than an actual human doing the same thing.