maybe it happened

  • vrek@programming.dev
    link
    fedilink
    English
    arrow-up
    2
    arrow-down
    6
    ·
    1 day ago

    True, but actually random could be fun. Like yeah you like south east Asian food but tonight… Let’s try Greek!

    • A Sharky Anthro@fedia.io
      link
      fedilink
      arrow-up
      10
      ·
      1 day ago

      If you want true random you need the dice and ditch the LLM; like another commenter pointed out LLMs are pseudorandom with varying degrees of randomness, that can be set by the one who created the tool in question. It’s never going to be as random as basic, unaltered dice can be. A bit of human curated randomness has a lot more warmth and meaning than the coldness of a fancy autocomplete warped in it’s basic function. Which a person who thinks and talks to their partner could put some thoughtful surprises in the mix easily enough. Meh, to each their own.

      • ZDL@lazysoci.al
        link
        fedilink
        arrow-up
        1
        ·
        1 day ago

        The dice aren’t going to be meaningfully random either. The number will be random, sure. But … the choices the dice are rolling against will have to come from somewhere.

      • vrek@programming.dev
        link
        fedilink
        English
        arrow-up
        3
        arrow-down
        4
        ·
        1 day ago

        Agreed it’s not random. But do you really want random? Do you want to be told to go to Gary’s apartment for a manwich? The point is not randomness, it’s being random with a heavy bias towards pretty good. Same idea as a movie wining a award, may not be amazing but probably better than your 5 year old and 3 of his friends making a Sci fi movie

        • A Sharky Anthro@fedia.io
          link
          fedilink
          arrow-up
          6
          ·
          1 day ago

          Ultimately, I am leaning towards human curated activities that dice rolls will decide for the date night in question (just a few so it is achievable). Otherwise, it would be fucking weird and make you seem particularly unsafe to be around. I am also advising against LLM reliance as…Even for seemingly harmless things because the way these models are set up, it’s designed to get people TOO invested in them and validate their worst ideas. Given we’re on Fuck AI (really they are warped LLMs), this should be a sane default recommendation to avoid using those sloppy things.

          • vrek@programming.dev
            link
            fedilink
            English
            arrow-up
            1
            arrow-down
            2
            ·
            1 day ago

            Ok, i mean I see it no different than picking randomly… What made you think I would be “particularly unsafe to be around?”

            I wouldn’t use llm for my job or anything important but what’s the harm in this? Worse case it recommends a restaurant that doesn’t exist, fine laugh and see another local place.

            Same idea as a dice roll but with all options instead of a few

    • fif-t@fedia.io
      link
      fedilink
      arrow-up
      7
      ·
      1 day ago

      LLMs (and most computer applications) don’t operate with true random number generators, mostly pseudorandom generators of varying degrees of strength of randomness.

      Physical dice are more likely to be actually random than most computer stuff. Unless they’re loaded or otherwise biased.

      • degenerate_neutron_matter@fedia.io
        link
        fedilink
        arrow-up
        4
        ·
        1 day ago

        LLM output is also not an unbiased PRNG; it will give you a result based on the probabilities of the next word according to the model, which will likely be heavily biased towards more common values.

        • ZDL@lazysoci.al
          link
          fedilink
          arrow-up
          2
          ·
          6 hours ago

          Heavily biased towards more common values of a particular subculture of a particular culture, in fact.

          Want to have some fun? Ask a culturally weighted question (say about gun ownership) in English. Keep the question as neutral as possible. Watch the pro-gun stance take hold, nine times out of ten. Now ask the same question in a new session translated to idiomatic German. Again make sure the question is as neutral as possible. Compare the answers.

      • vrek@programming.dev
        link
        fedilink
        English
        arrow-up
        2
        arrow-down
        2
        ·
        1 day ago

        Granted they don’t, but can you run a 2x10^500 die? I wouldn’t trust it with my login but choice of dinner tonight…