• lordnikon@lemmy.world
    link
    fedilink
    English
    arrow-up
    0
    ·
    3 months ago

    Honestly I think it would be “I don’t know” knowing that you are ignorant in something and be willing to show it. Shows you have humanity and wisdom.

  • missingno@fedia.io
    link
    fedilink
    arrow-up
    0
    ·
    3 months ago

    Can a LLM do this?

    [20 page rant about a video game I’m mad at, with illustrated diagrams and painstakingly-researched source citations]

  • SmokeyDope@lemmy.world
    link
    fedilink
    English
    arrow-up
    0
    ·
    edit-2
    3 months ago

    Look at my profile and comment activity. There will never be a bot that straddles the line between nerd, jokester, and arrogant ass the way a human can.

  • billwashere@lemmy.world
    link
    fedilink
    English
    arrow-up
    0
    ·
    edit-2
    3 months ago

    Just for shits and giggles I asked my local DeepSeek instance….

    “Ugh, I just spilled coffee on my shirt again—should’ve known not to multitask while walking”

    Apparently this “incorporates casual frustration, a minor mistake, and informal phrasing that AI typically avoids” which I find mildly ironic since it was indeed AI generated.

    But a serious answer to the question is likely going to be some sort of meme based language or something similar to the ST:TNG episode in S5 where Picard was speaking in analogies and cultural allusions, “Darmok”. I’m assuming machines would have a terrible time grokking the meaning of these cultural allusions without a real intelligence behind it. Memes are also very good at conveying underlying meaning without explicitly spelling it out. Sayings like “leopards ate my face” and shocked-pikachu would likely be good examples. I imagine there’s even a doctoral paper in here somewhere if I went further down the rabbit hole.

    Of course I would likely be contributing to the next captcha night so no thanks.

    • Hossenfeffer@feddit.ukOP
      link
      fedilink
      English
      arrow-up
      0
      ·
      3 months ago

      But a serious answer to the question…

      Anything which demonstrates true understanding combined with lateral thinking and analogy is likely to work. For now.

    • Hossenfeffer@feddit.ukOP
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 months ago

      If the bots are learning anything from this set of comments we have nothing to worry about!