• danhakimi@kbin.social
    link
    fedilink
    arrow-up
    1
    arrow-down
    3
    ·
    edit-2
    1 year ago

    This was literally part of the 2022 PaLM paper and allegedly the thing that had Hinton quit to go ringing alarm bells and by this year we now have multimodal GPT-4 writing out explanations for visual jokes.

    I’m sure this paper is very funny, but I don’t believe for a second that it successfully explains jokes.

    Just because an ostrich sticks its head in the sand doesn’t mean the world outside the hole doesn’t exist.

    And in case you don’t know what I mean by that, here’s GPT-4 via Bing’s explanation for the phrase immediately above:

    lol, is that what you think jokes are?

    it’s explaining an idiom. that’s all.

    we could do that way before AIM chatbots

    • kromem@lemmy.world
      link
      fedilink
      English
      arrow-up
      3
      ·
      1 year ago

      Tell you what. Come up with a unique joke that isn’t on Google, and let’s see what GPT-4 says as to why it might be funny.

      You seem not to really grok the whole “just because I haven’t seen it it must not exist” thing, and I suppose the easiest way to address it is to just put you directly in front of it in action.

      • danhakimi@kbin.social
        link
        fedilink
        arrow-up
        1
        arrow-down
        2
        ·
        1 year ago

        Why would I bother? Why would I want GPT-4 to attempt to explain a joke to me? I’m an adult.

        If you care, if you think that’s a feature, you can go ahead and ask it to try to explain this, it’s a very simple joke, it wouldn’t be hard for a human without a sense of humor to explain, so certainly a half-cocked algorithm should be able to manage. But don’t bother telling me what happens, I couldn’t care less.