ChatGPT has a style over substance trick that seems to dupe people into thinking it’s smart, researchers found::Developers often prefer ChatGPT’s responses about code to those submitted by humans, despite the bot frequently being wrong, researchers found.

  • @joe@lemmy.world
    link
    fedilink
    English
    8211 months ago

    A caveat: This user analysis involved just 12 programmers being asked to assess if they prefer the responses of ChatGPT or those written by humans on Stack Overflow to 2,000 randomly sampled questions.

    Nothing to see here.

    • @abhibeckert@lemmy.world
      link
      fedilink
      English
      1811 months ago

      What they should have done is asked those same 12 programmers to ask a common everyday question on Stack Overflow and then while waiting for a response, ask ChatGPT the same question.

      I’d bet 50 bucks almost all of them would get an acceptable answer to their question out of ChatGPT 4 in far less time than it takes the moderators at Stack Overflow to delete the question. I can’t imagine any of the questions will actually be answered on SO.

      • @joe@lemmy.world
        link
        fedilink
        English
        811 months ago

        Right. The problem with SO is that you don’t actually get to ask any questions; so reason would suggest anything is at least as good as SO-- even asking a house plant, or Siri, or whatever. Something that actually answers your question would obviously be a better option.

        Stack Overflow brought their irrelevance on themselves, I suspect.

    • 𝒍𝒆𝒎𝒂𝒏𝒏
      link
      English
      1411 months ago

      12 programmers? That ChatGPT sample size is way too small to be meaningful IMO

      • @joe@lemmy.world
        link
        fedilink
        English
        211 months ago

        This is the inevitable result of the decision to fund the internet at large via ads. And there would be (has been) tremendous friction from users when it comes to switching from ad-based to subscription, so we might just be stuck with it.