Edit: The post was probably heavily AI written and contains mistakes to that effect, which is unfortunate. The data in general is still interesting though.

    • sircac@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      ·
      edit-2
      9 minutes ago

      For me it is not only that they used AI for the writing, is that they did not care to review/recheck/polishing it before releasing it to the public, so my effort in consuming it will be reciprocal

    • magnue@lemmy.world
      link
      fedilink
      English
      arrow-up
      16
      ·
      16 hours ago

      Oh so you want them to do all that and gather all the data and do it themselves for free? What a dumb comment.

      I’ve run a honeypot for the last month and the data is near-identical to this. It’s definitely credible.

      • MerryJaneDoe@lemmy.world
        link
        fedilink
        English
        arrow-up
        2
        ·
        3 hours ago

        The issue with using AI is that the author doesn’t openly disclose the use at the beginning of the paper.

        Yes, I know this particular write-up isn’t for official submission to an academic journal, but sharing methodology is important.

        I would have no problem with AI-assisted writing IF the author credited the service used and, where applicable, included the prompts used.

        It should be similar to documenting any sourced material. It’s not just about giving credit where credit is due. It’s also about accountability.

        What a dumb comment.

        Why is this necessary? Does this add anything at all to the conversation?

        I’ve run a honeypot for the last month and the data is near-identical to this. It’s definitely credible.

        Ah, well then. Problem solved. Someone on the internet said it’s credible, therefore it must be credible. Tell ya what - when you create a webpage to display your data and then provide an analysis of said data, I’ll consider you credible. Until then, though, you are just some short-tempered, rude, anonymous voice shouting into the void.

        • magnue@lemmy.world
          link
          fedilink
          English
          arrow-up
          2
          ·
          2 hours ago

          I won’t be posting shit. I just submitted IPs and reports to the hosts (surprisingly some of them were using azure/cloud services).

          Kindly, go fuck yourself.

      • floquant@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        38
        ·
        edit-2
        13 hours ago

        Oh so you want them to do all that and gather all the data and do it themselves for free?

        Yes, that is what 90% of the internet has been about since it became a thing. Doing everything for profit turns everything into shit.

      • cecilkorik@piefed.ca
        link
        fedilink
        English
        arrow-up
        53
        ·
        16 hours ago

        Near-identical doesn’t make it valuable. Plausible but incorrect is still incorrect. AI creates plausible and credible but incorrect data.

        The plausibility and credibility is like a honeypot for your confidence. You read it, and understand it, and come to believe it. But it was false all along. You think you learned things. You actually learned nothing.

      • NotSteve_@lemmy.ca
        link
        fedilink
        English
        arrow-up
        6
        ·
        14 hours ago

        I initially disagreed but after actually reading the post, I’m with you. If it was only the article’s text that was generated and not the data or graphs then I don’t see why the whole thing would be written off. I mean, it’s really sad seeing people offload their writing to AI but I still found it interesting.