Comment on Ars Technica Fires Reporter After AI Controversy Involving Fabricated Quotes

nutsack@lemmy.dbzer0.com ⁨2⁩ ⁨months⁩ ago

• rodneylives@lemmy.world ⁨2⁩ ⁨months⁩ ago

  I’ve interacted with Benj Edwards on social media for some time. He’s done lots of good work! He’s on (or maybe used to be on) Mastodon and Bluesky. He runs Vintage Computing and Gaming, and has written good articles for several prominent places. I’ve said as much in multiple forums, I feel like I’ve maybe been going on a crusade.

  I haven’t seen many others defending him. I’m really torn up over this. They had a weak moment. They were sick (I mean, literally.). A few other people, notably Cory Doctorow and Paul Ford, have written LLM-defending places. And the AI hype has been deafening.

  It’s amazing though, that so soon after he used AI, that it immediately hallucinated something job-ending. I knew it was really bad, but I didn’t know it was THAT bad. You get the sense, with so many people talking positively about it, that the hallucinations must be something that happens, what, maybe 5% of the time?

  To me, it seems like the kind of mistake that he should be able to apologize for, promise not to do it again, and move on. But we’ve all had our good will taken advantage of for so long by malicious actors, like how Gamergate was used as a wedge to push loathsome politics onto a legion of young males. It feels like we can’t give anyone the benefit of the doubt any more.

  I don’t know. I know I’m influenced by all the good work he’s done. I feel like that shouldn’t all be thrown away.

  source

  • partofthevoice@lemmy.zip ⁨2⁩ ⁨months⁩ ago

    5% of the time? LLMs, from their own perspective, are only capable of hallucinating. There’s no difference in what they’re doing between cases we call “hallucinating” and “correct.” It’s the same process.

    source
  • mojofrododojo@lemmy.world ⁨2⁩ ⁨months⁩ ago

    Cory Doctorow … have written LLM-defending places.

    citation requested because everything I’ve seen is opposed

    source

    • rodneylives@lemmy.world ⁨2⁩ ⁨months⁩ ago

      Took a bit but found it, it’s not ChatGPT but a small self-hosted AI with an open source model: pluralistic.net/2026/02/19/now-we-are-six/

      source

      • mojofrododojo@lemmy.world ⁨2⁩ ⁨months⁩ ago

        TYVM!

        source
    • rodneylives@lemmy.world ⁨2⁩ ⁨months⁩ ago

      Trying to track it down…

      source

      • mojofrododojo@lemmy.world ⁨2⁩ ⁨months⁩ ago

        TY. Recently finished his book, Enshittification, it’s spot on. good read.

        source
  • nutsack@lemmy.dbzer0.com ⁨2⁩ ⁨months⁩ ago

    why the fuck wouldn’t a journalist double check the things that the AI is feeding to him? in what universe is this even considered journalism?

    source

    • rodneylives@lemmy.world ⁨2⁩ ⁨months⁩ ago

      He was sick and had a weak moment. He didn’t realize that it would just make the quote up.

      source