Login
You're viewing the front-end.social public feed.
  • Mar 20, 2026, 6:29 PM

    Sigh. I was just reading a post on #FreeBSD secure levels, as I'm trying to understand them more, at vincentdelft.be/post/post_2025, and in the last paragraph:
    "Would you like me to also add a practical demo section (commands showing...."

    And now I don't know if any of it is believable or not. The rest of the site and #blog doesn't look mass produced, and so I'd be inclined to think it's a real person writing about real experiences, but now I don't know.

    A lot of people apparently have the ability to quickly spot AI #slop produced articles, but I can't seem to notice the tells. At least I found out in the end?

    To be clear, the author hasn't done anything *wrong*, they don't claim it's not written/assisted by AI, so I don't think I can really complain? I guess I just feel a bit disappointed that I thought it would be human communication on a topic of mutual interest, but then it isn't.

    💬 3🔄 1⭐ 1

Replies

  • benbenjamineskola@hachyderm.io
    Mar 20, 2026, 6:33 PM

    @hl I think what bothers me, at least in part, is the carelessness.

    The information might be correct — but the site owner clearly hasn’t taken the time to check whether it is before posting it.

    💬 0🔄 0⭐ 0
  • Mar 20, 2026, 6:45 PM

    @hl To me, it looks lifted from a handbook or guide of some sort. There are plenty of mass-produced pages in the Linux realm describing how to do this or that on Ubuntu<current-version>. The most glaring tell is the absence of idiosyncratic language. This doesn't mean that AI was used, it simply means that one person "wrote" it, and it got distributed as unique content across several sites. This may or may not be the case in your example, just a heads up. The "author" does ask if he should append some examples at the end, so tweet him and find out I guess...

    💬 2🔄 0⭐ 0
  • Mar 20, 2026, 7:05 PM

    @wendigo Yeah, so it could be a guide copied and pasted, or regurgitated. In this case though, the way he/it ends by offering some, or more, examples is exactly how Google Gemini, which we have at work, ends most technical answers - which is what makes me 99% certain it's AI generated (perhaps other bots do that too? Got to keep people spending those tokens).

    💬 1🔄 0⭐ 0
  • Mar 20, 2026, 7:11 PM

    @hl Gotcha. Good to know. I always appreciate manna from dudes who are really passionate about a subject. Hard to find these days. I wish you abundance in this regard. A lack of style is the biggest give-away. As you know, Jake and Chris write a certain way. You do, I do. I always try to be as grammatically structured as I can be while also using the turns of phrases that amuse me and the person I'm talking to... ;>)

    💬 1🔄 0⭐ 0
  • Mar 20, 2026, 7:44 PM

    @wendigo Yeah, no one would pay for a bot to write as badly as I do 😅.

    At work Gmail has rolled out a 'purple underline', which in addition to spelling (red) and grammar (blue) offers to re-write that section for you with AI to make it 'clearer' or 'shorter', and that I hate. Perhaps if English isn't your native language, and for a lot of people at work that's the case, it's a useful additional help. But when I've tried it, it's completely removed my voice from the text. Perhaps it is somehow more readable or compact, but it's not how I wanted to say it.

    💬 1🔄 0⭐ 0
  • Mar 20, 2026, 7:49 PM

    @hl Exactly, buddy! It's funny, people used to say I speak like a textbook, but hey, that's how I prefer to roll... ;>) And that combined with my accent and other well timed phrases, is my voice. I love to hear other people's voices so I can read what they say in that voice. It's an extra dimension to meaning...

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Mar 20, 2026, 8:24 PM

    @hl if they can't even be quality checking by reading through what they intend to post, what does that say about the QA of the actual content of the post?

    Would be a pass for me. There are certainly other sources for that info.

    💬 0🔄 0⭐ 0