8 comments

[ 0.21 ms ] story [ 14.0 ms ] thread
I think this is going to be a problem for a while, with lots of self-proclaimed expert "LLM detectors" accusing others of using LLMs to write, without any actual evidence other than "I know it when I see it" or some online black box tool that is likely designed to reinforce the user's existing biases.

That said, lots of people do use LLMs to write, sometimes for legit reasons (like English being a second language, or as an assistive device), and sometimes just because they're lazy (or doing some automated BS).

Sounds like OP wasn't banned for writing like an LLM but rather because he doesn't respect the sites rules at all.

> By chance, I found a way to post a blocked domain. It looked like a small bug.

> If you try to open the thread now, you won't be able to unless you have an account. So how do I have screenshots? Well, that's a topic for another article.

> Your story has been removed for breaking the community rules. Self-promotional content is not allowed.

I mean, take a hint.

No. I've been reading your activity on Lobsters and in the Spanish Emacs Telegram group. You were banned in Lobsters primarily because you're a spammer. You don't interact with the community unless it is about something you did or a post you wrote. That behavior is not allowed on Lobsters and I agree with the rule. You said you "found a bug to allow posting" after the first ban. That also goes against the rules.

It's very strange of you because on your blog you're constantly talking about creating communities and then, you disrespect and fail to engage in communities not created by you.

Am confused. He said LLM like writing got him banned and the bottom of this post admits he uses LLM. So did he do the same for comments? And does he assume this is also NOT LLM written when it is clearly used? I have questions. OP should clear this out. You givin mixed signals. :)
All these restrictions on the use of AI are silly; like Torvalds said, submissions should be judged on their quality or merit, nothing else matters.

Eventually this will all sort itself out, but there will be many casualties along the way. And there will be a small group of die-hards, as there always is.

Is it just me or does that site seem to feature an overabundance of articles filled with anime cat girls and anthropomorphic animals for some reason?

Not sure if the invitation-only nature of the site is part of this, but I feel like the quality of comments there is even more gatekeepy and dogmatic than HN even, with more colorfully/arbitrarily interpreted rules.

And I always found it hilarious that they rail on DHH so hard while posting on a server that proudly runs Rails.

Well this very article discloses at the bottom that it is in fact AI written or at least AI-edited.

Perhaps just use your own voice next time?

When AI edits writing it doesn't polish your voice, it erases it.

Over time we mould to the AI voice, we can't help that, but there's still a big difference between human writing that's influenced by AI, and AI edited copy.

From what I can tell:

* 7/8 of your submissions were for your own website or projects (https://lobste.rs/~andros/stories + this deleted one)

* You evaded a domain block ("By chance, I found a way to post a blocked domain") to do this

* The projects and blog-posts, while maybe not "slop", do make use of AI assistance (going by "Claude" being a contributor, and what you'd admitted in the comments)

* You'd recieved warnings from moderators and other users about what you were doing

The ban seems reasonably justified.