Tons of people on the internet write in this style. I did, frequently, well before ChatGPT.
That’s where LLMs got the habit from, then inbred it into each other. Dare I say it may have bled back into human writing too.
Personally, I’d want to block “is this or is this not AI” threads on Lemmy. I’m sick of people thinking they’re human LLM/diffusion detectors when that is (outside of obvious cases, or logit watermarks) physically impossible.
I’m not being condescending at all and your comment makes perfect sense in isolation. It’s stupid to call out a single construction (or punctuation mark) as you’re interpreting me as having done. I’m just assuming you’re smart enough to see a pattern when you see it repeated often enough; so sorry, I guess?
Same with some replies in my inbox. I see a lot of semicolons, “not X but Y” formats, OpenAI purple prose, stuff I might suspect as slop in specific contexts, but I’m not pointing fingers over it unless it seems like they are actually trying to spam something. There’s no way I can tell if an LLM was used with reasonable certainty.
Perhaps I’m unreasonably agitated by this. But I don’t like the idea of a “purity test” for writing style; the Tech Bros that are forcing that paranoia on us don’t deserve one ounce of our attention.
Tons of people on the internet write in this style. I did, frequently, well before ChatGPT.
That’s where LLMs got the habit from, then inbred it into each other. Dare I say it may have bled back into human writing too.
Personally, I’d want to block “is this or is this not AI” threads on Lemmy. I’m sick of people thinking they’re human LLM/diffusion detectors when that is (outside of obvious cases, or logit watermarks) physically impossible.
I’m not screaming about an em-dash. Look at the first comment they make on their posts and come back and tell me what you think.
I also dislike this sort of condescending tone on forums. I’m guilty of it too, but it never makes me want to engage.
I’m not being condescending at all and your comment makes perfect sense in isolation. It’s stupid to call out a single construction (or punctuation mark) as you’re interpreting me as having done. I’m just assuming you’re smart enough to see a pattern when you see it repeated often enough; so sorry, I guess?
I wasn’t involved here, got curious, did look and… Yeah it’s odd. An awful lot of commonality.
So is there in my history.
Same with some replies in my inbox. I see a lot of semicolons, “not X but Y” formats, OpenAI purple prose, stuff I might suspect as slop in specific contexts, but I’m not pointing fingers over it unless it seems like they are actually trying to spam something. There’s no way I can tell if an LLM was used with reasonable certainty.
Perhaps I’m unreasonably agitated by this. But I don’t like the idea of a “purity test” for writing style; the Tech Bros that are forcing that paranoia on us don’t deserve one ounce of our attention.