I'm starting to think that a way we can easily filter what's being heavily composed by (SOTA/mainstream) LLMs or not is by how "polite" the public sees their published blog post.

If everyone sees the post as "polite", most of it probably been written by LLMs, as they remove anything that could be seen as "nonpolite" and human. Meanwhile, engineers who just want to publish their own thoughts and feelings on a subject, will be filled with stuff the public sees as "nonpolite", and since those hard edges weren't trimmed before the publishing, we can then assume this is actually a genuine person's thoughts and feelings.

▲

Anon1096 2 hours ago | parent [-]

Your post seems pretty polite. Did you have Claude write it for you?

▲

embedding-shape an hour ago | parent [-]

That'd have been fun, wouldn't it? No, I'm too lazy for that, HN gets my raw and unfiltered disgustingly human thoughts and feelings, unfortunately for all of you.

For shits and giggles, I asked Sol xhigh what it thought about my previous comment, giving it a "6.5/10 for politeness", saying "it’s polite in tone, but somewhat provocative and reductive in substance.".

Maybe this filter should also include provocativeness and reductiveness, and if it isn't provocative and reductive enough, surely it's a LLM? ;)

	▲	postalrat 9 minutes ago \| parent [-]
		To be useful the filter should give a positive on content I personally disprove of.