[brendoelfrendo]

I suppose it stands to reason: LLMs were trained on human writing, and overuse certain tropes and patterns because those patterns are commonly represented in human writing. But many people aren't particularly adept writers, and they're going to turn to AI to either do their writing or inform how they write. The trope ends up reinforcing itself as people just start to think that AI output is just what normal writing looks like.

[lelanthran]

> overuse certain tropes and patterns because those patterns are commonly represented in human writing.

That's factually untrue. Give me a link to a pre-2020 piece of writing that sounds like an LLM.

[throwaway219450]

The actual prose is ham fisted, but the structural bits - “why it matters” headings and bulleted content are traditional clickbait. Compare to “the results may surprise you” and similar stock phrases that we roll our eyes at. This stuff has always hooked people so it’s unsurprising that RLHF got its way.

If you want examples look at popular DIY magazines from a decade ago, they’re full of this sort of material where every article subheading has a catchy sentence.

[lurkercodemnky]

There's something called RLHF.