LLMs originally learned these patterns from LinkedIn and the “$1000 for my newsletter” SEO pillions. Both accomplish a goal. Now that's become a loop.
There is a delayed but direct association between RLHF results we see in LLM responses and volume of LinkedIn-spiration generated by humans disrupting ${trend.hireable} from coffee shops and couches.
// from my couch above a coffee shop, disrupting cowork on HN. no avatars. no stories. just skills.md
The titles are giveaways too: Comfort Over Correctness, Consensus As Veto, The Nuance, Responsibility Without Authority, What Changes It. Has that bot taste.
If you want I can compile a list of cases where this doesn't happen. Do you want me to do that?
Neither is Vonnegut's (which your short, choppy sentences reminded me of), but he was a very successful and beloved author. I'm in no way comparing myself to Vonnegut, but my point is just because it doesn't appeal to you, it doesn't mean it isn't good.
Writing is art. Does it get the intended point across? Does it resonate with the reader? Does it make them feel something? Then it is good.
I disagree on Vonnegut. Most human authors at least have a voice, even if you don't like it it's recognisable and theirs, and I would rarely think to criticise that, it makes the writing come alive. If you truly write like an LLM (there is little evidence here of that) it would not be the same.
LLMs serve up a sort of bland pap with sugary highs of excitement which resembles a cross between manic advertising copy and a breathless teenager who's just discovered whatever subject they're talking about. They also sometimes confabulate and generate text which is at best tangential and at worst completely misleading.
It's exhausting and if you haven't carefully read what they generate (which most people clearly have not), you should not expect another human to read it.
Just as an interesting taste, here is my copy above rewritten to sound even more EXCITING and ENGAGING.
"They deliver a horrifying concoction – a sickly sweet, manufactured echo of thought, a grotesque blend of relentless advertising whispers and the manic, unearned enthusiasm of a teenager just discovering a world they don't understand! But the truly chilling thing is this: they fabricate. They weave elaborate lies, constructing text that’s not just tangential, but actively, dangerously misleading!
It’s a psychic assault, a draining vortex of intellectual despair! And if you haven’t wrestled with every single word, dissected it, exposed its flaws – and frankly, I suspect most haven’t – then don’t dare expect anyone else to salvage this wreckage! This is not a passive observation; it’s a desperate plea against a future where genuine thought is suffocated by the cold, sterile logic of a machine! We must guard against this, or we risk losing everything!” -- gemma3:4b
I don't disagree with your take on what how LLM copy is awful; I just disagree that this was written by an LLM. For example, this paragraph at the end:
> If you're in this position (relied upon, validated, powerless), you're not imagining it. And it's not a communication problem. "Just communicate better" is the advice equivalent of "have you tried not being depressed?"
I've seen "you're not imagining it" countless times from LLMs, but always as the leading sentence in the paragraph; for something like the above, they tend to use em-dashes, not parentheses.
FWIW, Grammarly's AI Detector thinks that 17% of it resembles LLM output, and ZeroGPT thinks that 4.5% of it resembles LLM output.
Stopped here. That pattern.
I recognize this pattern from this AI "companion" my mate showed me over Christmas. It told a bunch of crazy stories using this "seize the day" vibe.
It had an animated, anthropomorphized animal avatar. And that animal was an f'ing RACCOON.