Hacker News new | ask | show | jobs
by svnt 2 days ago
It is mostly written by llm. “narrower” and “I want to put a fence here” hedging, etc. This is very 4.8. Maybe llm that has been somewhat massaged by a human to sound less ai.
4 comments

I wonder if you may be seeing ghosts? At least to me, this sounded so clearly like an authentic human voice, at least the parts I've read (haven't finished yet).

This strikes me as the good writing that LLMs very poorly try to model (or have been forced into through brutal fine-tuning), and I think we should be cautious not to miss the distinction.

I don't suppose you're someone who tends to dislike metaphorical flourish and narrative elements in articles even before all this? I ask, because I've been wondering lately whether people who like clear information-based writing might have a less developed pallete for writing styles, and "humans writing with flourish" might kinda blend with "LLMs writing"..?

It has very human aspects, such as the beginning. So people switch off. And then it has long stretches where it is bulked up by claude and opus-4.8’s obsession with “honestly true,” “narrower” claims, how concepts “rhyme” etc.

I guess it is also possible this person has internalized claude, but I think their writing pattern is: short pieces: fully human voice; long pieces: ai-supported.

As to my personal views, I am sad to have lost the emdash and the antithesis, among other things, to the llm-cliche dustbin.

It's possible we're at the point now where it fools me, but I didn't see it that way. I think more evidence against would be the fact that the author discloses genAI usage in another article [0] and provides their own version of the same [1].

[0] https://wilsoniumite.com/2026/06/14/labor-pressures-causing-...

[1] https://wilsoniumite.com/2026/06/07/labor-is-a-market-distor...

The one they claim is in their voice [1] has zero of the tells I am referring to. You can compare and contrast it with your fully/generated reference 0 (although that is fable and I have little experience reading its writing it seems to use many similar comventions) and the title piece and see what I mean.

[1] https://wilsoniumite.com/2026/06/07/labor-is-a-market-distor...

you are so wrong. this is not ai.
Parts of it seem not to be, but the bulk of it is. Here is a particularly clear example of opus-4.8-speak.

> Now, I want to be careful here, because this is the part where it would be very easy to start waving my arms around. Every single one of these has many causes. Suburbs and cars. Television, and then phones. A long list of things that have nothing to do with me at all. I am not going to claim I have found the one secret root of loneliness, and you should be suspicious of anyone who does. We cannot cleanly untangle these. That is just honestly true.

Watching trust in online content of any kind disintegrate in realtime due to AI in a forum that on balance breathlessly touts AI is surreal.
You're conflating different things here. You would have to look very hard on HN to find someone who "breathlessly touts" the idea of publishing AI-generated blogposts under your own name.
In for a penny, in for a pound. "Well it's ok when I do it" doesn't wash.
i wish i could codify voice.

probably someone has.

i’m not beyond being fooled, but when there’s voice there’s a human.

this is especially true with certain types of awkward phrasings that LLMs love to correct but humans don’t.

one day my internal filter will fail, but not today.

Right, this is cyborg text. The opening is in human voice. I am acknowledging that. Then it ebbs and flows over llm rocks.

The piece would be maybe 6-8 grafs without claude, and much of claude is papered over.

> Now, I want to be careful here, because this is the part where it would be very easy to start waving my arms around.

Very strong LLM signal there. I don't mind people using LLM in their writing, but when there are LLMisms like that in the text, it takes away from the reading experience in multiple ways. Firstly, it screams out LLM use and changes the reader's focus from the content to the content creation. Secondly, it's just bad writing that reduces reading enjoyment. I'm looking forward to improvements that eliminate these obvious problems.

How did LLMs end up doing this anyway? I wasn't seeing this kind of thing before LLMs. Was there a large corpus of training material with this kind of thing is common?

That’s not really an LLMism. It’s a phrase that ordinary writers use and was perfectly fine, but LLMs started overusing it, so now you see it as a “tell.” People who haven’t read enough LLM-generated writing to see the pattern won’t notice anything wrong.