I think classifying it as an allergy or a status thing is a little too glib. I’ve read and reviewed, conservatively, hundreds of AI generated documents for work, and “written”/commissioned a bunch too. My biggest issue is that it’s impossible to engage with and give feedback on an AI written document, because it’s impossible to know whether misconceptions or gaps are because the author doesn’t understand the material deeply enough or the author does but the AI doesn’t and the author’s not proofreading carefully enough. Or if a surprising idea is raised — is it the authors insight, can they elaborate on it, where did it come from, etc?
Hackernews isn’t work, obviously, but “it’s impossible to engage deeper with the material because the author doesn’t really exist” is sort of a problem for a discussion site. If the human coauthor puts in enough work, they can make sure the doc really reflects their views and their understanding, but in my experience that’s much less common.
You can engage with an AI-written document! Simply paste the document into the AI (same or different as the AI that wrote it) and ask questions. You can get an answer in seconds compared to an hour (if it's a blog post by an HN reader), days (if it's a professional blogger), or never (if it's by a journalist).
Perhaps AI-written articles should include the prompt as hidden text on the page. Or include a link with the prompt embedded, that causes the AI to spit out the same article that you could then ask questions about.
>My biggest issue is that it’s impossible to engage with and give feedback on an AI written document, because it’s impossible to know whether misconceptions or gaps are because the author doesn’t understand the material deeply enough or the author does but the AI doesn’t and the author’s not proofreading carefully enough.
FWIW, same problem with PRs or PoC that I have to work on; now my first question is, "did you know about his behaviour?". The first step, getting a decent spec, is delayed to after a first draft implementation is already pushed...
This exactly. Whenever someone shows me their (or someone else's) AI-assisted-coded (vibe coded) thing, I can only ever engage with it on a very superficial level, saying "oh cool".
You can't have a discussion about it, how it was done, because for 80% of cases it's "the AI did it".
I don't think this is that interesting or useful for HN, because indeed it's a discussion site.
It would be different if the vibe coded thing didn't come as a "look what it made and its vibe written README" but instead as a human written blog article about "how I made this thing using AI, what models, prompts and harnesses, how the experience was etc etc etc" --- again that would be something interesting to have a discussion about.
But otherwise what is there to say except "yeah cool that's cool that the AI made you that thing"
> it’s impossible to know whether misconceptions or gaps are because the author doesn’t understand the material deeply enough or the author does but the AI doesn’t and the author’s not proofreading carefully enough. Or if a surprising idea is raised — is it the authors insight, can they elaborate on it, where did it come from, etc?
That's interesting, thank you! My only pushback is that my bit about "allergy" and "status" is about the audience response here, whereas you seem to be talking on a somewhat different level.
Oh for sure. My only point is that IME (and I’d assume many other commenters) the quality of discussion and engagement around an imperfect ai-written doc is considerably worse than the same doc written by a human. Work is where it’s most obvious to me… because many of these are design docs for systems I care deeply about us delivering, but I’ve noticed the same thing elsewhere.
But you wouldn't have called it "allergy" or a "status" problem when considering the audience's response to human written posts of a similar low quality level.
It's always been the case that the HN audience has been "allergic" to low effort slop (including human generated).
The OP (and my) point was trying to point out how the AI generated stuff might masquerade as something "interesting" if it had been human written, but actually is much of a nothing burger when there's no human behind it:
if a human made a strange, weird, or even sub-optimal design decision, that's something to possibly have an interesting discussion about.
if an AI does the same, first of, I don't care and don't wanna guess why its weights did the thing it did, it's just not as interesting as human motivations, UNLESS the main meat of the article/link is in fact about (again the human experience of) "how I made this thing, using AI, what prompts, harness, models, etc".
I hope it makes sense and allows you to no longer dismiss this as an "allergy" (aka the audience's problem), but as something that is actually a quality problem with the post.