id":48405497,"user":"dang","time":1780612071,"time_ago":"a month ago","type":"comment","content":"

I responded to the GP comment because it was your most recent post, but there have been quite a few other posts that got classified as genai - e.g.:

As I said, it's impossible to know for sure - but we have no choice but to use imperfect measures because (1) the site is getting inundated, and (2) the community feels strongly about it.

In most cases when I've traced through similar cases with specific users, it turned out that they were using some tool (e.g. a grammar checker) that they didn't think of as AI-editing, but which was leaving LLM imprints on their text. So far it doesn't seem possible to run text through LLMs and not have that happen, but we have no idea how that will evolve over time.

For anyone wondering, I suppose I should add that this is in no way anti-AI! LLMs are amazing and we rely on them heavily; we just don't use them on text that we post to HN, since HN conversation is a social activity for humans. There was an interesting PG post about this the other day:

https://x.com/paulg/status/2058844147092488401, https://xcancel.com/paulg/status/2058844147092488401

Although he was writing about emails, not forum comments, the argument is much the same.","comments":[],"comments_count":0,"url":"item?id=48405497"}

Comment by dang