←back to thread

882 points embedding-shape | 1 comments | | HN request time: 0.207s | source

As various LLMs become more and more popular, so does comments with "I asked Gemini, and Gemini said ....".

While the guidelines were written (and iterated on) during a different time, it seems like it might be time to have a discussion about if those sort of comments should be welcomed on HN or not.

Some examples:

- https://news.ycombinator.com/item?id=46164360

- https://news.ycombinator.com/item?id=46200460

- https://news.ycombinator.com/item?id=46080064

Personally, I'm on HN for the human conversation, and large LLM-generated texts just get in the way of reading real text from real humans (assumed, at least).

What do you think? Should responses that basically boil down to "I asked $LLM about $X, and here is what $LLM said:" be allowed on HN, and the guidelines updated to state that people shouldn't critique it (similar to other guidelines currently), or should a new guideline be added to ask people from refrain from copy-pasting large LLM responses into the comments, or something else completely?

Show context
tptacek ◴[] No.46208058[source]
They already are against the rules here.

https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

(This is a broader restriction than the one you're looking for).

It's important to understand that not all of the rules of HN are on the Guidelines page. We're a common law system; think of the Guidelines as something akin to a constitution. Dan and Tom's moderation comments form the "judicial precedent" of the site; you'll find things in there like "no Internet psychiatric diagnosis" and "not owing $publicfigure anything but owing this community more" and "no nationalist flamewar" and "no hijacking other people's Show HN threads to promote your own thing". None of those are on the Guidelines page either, but they're definitely in the guidelines here.

replies(2): >>46208286 #>>46212157 #
embedding-shape ◴[] No.46212157[source]
Thanks for a lot of references!

One comment stands out to me:

> Whether to add it to the formal guidelines (https://news.ycombinator.com/newsguidelines.html) is a different question, of course. I'm reluctant to do that, partly because it arguably follows from what's there, partly because this is still a pretty fuzzy area that is rapidly evolving, and partly because the community is already handling this issue pretty well.

I guess me raising this question is because it feels maybe slightly off that people can't really know about this unwritten rule until they break it or see someone else break it and people tell them why. It is true that the community seems to handle it with downvotes, but it might not be clear enough why something gets downvoted, people can't see the intent. And it also seems like an inefficient way of communicating community norms, by telling users about them once they've broken them.

Being upfront with what rules and norms to follow, like the guidelines already do for most things, feels more honest and welcoming for others to join in on discussions.

replies(1): >>46212196 #
1. tptacek ◴[] No.46212196[source]
The rules are written; they're just not all in that one document. The balance HN strikes here is something Dan has worked out over a very long time. There's at least two problems with arbitrarily fleshing out the guidelines ("promoting" "case law" to "statutes", as it were):

* First, the guidelines get too large, and then nobody reads them all, which makes the guideline document less useful. Better to keep the guidelines page reduced down to a core of things, especially if those things can be extrapolated to most of the rest of the rules you care about (or most of them plus a bunch of stuff that doesn't come up often enough to need space on that page).

* Second, whatever you write in the guidelines, people will incline to lawyer and bicker about. Writing a guideline implies, at least for some people, that every word is carefully considered and that there's something final about the specific word choices in the guidelines. "Technically correct is the best kind of correct" for a lot of nerds like us.

Perhaps "generated comments" is trending towards a point where it earns a spot in the official guidelines. It sure comes up a lot. The flip side though is that we leave a lot of "enforcement" of the guidelines up to the community, and we have a pretty big problem with commenters randomly accusing people of LLM-authoring things, even when they're clearly (because spelling errors and whatnot) human-authored.

Anyways: like I said, this is pretty well-settled process on HN. I used to spend a lot of time pushing Dan to add things to the guidelines; ultimately, I think the approach they've landed on is better than the one you (and, once, I) favored.