Hacker News Reader

Comment by anigbrowl | Hacker News Reader

Okay, this is classically nonsense that a lay person like me could never find the time or pleasure understanding.

I guess I'm a lay person who's not like you then.

I've tested one of the LLM's on this problem

Please do not. If any HN reader is interested to get an LLM interpretation, that's well within their capability as is, but unless the topic is itself about LLMs (which this is not) then LLM output makes for utterly uninteresting comments. Without prompt and other technical details it's not helpful, and just creates clutter.

View on HN

> Please do not. If any HN reader is interested to get an LLM interpretation, that's well within their capability

@dang, we hereby propose a guideline amendment: No comments that are entirely or substantially the output of an LLM.

When a user posts an LLM-output comment, consider them as just wanting to share it. It may or may not be interesting. I imagine there are many examples of interesting as well as uninteresting (to HN) LLM outputs. Then, use flagging or downvoting to counteract uninteresting comments specifically. (If users widely rate a comment as uninteresting, it will be successfully flagged/downvoted.)

> When a user posts an LLM-output comment, consider them as just wanting to share it

I don’t think it’s in bad faith. It’s just useless and derailing, analogous to the “don't comment about the voting on comments” rule. If something is routinely downvoted, it’s fair to give folks a head up.

I just meant that I think there are many such comments that won't be downvoted, and I think many such comments are not useless and derailing. Just downvote those that are.

People often get confused about negative feedback without specific details. I understand that people just want to share things, but do that with friends. IT's like commenting with a list of Google search results.

Suppose I had not introduced my comment as being primarily generated by an LLM, how would you have policed this or reacted? And to be clear, my comment is doing quite poorly which seems to show the system is working, but I wonder if the summary, judged exclusively on the merits of its content, would have done as poorly? In the absence of my own personal time and alternate summaries in the comments at the time I posted, I found the two paragraphs from this LLM to be pretty damn good.

And yes, I understand the concern about machines batting around ideas with humans, but it seems like up and down votes are working, admittedly to my disadvantage today.

> Suppose I had not introduced my comment as being primarily generated by an LLM, how would you have policed this or reacted

By calling it out as inaccurate [1].

[1] https://news.ycombinator.com/item?id=36133176

See previous comment by Dang at https://news.ycombinator.com/item?id=33950747

Reply on news.ycombinator.com