Ask HN: How do you auto-moderate FUD? | Hacker News Reader

Ask HN: How do you auto-moderate FUD?

9 pointsomniglottal14 comments

I've seen an increase in bad argumentation styles, with faulty/poor rhetoric and persistent what-about-ism here lately. Stuff like a purely cynical, hyperbolically dystopic, or selfish and nihilistic perception which fails to align with the moral community I previously witnessed on this site. I.e., in a conversation about the US or western democracies in general, there inevitably will be a new account asking simple, distracting questions for which the most appropriate gr retor is often something like "but, in this context, we actually don't want authoritarian governance..."

Engaging with these accounts shows that they are unwilling (or else not employed to) engage in good-faith dialogue. Although this guideline applies to all, there's only one "dang" and he can only do so much.

My question: are there tools/plugins/front-ends which you use to, i.e., share lists of known "shill" accounts to filter stuff like foreign state actors out of an otherwise valuable HN thread?

HN has a problem that it rejects "shaming" in the community.

I think you can shame people in a way that isn't completely abusive, and even instructive.

Shame exists for a real reason in communities - but it only works if the shamed (and everyone else) understands why they were shamed. This takes more communication than a downvote ...

The problems you want to address are real. I also have seen the drift in HN over the years, as zealots start to outnumber (or at least out-post) those who will actually think.

But I suspect that tools aren't good enough to reliably work. (Do you want this post, say, marked as FUD because of a negative take on HN posters?)

The best answer I have is to parallelize dang. Downvote it when you see it, and don't feed the trolls (that just gives them a chance to make another comment). If it's particularly bad, downvote and flag. (I'm not sure what the threshold is to kill a comment. It's more than one flag, but it's not that many. I suspect it's in the low single digits. If very many people see the problem, and will bother to flag it, bad comments will quickly die.)

But there are at least two problems with my solution. First, it hasn't been working so far, or at least it hasn't been working well enough. It working better requires a change in behavior (and more effort) from the established users, which is kind of a tough ask. Second, this is something that could be turned against us if a block of zealots decided to mass-downvote and mass-flag things. I mean, we can vouch for dead comments, but that's also work...

I get the impression that most people downvote a bad (or bad-faith) comment, flag a nasty one. I find that one time in 4-5, my flag seems to kill the comment, so I guess it takes around that many.

There's big problems with both downvoting and flagging: you don't know WHY people are doing it. Without that, you can't correct behavior.

I'm regularly downvoted and I know exactly why: because I'm mocking Rust's curl | sh or Python's appalling packaging or whatever, no-one is going to "correct" that behaviour and it doesn't matter anyway; it's just HN karma.

Is that "why" though?

The simplest correction that might work is don’t do what was downvoted or flagged, again.

Right. But there's a difference between that thing and why it was downvoted.

It could have been downvoted because it is factually incorrect, which means you should be careful about posting incorrect information.

It could have been downvoted because it's inflammatory.

It could have been downvoted because it is redundant (doesn't add to conversation).

Sometimes it's not clear, and you only can speculate.

Inflammatory is self evident.

Non-contributory is self evident.

I didn't say it wasn't obvious in ALL CASES. And even in those cases there are language and communication issues that make it not obvious.

Someone who doesn't speak English for example may use a tone that comes across as inflammatory when they really are just trying to lightheartedly say "wtf mate" which is sometimes a perfectly fine response.

Do you have the means to hire a secretary that can give you a run-down of the latest hacker news at your convenience and censor out the posters you don't want to hear? That's probably your fastest solution.

Your filters won't work against the new accounts that are plaguing you, and an AI will never be better than people at detecting sarcasm, or be able to sniff out unpleasant states of mind such as nihilism, selfishness etc.

You could create a community of hackers to together censor and weed out questions and comments that aren't kosher, but that group will soon disintegrate with in-fighting, as you will all personally have your strict opinion of where to draw the line of what is acceptable or unacceptable.

So I say go with a smart secretary. He or she will do the job as you demand, and since you pay there won't be any shenanigans or all night debates on what question is honest or bait.

The auto-mod is flag and downvote. However the community has a particular bias.

Only a mod can overturn this bias. I've seen it a few times, e.g. people celebrating Queen Elizabeth's death and another where everyone flagged some legit advice from a female founder because "female" sounded like clickbait and sexism. And in these times the action taken was controversial because it ran against community expectations.

Some of the problems you list runs further into the gray area, and I don't think there's much that can be done.

idk dude, sounds cringe, maybe even cope.

Just use HN for what it is: a stage for performative erudition, and erudite shitposting. Just have fun and don't take it so seriously.

in a conversation about the US or western democracies in general

In cases like that, self-moderation works best for me.

I know from experience the comments will tend towards a dumpster fire.

And I am flammable so I stay away to the extent I can by reminding myself there will be another chance to get upset about politics tomorrow.

https://xkcd.com/386/

Reply on news.ycombinator.com