Comment by dang | Hacker News Reader

Comment by dang | Hacker News Reader

Parent Full thread

Ah, the contrarian dynamic strikes again:

(1) an initial wave of objections to the article;

(2) a second wave of objections to the objections;

(3) those get upvoted, so that

(4) the most popular comment becomes the one about how the site is so negative, all people do is complain, etc., producing

(5) irony!

https://news.ycombinator.com/item?id=24215601

https://news.ycombinator.com/item?id=25434665

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

If you study this phenomenon, the difference between (1) and (2) is not negativity, just timing. And the upvotes of (3) are negative in the same way. Negativity about negativity is not positive. It's an idempotent operation.

(Edit: I hope it's clear that I don't mean to pick on you or anyone else personally! This is a systemic problem that we're all part of—that's kind of the whole point actually.)

If we want a solution to the ambient negativity that can afflict HN threads—which we do—we need to tackle it a little more deeply. We all need to become aware of how the same negativity that we perceive in others exists in ourselves, and without cheap avoidances like "well, the others do it worse". It always seems like others do it worse; everyone experiences that. It is the chief way we avoid looking at ourselves.

If we admit that we're all just mirroring our own denied negativity to each other, we can start taking steps to a solution. Not that we'd never be negative any more—but maybe we can get less mechanical in our responses if we learn something about how the mechanism works in ourselves. Denouncing it in others doesn't work—that's how the problem recreates itself: all this disowned negativity keeps circulating through the system, when what's needed is for people to work with it internally so that it can start to shift a little.

That's why the site guidelines now include this line: "Please don't sneer, including at the rest of the community." – as a baby step in that direction. The HN community has been around for long enough that I think we can take this as a task to work on together. It would be a big step towards optimizing this place for curiosity (https://hn.algolia.com/?dateRange=all&page=0&prefix=true&sor...), which is what we're all here for.

https://news.ycombinator.com/newsguidelines.html

Sometimes I would like to see comments divided in two sections: 1. discussion strictly about the contents of the submission and 2. all the rest, which I would call meta. Meta would be hidden by default and often would be a catch-all for many types of negative comments in type of ("the page is unreadable", "I hate marketing", "I don't know what this is"). I know that many of my comments, like this one, would go to the meta section, but sometimes when I think about it like that I decide not to write a comment. Interesting thing would be to divide karma and weigh content comments by content karma and meta comments by meta karma. The problem would be that a reply to a content comment could be meta, so maybe every thread would have its own section?

One question is how would you categorize them. It could be some combination of software, community input, and moderation, and there are problems with all three.

I would use the word 'generic' rather than 'meta'. Generic includes meta but also all other predictable themes, and a subthread that goes off-topic isn't necessarily bad—whimsical tangents can be interesting when unpredictable, as long as they're not done too often (at which point they'd count as generic).

We downweight generic subthreads and the current certainly counts as that, but I haven't downweighted it in this case because it seemed more important to communicate to the community about this.

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

It is certainly the biggest issue with the idea. I thought about having a checkbox in a reply. Also there could be a side pointing vote arrow. So if you place your comment in generic category you would only have a side- and a down- vote buttons. If your comment gets more side-votes then up-votes it becomes generic (maybe after a threshold). Also people who's comments are too often made "generic" by community or moderators would not be trusted as much on "on-topic" comment votes.

Without introducing more UI/UX it could then work like this: On-topic comments would go first, then side/generic comments and then down-voted comments. Side-vote would be a little like down-vote, but side-votes would count against other side-voted comments.

As always it would go down to many checks and balances. So it it easier to say then do.

Your manual intervention as you describe is often visible and welcome.

Is your downweighting process driven by manual reading through threads or are there signals that draw your attention to potentially-downweight-worthy subthreads?

edit: that is to say, I wouldn't expect user-driven signals to automatically partition or downweight subthreads, I was more curious if they are useful in existing moderation actions and how they might extrapolate.

It's mostly manual. Readers often email us cases that need attention, though, and that's super helpful when the top subthread is a generic black hole. If you see one, and have a minute to alert us at hn@ycombinator.com, we'd be grateful! It's probably the single biggest thing moderation can do to improve thread quality, and we can't read all the threads.

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

That sounds amazing!

That would get the content separated from the noise. I'd love it. The problem with these meta topics is that they are like the bike shed problem--emotionally driven, easy to get sucked into, everyone has an opinion, but ultimately, a distraction that takes away from the important stuff.

Here's an implementation. We have up and down voting. Let me vote these sideways.

This would be a left arrow, which hides them for me--conceptually like kicking them off the screen--and if a quorum develops past a threshold, by default for anyone with "meta" turned off.

I was thinking about side-votes and thought the same before I read your comment.

Well, a lot of what you describe as meta (e.g. "this page is unreadable") is explicitly against the site guidelines, and much of the rest is more-or-less implicitly against the guidelines (ie, HN is for discussions which gratify intellectual curiosity).

The right thing to do is to downvote it. If you can't downvote comments yet, well, they'll usually get downvoted by others soon enough. IMO, the nice thing about HN is that for all that it's imperfect, compared to most of the internet the community moderation (and official moderation) works very well!

There are also comments that are perhaps not exactly related to the story, but are adjacent enough to be on-topic or arise naturally from conversation. It's natural for conversations to move between topics, and many of these discussions are super interesting!

I have gotten so much more information reading the comments than the article.

A lot of the "meta" is useful to me.

I come here for the comments, and like all the comments, even most of the silenced blurry ones.

Sometimes it sounds like a echo chamber, but I like the diversity of this crowd.

I think most of the regulars here realize HN has some strong rules. A bit too strong for myself, so I refrain from saying what I really feel.

I think most of us know the culture at Hacker News. Would I like to see the boundaries expanded--yes, but I understand most of you are afraid of this site becoming Reddit, and so am I.

A bunch of niceties over a article, or controversial subject, is not helpful to me.

(More meta coming up. Does anyone know the exact source code to HN. I've seen the old source code at Git. I'm looking for something current. I want to learn Lisp, and programming a site I like would be nice. Or, is there another language better suited (I guess more libraries) that would be a better fit for a HN clone?)

Same here, but sometimes the meta is overwhelming. Most big technical discussion boards have hit the same issue: the signal/noise ratio falls with popularity, leading to excessive negativity and a lot of meta-discussion to clean up the negativity.

Maybe @dang is right that the negativity actually results from both popularity and timing of comments. That's an intriguing hypothesis. It wouldn't be right to kill the popularity of HN, but maybe there's some way to fix the timing or visibility of comments. I hope that doesn't mean a complicated moderation system. Maybe a new algorithm could help.

Is the solution that no one ever issues legitimate criticism, or that no one ever posts anything worthy of legitimate criticism? Personally, I don't find it ironic or hypocritical when people issue legitimate criticism of anything, including legitimate criticism of the preponderance of illegitimate criticism on this website. These are not two instances of "the same negativity." Issuing legitimate denouncements of people who are issuing illegitimate denouncements is not "the problem recreating itself." This is pretty much a textbook example of the paradox of intolerance.

I would say you're too focused on the legitimate-vs.-illegitimate binary. HN's goal is to be interesting, a.k.a. to gratify curiosity (https://hn.algolia.com/?dateRange=all&page=0&prefix=true&sor...). For that we want comments that are thoughtful, substantive, and above all unpredictable.

Can there be comments which are legitimate but nevertheless also predictable, unthoughtful, unsubstantive? Sure there can. We don't need those here. They don't affect threads in a good way.

Can there be comments which are illegitimate, but nevertheless manage to be thoughtful/substantive/unpredictable? Surely this must also be possible. Such comments could well make for interesting conversation, in which case they'd be fine.

There's no rule requiring commenters to be right (which is handy, since to a first approximation we're all wrong about everything and rarely have any idea what we're talking about—this is what they used to call "the human condition"). When people approach conversation with a strong need to be right, that doesn't lead to curious conversation, it leads to the situation in the famous xkcd classic.

I'm not saying that the truth doesn't matter and certainly not that people should just bullshit. Bullshit is a lot less interesting and more predictable than it assumes it is. I'm saying that an open-minded, exploratory, conversational way of looking for the truth yields best results here, and trying to categorize things as legitimate vs. illegitimate isn't in that spirit. As for "legitimate denouncements", god help us, that covers every denouncement ever made, as far as the denouncer was concerned.

Small observation on this great comment :)

I noticed on twitter that the culture (at least in my bubble) valued „smartness“ over empathy. In attempt to say something „smart“ it lead to many snarky responses.

I also noticed this pattern in myself, which is why I stopped replying to tweets. (It was easier for me to stop replying than stopping to reply in a snarky way).

For twitter I think this is not only a question of community, but also a question of design (ui/ux). And how the ui/ux shapes the community.

If I ran HN, I would consider if there are any (small) UX changes in the comment flow that could improve the „well, actually“/contrarian metric.

Do you have any ideas for how that could be done on HN? What about Twitter UX in particular do you think leads to that dynamic?

Seems some debouncing is required :)

The mmm.pages post yesterday was really refreshing positive, until a comment floating to the top ranted about geocities style/gifs being schizophrenic, nostalgia being toxic, and real creativity = having the same cookie cutter modern web design as everyone else (now that’s irony).

But based on your comment, you either: a) were inspired by the post itself and decided to have a little fun roasting this comment; or b) need to take a vacation.

or c) dang just has their finger on the pulse and has seen this every day for years.

Reply on news.ycombinator.com