Kind of amazing to me how it works that way: once there's graffiti on the wall, suddenly the perception is that it's okay to put graffiti on the wall.
I'm curious if you agree that this is a problem, and what your plans are to address it, if you do.
One tactic for not making it worse is to introduce changes slowly and with lots of time to assess the effect of each change. We're also willing to get procrustean, if we have to, about lopping off any changes that have bad effects. We've done that twice that I can think of.
Any new form of community moderation will come with a mechanism for overseeing it. For example, when we added vouching for users to rescue dead comments, we also added a review mechanism for moderators to catch vouching being abused—e.g. dead comments being rescued when they ought to stay dead because they violate the site guidelines. This combination, which was inspired by the flagging mechanism, turns out to work well. HN's userbase has vastly more capacity to sift through comments than moderators do, and makes good decisions most of the time. Meanwhile it's orders of magnitude less work for moderators to review vouches and flags than to read all the comments directly, and doing so lets us correct the community decisions that don't align with the site guidelines. Such an approach benefits from the strengths of both sides, users and moderators.
I joke that our goal is to turn HN into a pyramid scheme such that the community does all the work and we can just monitor things from time to time, but the real joke is that a 'pyramid scheme' in this case would amount to the holy grail of a self-regulating system. In my wildest dreams it would work à la the Tao Te Ching and simply run itself with the occasional tweak. But I don't expect to get off that easily.
The kind of community moderation we want has to do with protecting the values of the site—intellectual curiosity and civility—rather than particular content people post or views they hold. If you were to argue that those things can't completely be separated, I wouldn't disagree. But I do think we can influence it by how we design the software, and more importantly by how we clarify what the values of the site are. HN is getting to the point where we need a new round of specification about what the site is and what it is not. That's coming soon.
Say something completely wrong, but you are polite about it? No problem, it stays up.
Say something right, but insult those who disagree? Take it down, even if it's right and the others wrong.
Besides, there's also the question of "for whom is this a good idea?" especially in the context of things like Uber and SF property values discussions. It matters which side of the proposition you're on.
Rationality can tell us about likely means to achieve goals and give a probability distribution of outcomes. It can't tell you what to want, what is right, or what to risk.
Really? The majority of academic philosophers disagree with you, and they have very good arguments for moral realism.
One thing to think about: it's not possible to affirm objective standards of rationality and deny objective standards of morality at the same time without being inconsistent, since both epistemology and ethics deal with normative statements (what one ought to do), not fact statements.
The closest thing I can think of in computing theory would be that the statement "A perfect player playing Go will never lose when going first" is obviously accepted as true, despite the fact that creating a perfect Go player is an open problem.
Is this one of those troll comments to which we should respond by saying something nice? Because if it were an American making a similar comment in reverse, then it would be buried as jingoistic/xenaphobic/racist/etc.
You look great today.
On a more serious note, I too have noticed a shift in the comments on this site; I really enjoyed the fact that HN was a more "serious" place for sober discussion, where most comments were useful (larger signal-to-noise ratio). This culture made me reconcider my own comments on a few occasions.
EDIT: Not saying that this is a huge problem - just a development I don't enjoy.
Both of these things put me on edge with americans who, on average, despise passive aggressiveness and use sarcasm as if it were the salt of conversation. The result is that i have to think in a completely different mindset to talk to US people, to avoid misunderstanding them or having them be disgusted by the way i talk.
Being a marketing platform, I would imagine something like the maximum number of people within their target audience engaging.
And two: not having facts/discussions which tarnish their company's image on the platform.
This does kind of make sense; excepting some special moderation view that allocates posts directly to each mod, they're all viewing the forum in the same way, posts bumped to the top or sinking to the bottom, so things will get the same amount of attention or outrage either way. Speaking again from experience, it's all dependent on how long that fire's burning before a mod sees it. A crappy comment deleted quickly causes little fuss, a crappy comment left to fester for a couple hours then deleted will still have the aftermath hanging over it. Even if the entire thread is purged, it still exists in the memories of users and effects the overall tone/behavioural standard of the forum.
1) What is the right number of moderators per comment for HN
2) Do you think that number would generalize to other forums.
Then there are many psychological costs. I'll list three. Often the people whose comments are being moderated do not feel that their comments are bad. They can become harsh, which can push your buttons. With the law of large numbers, any button you have will eventually get pushed sometimes, and when that happens it doesn't feel like a mere statistical process that's afflicting you. It feels more like an evil genius or demon that has perfect knowledge of how to drive you crazy. So you literally can feel driven crazy every day. (Edit: more on 'demons' here: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so....)
Second, an even harder problem is that people want explanations for why their comments are being moderated. It's hard to provide precise explanations in every case. When you get it right, it's helpful—at least to the general audience, and often to the commenter too—but it takes a ton of energy to get right because there is endless variation and nuance in these things, and the cost of being even a little wrong is high: it causes hurt feelings and a general sense of injustice. On the other hand, if you don't provide explanations, people feel that you're being arrogant and heavy-handed.
Third, every mistake you make will be jumped on by lots of people. Their intentions are largely benign, but it's hard to remember that when your mistake is being jumped on. I could continue!
There are different strategies you can take to this, but they all involve tradeoffs. When PG was moderating HN he relied on software to do as much as possible and did the human part with a minimum of (i.e. usually no) explanation. That led to complaints about lack of transparency and so on, but given that he was both building YC and raising small children at the time, there wasn't an alternative. By the time he handed HN over, the community had reached the point of needing a different approach. But no matter what approach one takes, there are these sorts of costs. I think the trick for keeping going is to design the system to have rewards as well as costs, but that's a different conversation.
(Edit from when I ran across this years later: from a 2020 perspective I would say that of the three psychological costs, #2 is the hardest because it consumes by far the most energy. #1 and #3 can be worked on internally, but #2 requires a lot of careful external communication, often under high-pressure conditions. I'm still thinking about what can alleviate that load. It's a bit of a double bind, because if you answer every objection and protest, it's draining—actually it's physically impossible—but if you don't answer, then people will imagine their own answer, invariably a hurtful one, and go away firmly believing it and feeling aggrieved, which is a bad outcome in its own right and often comes back to bite you later.)
As I'm sure you know, it's a more complex problem than just being willing or unwilling to put up with bad behavior. There are costs to addressing it—quite a few, it turns out, no matter what approach one takes.
Our plan is to move HN toward more community self-regulation. Each step we've taken that way—e.g. when we added vouching for dead comments—has worked well. I doubt the community can become completely autonomous (though a mod can dream), but we're pretty sure it can go further in that direction, especially if we do a good job of specifying what kind of site HN is, and isn't, supposed to be.