4 ms·

IMO the biggest challenge of self-moderating communities is avoiding the echo chamber effect: ideas surviving and spreading due to their popularity, rather than

IMO the biggest challenge of self-moderating communities is avoiding the echo chamber effect: ideas surviving and spreading due to their popularity, rather than their merit. I'm curious if you agree that this is a problem, and what your plans are to address it, if you do.

Retra 10y ago
I think you're really overestimating the relevance of moderation on that, in lieu of, (for example,) the education level, ages, and number of people and their investments in the community.
- TeMPOraL 10y ago
  Communities run in a tight feedback loop with their populations; when a community goes in any direction, some people who don't like it will leave (and some outsiders who observe it may refuse to join), which pushes a community further into that direction. People within the community like HN are all self-selected, so the education level, ages, etc. depends strongly on the direction the community is going.
niccaluim 10y ago
Indeed, one need look no further than Slashdot to see this in action. 100% community moderation produced a comment section that is all confirmation and no information.
ars 10y ago
You need to moderate based only on civility. And emphatically NOT content (which slashdot does). 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.
- dang 10y ago
  That's a fundamental point and one we can do more to clarify. It's in everyone's interest not to discredit their own argument by going about it uncivilly.
- BaronSamedi 10y ago
  I wonder how good or useful a civility algorithm might be in this case? Aren't sentiment analysis tools pretty good now? Finding examples of rude and uncivil comments on the Internet certainly won't be a problem.
dang 10y ago
I agree that it's a problem, one HN already has. I doubt we can solve it. We can maybe mitigate it, or at least not make it worse. 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.
pjc50 10y ago
How on earth do you propose to separate merit from popularity? 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.
- KaliYugaz 10y ago
  Are you really arguing that there are no objective standards by which we can judge rationality?
  - pjc50 10y ago
    To the extent that there are, that would really limit the scope of discussion. There's no objective morality and no objective politics. Even if we try to limit discussion to "facts", things fall apart fairly rapidly once we start talking stochastic events and statistical morality. 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.
    - KaliYugaz 10y ago
      >There's no objective morality and no objective politics. 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.
      - pjc50 10y ago
        Really? I wasn't aware that morality was a solved problem! Now that we have the one true moral code, why aren't copies being distributed everywhere?
        10y ago

Retra 10y ago

I think you're really overestimating the relevance of moderation on that, in lieu of, (for example,) the education level, ages, and number of people and their investments in the community.

TeMPOraL 10y ago
Communities run in a tight feedback loop with their populations; when a community goes in any direction, some people who don't like it will leave (and some outsiders who observe it may refuse to join), which pushes a community further into that direction. People within the community like HN are all self-selected, so the education level, ages, etc. depends strongly on the direction the community is going.

niccaluim 10y ago

Indeed, one need look no further than Slashdot to see this in action. 100% community moderation produced a comment section that is all confirmation and no information.

ars 10y ago

You need to moderate based only on civility. And emphatically NOT content (which slashdot does). 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.

dang 10y ago
That's a fundamental point and one we can do more to clarify. It's in everyone's interest not to discredit their own argument by going about it uncivilly.
BaronSamedi 10y ago
I wonder how good or useful a civility algorithm might be in this case? Aren't sentiment analysis tools pretty good now? Finding examples of rude and uncivil comments on the Internet certainly won't be a problem.

dang 10y ago

I agree that it's a problem, one HN already has. I doubt we can solve it. We can maybe mitigate it, or at least not make it worse. 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.

pjc50 10y ago

How on earth do you propose to separate merit from popularity? 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.

KaliYugaz 10y ago
Are you really arguing that there are no objective standards by which we can judge rationality?
- pjc50 10y ago
  To the extent that there are, that would really limit the scope of discussion. There's no objective morality and no objective politics. Even if we try to limit discussion to "facts", things fall apart fairly rapidly once we start talking stochastic events and statistical morality. 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.
  - KaliYugaz 10y ago
    >There's no objective morality and no objective politics. 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.
    - pjc50 10y ago
      Really? I wasn't aware that morality was a solved problem! Now that we have the one true moral code, why aren't copies being distributed everywhere?
      - 10y ago