Hacker News Reader

Comment by joe_the_user | Hacker News Reader

The thing is the "moderation" that's being discussed just bare content filtering, more or less one step past the scripts and neural nets.

Which brings up the point that good moderation is something entirely different. Good moderation involves a person actually emphasizing civil discussion and providing an alternative to argument by shock.

Of course, some fair portion of larger social networking systems today might well be described as "what happens when you remove all moderation" in multiple senses of the term. With that situation actually spilling over into a lot of society with attendant problems (as well as the more invisible problem that real problems get ignored).

Some systems now seem to be aiming for an opposite to this. Kialo is one effort but their approach to debate seems to rigid. A more "personal" approach where the debate is open but assumption of reasonableness is simply implicit would work better imo.

View on HN

But thats the point of the article - that at this scale, there is only industrialized moderation possible.

They describe 3 models by which this can happen, and the type of moderation you describe is classified as "artisanal" or work shop level moderation.

And I guarantee, every system has been tried, and found wanting.

Because every system which gets anywhere, will attract value and then those who prey on those systems and target the weaknesses of those systems.

In essence it’s the difference between what’s done here on HN, and the travesty that is Twitter.

Don't fall into the trap of thinking that HN is good. All societies are biased in various ways, and there are plenty of bigotries and stupidities uttered here on a daily basis.

What you are doing is drawing an us-vs-them distinction. It's not healthy for discourse.

HN is objectively better at content moderation than Twitter is. Not perfect, but better.

"Objectively?" Could you please define your metric and then carry out a measurement, hopefully in a preregistered and reproducible manner? If not, why so confident? Because your tribe, the HN tribe, is objectively better than the Twitter tribe?

HN has guidelines, which are explicitly stated. Twitter (I gather) has guidelines or standards. So you measure content moderation by the number of posts that break the guidelines that are not moderated out.

Your metric is not scale-invariant.

It is trivially gamed by malicious actors. It does not reward higher-quality discourse. It rewards overzealous moderators. It rewards edgy or controversial discourse which does not break the guidelines. It rewards users who contribute a bare minimum of information while punishing users who write longer, more informative or interesting posts.

Worst, by what objective measuring device can a post be compared to the guidelines? There are many extremely-subjective suggestions in HN's guidelines alone:

* "Don't say things you wouldn't say face-to-face."

* "Assume good faith."

* "Don't introduce flamewar topics unless you have something genuinely new to say." (Wanna give us a list of flamewar topics, mods?)

* "Please don't use Hacker News primarily for political or ideological battle." As I write this, there is an ideological battle between JS frameworks on the front page.

* "On HN, users should have an identity that others can relate to."

Could you write a script that measures whether an HN post breaks the guidelines?

> It is trivially gamed by malicious actors. It does not reward higher-quality discourse. It rewards overzealous moderators. It rewards edgy or controversial discourse which does not break the guidelines. It rewards users who contribute a bare minimum of information while punishing users who write longer, more informative or interesting posts.

To the best of my ability to determine, every single thing you said in this paragraph is false. Could you explain how you think some of that works? Because, quite frankly, your comments here look like you just want to complain about HN.

> Could you write a script that measures whether an HN post breaks the guidelines?

No. I could evaluate a bunch of posts, though, and I suspect that my evaluation of that bunch of posts would not be statistically different from others' evaluation of them (assuming fairness and no malice on all the evaluators' part, of course).

It's not automeable, and it's not perfect. But it's also not rocket science.

I would argue that's truly subjective. What's the definition of good content moderation?

Each platform has its purpose.

Reply on news.ycombinator.com