> The end result is a sort of artificial consensus, which has been quite controversial and caused many long time users to leave.
As a long-time user of the site and an occasional commenter (I seem to be sitting at around 1500 comments after 8 years of having an account, which marks me as a bit of a noob), this feels to me like a true but perhaps misleading statement.
MetaFilter demonstrates that intensive, situationally attuned, full-time moderation can be very effective at maintaining a functional community. It also demonstrates that this process can (and almost certainly will) create its own style of drama and discord. Whatever else it is, MetaFilter commenting is a game system with rich, subtle, and frequently unsettling dynamics. It's somewhat driven by what I think of as a correctness ratchet, where locally established positions and norms are both ruthlessly enforced and endlessly subject to critique or reversal (except for certain positions which have attained the status of axiom or common-law moderation practice). As far as I can tell, as many people leave the site because they are on the _leading_ edge of this process (and view the overall state of things as regressive) as because they trail the consensus.
This can all be pretty frustrating. At its worst, it devolves into hyperbolic groupthink, and occasionally edges into the kind of emergent self-satire that you see so much of on, say, tumblrs produced entirely by socially isolated teenagers. I'm currently disinclined to participate in a lot of MeFi threads for the simple reason that I don't see any real gain for anyone in litigating positions contrary to the current mefite correctness state machine.
I don't think any of that serves to undermine the basic point, or demonstrate that the MetaFilter comment culture is a failure. (Participating in MetaFilter, and engaging in good faith with the really difficult aspects of MetaFilter-style discourse, has probably done more for me as a thinker than any other single thing I can point to.)
It rather serves to highlight that no discourse is _without_ its significant problems, or statically assured to remain productive & humane. Humans fail. Systems fail. There are very definitely other kinds of community you might want to build than MetaFilter, but we would be doing a lot better if we built more systems that fail as well as MetaFilter does.