He makes a really good point that there are lessons to be learned from, "disciplines like urban planning, zoning regulations, crowd control, effective and humane policing, and the simple practices it takes to stage an effective public event".
But I think what he's leaving out is hierarchy. In 2011, I think a "moderator class" and a "commenter class" was all you needed. But if you want people to treat each other well online, you need something more than that. There needs to be stakes. There needs to be value in building up your standing in the community, there needs to be something to be gained and lost in exchange for the community policing itself.
Right now, a community like Twitter has a sort of chaotic hierarchy. There is the average user, the influencer, the blue checkmark, and the Twitter employee/algorithm doing moderation. Who is answerable to who? How does one climb up the hierarchy? What value does a blue checkmark have in moderating an influencer? Can they? Are average users ultimately incentivized to be dicks in order to get the attention of influencers and blue checkmarks?
Maybe it's time for communities as big as Twitter to make some decisions about this and tame their userbase with more rigid and well-defined hierarchies.