zlacker

>>feross+(OP)
I'm not sure I buy that the MVP here is actually "viable". Suppose you're Reddit, and you have FatPeopleHate on your site, and you "ban" it, in the sense that you hide it from users who have not opted-in. Does that really provide the same level of enforcement as a true ban? It seems to me that the presence of that community on your site has effects that spread beyond the community itself, it shapes the way people interact even outside the soft-banned subreddit.

I'd be willing to bet that if you could somehow run an experiment in parallel where you had one Reddit with real bans, and one with soft bans, the quality and nature of interactions on the soft ban one would be much, much worse even outside of banned communities.

>>Imnimo+kX1
> It seems to me that the presence of that community on your site has effects that spread beyond the community itself, it shapes the way people interact even outside the soft-banned subreddit.

This is sort of talking around an argument. You could say the same thing about a subreddit dedicated to re-electing a local alderman because of his policy on the maintenance of public parks. Speech is meant to inform, or to affect change.

The question is whether you're going to use an online annoyance argument to moderate controversy on a platform. If the justification for why you're going to moderate speech is that people who are not annoyed by that speech might react to it, you've moved squarely into making "genuine arguments for true censorship: that is, for blocking speech that both sides want to hear."