I'm not yet convinced. The whole self moderation thing is so far incomplete and trivially circumvented with sockpuppets. The moderating tools seem to be merely encouraging a situation where there are (for example) people literally using hubski as a platform on which to incite racial hatred which those who've moderated out those folks won't realize the extent of, but they're nonetheless associated with in the eyes of potential community members who are then going to self select whether they want to be here.
What we have today isn't what we will have tomorrow. mk might want to jump in here, but I would expect that as problems arise there will be discussion and thought about how to deal with specific issues. For example, take a look at anti-spam solutions (I work at an anti-spam company). You can look at the entire user base and see how well they moderate, and whether they are accurate at moderating against obvious spam and other evils, or whether they just follow the herd. You can then weight those users higher, lets call them a canary. They are an early indicator that someone is a fuckwit. You could then use that data to generate a list of "low quality" submitters, that a user can opt into hiding posts from on their profile. "Hide low quality submitters" checkbox. The default "public" not-logged-in feed would show only "high quality submitters". Hubski could also generate a internal list of "high quality" members. These are members who consistently get badged for their content, and are rarely hidden or muted. A user could opt to sort "high quality members" to the top of threads. Again, the not-logged-in experience could have this enabled by default. This is just one idea that I came up with after reading your comment and thinking for a few minutes. mk etc are likely to be able to do better. I'm sure it has obvious flaws that haven't occurred to me yet. If you're not restrained by trying to make money and just care about the quality of the community, you can do things that you ordinarily wouldn't do.