A saner way to do this will be to trust people to do the right thing, but verify that they are not doing anything malicious.
Absolutely. And Mefi's definitely been of a size for a long while now where trust-but-verify has been an essential part of keeping it working. Our approach these days is pretty much the same as Matt's approach years back when it was just him running the place: welcome people in with the benefit of the doubt but keep an eye out for trouble -- whether of the malicious or the clueless variety.
And we see both: keeping spam in check (even after the $5 speedbump) is a daily project, though thankfully not a high-volume one (we probably ban a couple of spammers/linkfarmers/astroturfers/SEO-marketers a week), and we're daily working with new users who are varyingly conspicuously not getting how the place works or what it's for or what the local tone and community expectations are.
One nice thing is that we've been able to take long experience with dealing with spammy behavior in particular and use that to build a pretty decent local toolset to keep a close eye on problem areas with a minimum of effort, so what spam does actually try to manifest these days tends to get spotted and nixed with quick turnaround. Which by itself is good news but doesn't say anything directly about the community, but the implication is important: because we can get on top of one of the major sorts of bad behavior very responsively, it leaves us free to continue to let everybody else on the site have the sort of general community freedom that Matt talks about in those guidelines re: trust. Because our small administrative team can effectively and promptly corral the problems that do occur, we don't have to clamp down on the basic posting and commenting rights of the vast majority of users acting in good faith.