Note that online video sites, Craigslist, Flickr, and just about every other social network have mechanisms like this. It can be poorly implemented or need incremental improvement but I think outrage is misplaced.
Flagging can be misused, especially if it's done wrong, and Digg-like behavior ensues. If LJ fails to code this right there will be problems, and possibly very bad problems. Ideally it should work the way Craigslist does, or AIM's "zapping," where one person can't cause a lot of damage.
The tag problem is worse. They should have set an arbitrary limit when tags were launched, but they didn't think about scaling and now they have to do it retroactively and annoy people. I would guess (and hope) that their engineers will find a better way of handling tags and the limit will be raised or removed at some time.
I'm not talking out my ear here. I've been on both ends of this argument since 1991, on a variety of services that serve radically different audiences. The "flag content" system is imperfect and sometimes maddeningly broken, but top-down approaches are a far worse failure. Letting the community flag things based on their own biases and then sorting out the disagreements is the only think I've seen work, at all.
With the current U.S. legal environment I doubt LJ has many alternatives. If someone wanted to check out the LJ code and build one in Belarus or something, it might take off as a refuge from this kind of thing.