It allows people to read only the things they are interested in; essentially it could provide multiple topic-based discussion areas.
I don't think one can look at how tags have previously been used on LW to answer "why would accurate tags be useful?", since the tagging situation is pretty horrible. On StackOverflow the tags can be edited (and are, all the time) by "high"[1] reputation user and they are used to filter content extensively. However the infrastructure there is a little different, since users can select "Favorite Tags" and "Ignored Tags" to control what shows up on their front page.
(Granted that the comparison isn't necessarily a good one, since StackOverflow is much higher volume, and it has a slightly different purpose.)
[1]: 500 reputation, but reputation is basically received 5 or 10 per vote (depending on the situation), so it is actually quite low. But the possible tags come from a finite set, and one needs 1500 reputation to add a new tag to this set.
The recent implementation of a -5 karma penalty for replying to comments that are at -3 or below has clearly met with some disagreement and controversy. See http://lesswrong.com/r/discussion/lw/eb9/meta_karma_for_last_30_days/7aon . However, at the same time, it seems that Eliezer's observation that trolling and related problems have over time gotten worse here may be correct. It may be that this an inevitable consequence of growth, but it may be that it can be handled or reduced with some solution or set of solutions. I'm starting this discussion thread for people to propose possible solutions. To minimize anchoring bias and related problems, I'm not going to include my ideas in this header but in a comment below. People should think about the problem before reading proposed solutions (again to minimize anchoring issues).