I've been doing this on and off, and encouraging some colleagues to do so as well. You can't mass-upload a ton of articles without attracting unwanted attention, but you can throw up a few interesting articles in unsecured directories. People do it accidentally all the time (Google Scholar finds lots of stuff from course wikis and accidentally open directory listings and such), so you could do it "accidentally" as well.
It's also a minor way to have some influence on what gets read and cited. A lot of academics have a favorite classic paper they feel is unjustly underappreciated in their field. Why not put it online? Getting the paper into Citeseer and Google Scholar, where someone might stumble across it, is a small way to help promote underappreciated work. An underappreciated paper nobody can find is going to stay underappreciated!