It's a nice sentiment, but thanks to the Open Access movement, most recently published papers are already available online somewhere, and the people talking about #pdftribute are the people most likely to have
already shared their papers. Important historical papers, even those that are now public domain, are still behind paywalls. As Aaron said, "even under the best scenarios, their work will only apply to things published in the future. Everything up until now will have been lost." [1]
Sharing your own papers is nice, but it's also safe. It's not really challenging the status quo.
[1] http://pastebin.com/raw.php?i=cefxMVAy
Edit: I overstated my point. I don't know that "most" papers are already available; it certainly varies a lot by field. I kind of doubt that many #pdftributers are people that weren't previously sharing their papers, though.