My mother commented that they used to publish whole *books* of paper abstracts so you could figure out what papers would be worth reading. I wonder if those exist somewhere? I know IEEE has some sort of summary thing they send out now and again.

As far as finding stuff, I get a lot of my favourite papers via recommendations from other academics, and more from being out at conferences. Both of which aren't really viable avenues for discovery unless you're already in the community, sadly.