I don't. If I'm working on something and need (or want) the latest cutting edge algorithms then I search for papers in that area as I need it. Otherwise, there's simply too much stuff going on to try reading through everything, or even a filtered down subset. Only a very small portion of it will be remotely relevant to my work or my interests.

If there's a fundamental new result in basic CS or something like that, I figure I'll hear about it on HN or another news site.

I can imagine it's different for people actively working on new research, though.