My take is that ACM, IEEE and AAAI are all worth maintaining memberships with, especially if you want to stay up to date on what's going on at an academic level. Maybe the biggest advantage of being a member of these orgs, other than access to archived papers, is that subscriptions to their journals is much cheaper if you're a member. YMMV, of course.
On a related note, at least one of IEEE, AAAI or ACM have a discount program for journals published by other outfits, like Springer. Some deal like "join ACM, and your subscription to $FOO from Springer is heavily discounted." If you like to subscribe to journals and what-not, check out all these offers, as they can bring the prices down to where an individual can actually afford them, whereas normally you'd find the price to be so high that only an institution could really justify it.
If the ACM removed the paywall, I would link directly to the proper page for each of my papers - there's excellent meta information on those pages, like who we cite and who cites us. As well as author links which link to other papers written by those authors.
I really want to somehow lobby the ACM and IEEE to remove their paywalls (I am a member of both), but I have neither the time nor the knowledge of how to be most effective.
- http://citeseerx.ist.psu.edu/ has a public "cached" version of many papers that aren't publicly available on other sites.
Also, if there is some paper you just can't find, you could try posting its name (or a link to the paywalled site) here. There lots of academic people on HN. Someone here may be able to help you out.
I think all research should be made freely available. The current situation is just sad.