I joined the ACM in college. I found the library pretty useful as a "typical developer" and took good advantage of the free access that my employers provided. I paid for ACM library access out of my own pocket when I left for a smaller company, but let it lapse a couple years ago when I realized that most of the papers I was reading were also available elsewhere, and that the journals I was following no longer had as many compelling papers.
(Reading papers on a regular basis is something that I encourage developers and managers to do; it's a good way to expand your toolkit and get ideas for different ways to solve things. A paper a week -- maybe an hour a week, on average -- and that's like fifty new things you've been exposed to in a year. These add up, and eventually can make a big difference in how you approach problems).
yes. depends what you are working on. if I need an on-disk range search structure its pretty irreplaceably valuable to be able to do a literature search and catch up to decades worth of thought on the topic. I'm not going to do nearly as well with just a whiteboard marker and some coffee.
lock free data structures. static analysis. solvers for systems of equations. cache maintenance techniques. concurrency control. etc.
most jobs though don't afford the latitude/scope for you to be looking into details like that in the first place.
(Reading papers on a regular basis is something that I encourage developers and managers to do; it's a good way to expand your toolkit and get ideas for different ways to solve things. A paper a week -- maybe an hour a week, on average -- and that's like fifty new things you've been exposed to in a year. These add up, and eventually can make a big difference in how you approach problems).