Social networks really, really aren't nearly as impactful in the the productive sense. They've mostly made it easier to chat with more people, and browse through more photos - things that are intensely interesting to people, but not really useful.
To your latter point, I can understand that most people use social networks to merely chat and browse drunken photos of friends, but that is like saying that all the invention of the telephone did was make it easier to share mindless banter with your friends. There are fringe benefits that arise from these kinds of technologies that could be really impactful. Twitter's use during the Arab Spring comes to mind as the most recent example of this.
I admit that Facebook doesn't provide any great "technological"/scientific achievement, but it does have a strong network effect supporting it; and that is something that can be impactful in a very productive sense.
To first approximation, humans have a fixed amount of daily attention span.[1] Automated driving will effectively increase the workforce by ~5%. The base rate of "maintenance attention" is already supplied, so much of the 5% will go to the margins, increasing the rate of progress. Since progress compounds exponentially over time, this is huge.
[1]Research I have read on the optimal length of work week, and other research on performance after sustained vigilence.