As a rural resident who does most driving on rough, dirt roads and roads without lane markings, I think it would be cool to eventually have a car that could self-drive, but I don't expect them to work for me anytime soon. Electric cars, similarly, aren't practical for rural usage—there is nothing comparable to the ease of throwing a few more cans of gas into the back to extend your range when driving long distance in a remote area with no fuel/power. As batteries and range keep getting better, they become more useful for some rural users, but still not so much for long distance backcountry travel.
Furthermore, people who choose to live in rural areas end up needing to be more self-sufficient mechanically as well, able to repair vehicles themselves. Electric and self-driving vehicles are and will continue to become decreasingly user-serviceable. That alone is a big reason for me to hesitate in ever wanting one. The world is moving away from repairability, and for people who need to be able to fix things ourselves, or who believe that it's the right thing to do, that forces us to avoid adoption of certain newer technologies. But now I'm getting into a whole new subject.
"The needs of 10s of millions of commuters in America" could have been addressed with public transport years ago. The fact that they weren't should tell us that there's more to it than whether someone drives the car or not.
There was literally a post in one of these threads saying we should "screen" people so undesirables couldn't use public transit so it would be more attractive to yuppy scumbags.
But in rural areas, you'd have to wait too long for a centralized service's car to show up because of lack of economy of scale, and the larger distances it'd have to service. In theory, you could schedule them well ahead of time to arrive when needed, but that's far less convenient and the scale & distance will still significantly affect the price.
For everyone else self-driving will just another trickle-down feature their cars get; like electronic stability control. When the Ford Fiesta has had self-driving features for 3 years(so I can pick it up used) that's when I'll have it. And that's probably about when it will start seeing widespread adoption; when it's available standard in lower-end cars and used lower-end cars.
There are two issues that I think get talked about together, whenever the idea of self-driving cars comes up: self-driving cars themselves, and then self-driving cars as part of an Uber-like pooled vehicle business model.
I think that self driving cars as a technology are definitely on the horizon, and if they don't hit any major roadblocks (literally or figuratively...), some level of self-driving capabilities might become mandatory in the next couple of decades. It makes driving on freeways much safer, and also allows for types of interchanges that aren't possible with human drivers.
But the argument you sometimes hear tossed around, that self-driving cars will lead to Uber-like services displacing all or nearly all private car ownership... that strikes me as much more speculative. There are cultural-psychological aspects to car ownership, and I'm not convinced that the value proposition could be made sufficiently strong to cause people to not want to own their own car, while still creating a viable business for the pooling company, in a suburban or exurban (to say nothing of truly rural) area.
Owning a car in a dense city is a pain in the ass and a huge expense; owning a car in a rural area is basically a necessity to live as an independent adult. It'll be a lot harder to get people in that second group to switch to vehicle pooling than the people who probably detest having a car in the first place for all the expense and inconvenience it causes.