Without all the human drivers on the road, computer controlled navigation gets a billion times simpler and safer because all these cars can communicate with each other. They'll then know almost instantly when other cars around them go out of control or have a failure of some sort, and accidents like pileups will be a thing of the past.
Actually, with half decent algorithms all those things can be done even with only some of the car being computerized. Case in point: The Stanford & Google cars*. Google drove their cars for around 200k miles, no accidents**, in real traffic on real roads.
*More info, from the online AI class (skip to 2:30):
http://www.youtube.com/watch?v=kqDvbguZsAA#!
**Technically, there was 1 minor accident in a self driving car, but it was when the autopilot was off and a human driver was in control; they backed into a parked car, which ironically would have been prevented had the car been driving itself.
There was something quite intriguing I realized the other day. Assuming aerospace companies keep going with the space tourism thing, prices come down a little (not even significantly, really), and I were to save a large amount for 5 years to a decade, I could probably go to space at some point in my life. Huh.
Nothing of the space hotel sort unless I really managed to fall into a pot of gold somewhere, but something along the lines of Spaceship Two ($200k) or Lynx ($95k) will probably be low enough in price to actually be feasible. Hell, even at the $200k price, it's still only as expensive as raising a child; an expense most people can manage fairly well, all things considered.