Similarly the car can take you to some point 3 miles away, you switch to a larger car ("bus"), get within a few miles of your destination where another car takes you to you final destination. The wait times can be very small, and again you'll have a financial incentive (discount) or can pay full whack for a completely private experience. The time and places where there are the highest demands on transport are the ones where self driving vehicles and financial incentives will be most effective.
Deliveries will become way cheaper too. Instead of each place managing their own delivery as currently happens, a delivery company can pick up things from multiple unrelated places. Suddenly there is no need for businesses like fast food to all be in high traffic areas.
Everyone will have the ability to make decisions based on price for each journey. Want to spend more for privacy and luxury? No problem. Happy to spend way less and take a few minutes more? No problem.
I'm sure PricewaterhouseCoopers must have thought about this.
http://www.detroitchamber.com/wp-content/uploads/2012/09/Aut...
I have no idea what they were thinking. What I do know, however, is that there are more than 2.4 million people in the US expected to be at work or school between 8-9am. Using your example, there are few people that can afford $50 to and from work ($100/day - $2,000 per month for those on a 5 day work week - around the cost of leasing a very high end Mercedes).
This isn't remotely close to possible. While it is technically true that cars spend most of their time sitting idle, the demand for transportation is lumpy. A huge percentage of the population works, sleeps, shops, and eats at approximately the same time. That means that with only 2.4 million cars available for use at any one time, the US economy would almost completely shut down. If the majority of those were Uber cars, surge pricing would be in effect most nomral hours....literally only the reviled "1%" could afford to go to work or do anything else from sunrise to sunset.