It will likely be the same way with the self-driving cars. There are several companies (Apple being one of them) that are investing billions in this. They would not be doing so if they were not absolutely sure this is where we are heading.
The technology of self-driving cars today is probably as primitive as cell phones were 20 years ago. We will likely be astounded to see the technology in this space that will be soon brought to bear.
By the way, I don't envision a scenario where self-driving cars are just thrown out there on the roadways with regular cars at the beginning. I see lanes on highways dedicated to the self-driving cars with maybe jersey barriers separating them from the rest of the cars. Drivers will still need to take over to merge with the regular traffic. Over time, probably within the space of 10 years, it will be deemed safe to co-mingle the two and then the regular cars will quickly get squeezed out of the picture entirely.
You may be right about the process for adopting automated-vehicle technology, since that would be the most feasible and practical way to do it. This is one area of automated technology that could probably be implemented today, if the infrastructure was in place to deal with it. In theory, you could drive from the center of New York City to the center of Chicago in a car that operates most of the distance in "self-driving" mode on a separate roadway on the interstate system, but must be operated in conventional mode on the local streets. What's interesting about this is that it means a self-driving car isn't really competing against a conventional car as a mode of transportation, but against intercity trains and passenger airlines.
Something else to keep in mind is that the research being done doesn't necessarily have to culminate in a fully operational highway system for self-driving cars. Much of the technology under development has practical applications both inside and outside the auto industry, regardless of how far advanced the autonomous vehicle world gets.
7 Trillion is not unreasonable.
When you consider all the hours that people spent driving when now be converted to time that they can use productively, or rest, so they can be refreshed for work when they arrive.
When you consider when perfected the number of car accidents will be reduced by 99%, which reduced time spent in traffic (possibly thousands of hours for commutters per accident), the lives and hospital bills saved, the savings in less auto repairs, and the savings in lower insurance rates for all drivers. ‘
And the savings for people having to own fewers cars because they can be shared. Think of a husband and wife where the wife stays at home and he drives to work. Now they both need cars because she needs a car to get around during the day. But with driverless cars the car takes him to work and auto-pilots home, and is ready for her to use during the day before it auto pilots to work to pick him up.
I think driverless cars will be transformative for society, as much if not more than the smartphone.