“Second of all, and I can’t stress this enough, we won’t have fully autonomous driving cars for at least 50 years.”
I don’t suppose you’d care to place a small wager on that...? We could use a much shorter time horizon, say 15 years. I think in reality it’ll be less than 10.
(My definition of “fully autonomous car” is that you will climb in, select a destination reachable on the current fuel supply, then the car will drive to said destination and park with no human driving inputs.)
I work professionally in a field where we are dealing with the implications of these "self-driving vehicles" on infrastructure and on traffic operations, and I believe we're much closer to a 50-year horizon than a 10-15 year horizon for what you describe.
Keep in mind that the technology isn't really the biggest challenge here. The technology already exists to do this -- and it has existed for much longer than most people probably realize (my first exposure to automated vehicles was on a test track way back in the early 1990s when I was still an undergraduate in engineering school).
The biggest challenges are two-fold:
1. Getting all of the state laws and legal standards updated to reflect a major change in how the legal system treats automobiles (from personal/company liability to product liability).
2. Related to the previous point ... Getting the auto industry to figure out how to mitigate their legal exposure in a way that doesn't force these companies to build so many safety measures into these vehicles that the technology is basically useless.