I mean, we were able to do level 5 trucks in the desert two decades ago! But the human world is much less benign, and the perception problem is hard. And the history of AI is littered with limited solutions that failed to scale.
So I'm more curious about the purported GM testing of level 4 in Manhattan, it's a much more realistic and far less forgiving driving environment.
That makes these subscriptions much more convenient than having to pick up/drop off a share car in a fixed spot.
The user uses the car for a while in semi-autonomous mode (car can drive 90, 95 or 99% with user supervision) and then the user can return it by letting it drive back. If the car is stuck it just stops and a driver from the fleet company comes along to drive it.
I think a scenario like that could be viable in small scale for some metropolitan areas (good climate, limited user base etc) within 5 years.
I think you drastically underestimate the possible values of pessimism. In my opinion, truly autonomous vehicles driving in arbitrary, real world, public roads in my lifetime sounds optimistic ...