As for the rest, keep in mind we're comparing renting an autonomous car vs owning an autonomous car, not renting an autonomous car vs driving. So the actual rides are equivalent. The time difference is me getting in my car and telling it where to go vs me hailing a car, waiting 5-10 minutes for it to arrive (where my capacity to do any meaningful work is basically nil) and then tell it where to go, a minimum of 3 times a day. Maybe up to 12 times a day on a busy day. That's not an insignificant time loss over the medium/long term.
Granted it's all hypothetical and the actual value of the time lost would be highly situational, but it would be one of those small daily time-sucking inefficiencies, like walking into the other room to get paper towels as opposed to just putting a roll in the kitchen. When you do them every day, those add up.
Nope, but they will be commoditized. If you have a car that all you have to do is pass it an API key and a location, there's not much value a ride sharing service can add to that -- which means there will probably be a bunch of them. Which means that the price to the consumer should be something close to [ (marginal cost of ride) + (depreciation cost of ride) ] * 1.03. And the cost side of that will benefit from scale for the business, but not the consumer.
Basically, once it's commoditized, it will cost you more to do it yourself than to pay someone to do it for you. Just like with currently commoditized services like AWS, there can still be good reasons to do it yourself -- but those tend to be special cases rather than the norm.
> But in this scenario I'm losing a minimum 90 hours per year on the margin.
Err, I'll play along. If you were in an autonomous car (or even a current ride-share), you could just work while en route rather than drive. I'm willing to bet you spend more hours sitting in traffic today than you ever would waiting on a car to arrive.