You can have the on-demand voice-activated self-driving car experience right now: it’s called a taxi. Is that viable for all your use cases?
Perhaps for some of the use cases there could be better utilization (meaning cheaper) taxi, e.g. a taxi picking up my kids for school could also pick up a few more.
Eliminate the driver and taxis become way cheaper (probably cheaper than owning a car). That means people like alkonaut here will be far more interested in just not owning their own car that sits idle except when they are personally driving it. Which reduces the amount of space we are spending on car storage.
Why do you think they will be cheaper? Get ready for surge pricing being the norm. If you don’t have a vehicle of your own, you’ll have no choice but to pay. That’s why investors are willing to pour untold billions into this, they expect a massive payoff.
I'm sure that investors will extract a nice sum of money, but it's their ability to outcompete taxis on price that will allow them to do that. Or to put it another way they don't need to extract $30,000/car/year^1 and competition means they won't be able to.
^1 what ddg tells me the average US taxi driver salary is.
That's how it works. I walk 600m to my bus stop in my suburb neighborhood. So commuting from where I live is no problem. I don't need a car for that. But all the other transportation (driving kids to various sports practices, going to the forest, the lake, going to the school that isn't along a bus route with kids that can't walk very far etc) is harder. There will always be a gap between the transport I want/need and the transport that can be solved by public transport. So some kind of semi-private transport (e.g. shared autonomous cars) is probably required if the suburban lifestyle is going to be possible without cars.