Because that is already a possibility (carpooling is an established idea), and the critical reasons why it isn't more popular don't significantly change with automated cars. People don't want to spend their commute with random people who just happen to be going the same direction. Perhaps more important is that it creates a strict commitment to be prepared to leave on time and a reliance on others leaving on time.
> People don't want to spend their commute with random people who just happen to be going the same direction
Many, many people already do this on public transport. What we're talking about is general door-to-door public transport.
Car-pooling is awkward to arrange, particularly for short joined parts of the journey (e.g. picking someone up part way and dropping them off a little later), and it's extremely hard to arrange anything that overlaps such that the first passenger needs to get out before the last passenger (currently would involve handing over the keys to your car).
>Many, many people already do this on public transport
Total fantasy. Virtually nobody uses public transport. 5% of commuters use public transport. Relative to the car-pooling, who sees double the volume of users as public transport, 10%.
So what? The reason so few people use public transit across the US is very likely because it sucks at getting them door-to-door in a reasonable amount of time.
The question is: if that changes, what might people choose to do? The example of large cities like NYC at least teaches us that the American preference for solo car travel is not fixed.
Carpooling in private cars is very different from carpooling in self-driven cars. While you own the private cars, it's very reasonable to expect you not to share the ride with strangers or to occupy another's car. However, if it's a self-driven car which no one owns - it's just like public transit, but way more efficient. Instead of buying a monthly train or bus pass you just pay your self-driven car rental/lease, to be picked up door-to-door at your whim. This will happen and will bring about the 10x change.
As long as cars aren't prohibitively expensive and remain symbols of personal freedom and identity, then people will pay for the luxury of not riding in public transportation.
Some people have an aversion to riding on public transit when it is used by people large income/wealth disparities, i.e. the professional doesn't want to be on the same bus as the indigent person.
Those people might use a vehicle subscription, because the cost of private rides would still be lower than the costs of private vehicle ownership (depreciation, maintenance, insurance). Many people would pay for the privilege of a private ride (like they do with Uber/Lyft now).
If car subscriptions services succeed and are nearly ubiquitous, the personal-freedom symbolism of car ownership will degrade.
Then, if people want to own a car as a status symbol, they will probably be better served by a more expensive one anyway, the same way people buy other status symbols (rose-gold iPhones, etc). Ironically, most of those status-symbol cars will still have an autonomous mode.
I'm not talking about cars as status symbols, but as factors of identity. Teenagers want cars because it is a rite of passage. Half my coworkers want trucks because they are 'outdoors' people. And some do use it as a monetary status symbol. Some people are just car people who have strong opinions about the car they drive because it is part of their self-expression. This is a pretty big thing in American culture.
This is true with current car economics, but in a shared car economy, it will become increasingly expensive to have exclusive ownership, and people will vote with their wallets.
Shared-car design could also be drastically different. I don't want to share a Honda Accord with random people, but I wouldn't mind sharing a road-train that had private compartments for each set of passengers...
