If you're a believer, then the autonomous car is a gateway to a brave new world in which you'll never have to waste an hour looking for a parking spot, ever again.
An obvious challenge to this vision is whether it encourages traffic-choked, polluting and wasteful behaviour: instead of parking while you visit the supermarket, just set your car to circle while you choose between brands of cola.
But here's a different outcome of a world in which driverless cars exist, and operate: a sales crisis for the motor industry.
For this thought experiment, I'm using figures from Australia where numbers are needed; feel free to adjust to your local data.
According to the Australian Bureau of Statistics, there were around eight million car commuters in Australia last year, and around six million of those travelled alone. Various sources give the average driving commute time as around an hour.
What I don't have data for is the number of vehicles that travel outside the commute (because the owner uses it during the day), but let's assume that half of all commuter cars – four million – are idle from nine to five. and don't get many hours of use other than for commuting purposes. Those assumptions mean they're spending the overwhelming majority of their time doing nothing.
That presents a huge opportunity for the cars to be multiplexed: sent on other trips carrying other passengers while they're not needed for the commute.
“No way,” I hear you say. “Why would I let strangers use my car while I'm at work?”
To which my response is “why should it be your car?”
Why own a car at all?
Already, there's a trend among “millennials” away from car ownership: they use taxis for short trips. In Australia, many have accounts with car-share operators like GoGet for longer trips. IKEA Australia even has a deal with GoGet that allows shoppers to rent one of the company's vans for the hour or two needed to get flat-packed furniture home.
The driverless car offers a hybrid car-share/taxi business model: a fleet of half a million autonomous cars, backed by someone with funds, computer power and a smartphone app, operated under the same subscription-plus-usage model that car-share already uses, with the ad-hoc booking model of a taxi, and cheaper than both.
It's cheaper than a taxi, because you don't need to fund the driver's salary, the (in Australia) expensive taxi plates, the radio network and other expenses.
There are two reasons the model could be cheaper than car-share operators.
Right now, car-share operators have to get insurance for all kinds of drivers; if autonomy works, it'll be cheaper to insure a pool of known drivers. Second: car-share has limited “multiplexing” – if I take a car 20 km to a meeting and park it for four hours, I have to pay for the idle time.
If I subscribe to an autonomous car service with a big enough fleet, smart enough software, and the right trip price, I can treat it as a taxi: request my trip on the app, walk out the door, and get in the car. If it happens to be peak hour, I may see others in the car for the same trip. At other times, perhaps not.
It's more efficient than car-sharing or ownership. Trips can be shared on an ad-hoc basis (rather than by some formal arrangement), simply because the computer can identify requests that dovetail with each other; and the same computers can minimise the vehicles' idle time, meaning that I don't have to pay rental on a parked car.