Mass transit in an age of self-driving cars

Wired’s article about the nearing technical feasibility of self-driving cars makes several intriguing observations about the (possible) future of personal transportation:

[A]s we gradually relinquish our driving duties, it’s conceivable that we may also give up our sense of ownership of the car itself….Does anyone long for their teeming towers of CDs over the simplicity of Spotify? Similarly, we may well come to view the car less as an object to be owned than as a service to be streamed from the cloud. We already have the experience, through an app like Uber [], of summoning a car service with our smartphone, then watching as it moves toward us on a Google map, like the progress bar of a download. The final leap here is to envision a self-driving car that can be commanded like an elevator. When a car can drive itself to our door whenever we want it, why own something that spends more than 90 percent of the time simply parked? [emphasis added]

Most of Tom Vanderbilt’s piece dissects the current state of the art in self-driving cars (summary:  it’s incredibly good and getting better), but I think this speculation into some possible cultural effects is far more intriguing.  Many current conversations about transportation and its relationship to development assume an impasse between mass transit and private cars.  (See, e.g., Brian’s recent discussion of edge city Tyson’s Corner here.)  However, if cars drive themselves and can be “subscribed” to as a “service”, much of the efficiency of mass transit could be realized without changing America’s dominant infrastructure (i.e., roads and highways).  I imagine, for example, that each self-piloting car would be on the road for a far greater percentage of the time, lowering the overall demand for the number of cars.  Additionally, “auto-taxi” companies could offer varying classes of service, from expensive and on-demand (i.e., a car to one’s exact location immediately) to cheaper and shared (e.g., using algorithms to find other passengers demanding similar routes and combining them into one trip).  Perhaps transit authorities could even offer a greater choice (and variability) of bus routes , knowing that they would only have to actually run buses on routes and at times actually needed.

2 Responses - Add Yours+

  1. Neil says:

    1. Congestion. Even if robocars are made very small and bunched together, they’ll still take up more space than the most comfortable seat on transit. So commuter corridors/arterials will still want to run transit.

    2. However, this doesn’t mean they need human drivers. In transit, frequency is freedom. If a bus comes every couple of minutes, why even bother booking an “auto-taxi” – just stroll along to the stop and grab the next one. The big variable cost currently is that every vehicle requires a driver. Fortunately a LIDAR currently only costs one or two years’ salary and will only trend one way with scale; ditto software costs and solar-powered grid-balancing electricity costs. I imagine large fleets of 12-seaters on arterials, sitting charging if empty.

    The barriers are taxi oligopolies and bus driver unions (and mandatory-driver legislation) not technology or capital. Which, given the system-level environmental and wealth benefits, is a real shame.

  2. [...] bloggers are starting to tease out the social and infrastructure implications of driverless cars, including David Alpert over at the Atlantic: [Driverless cars] will bring many changes, but when [...]

Leave a Reply