Many people tend to think about robotic mobility as nestled within our existing world of household vehicles, potholes, congestion and complex intersections. Others describe it in the context of a far-Jetsons future when no one living will even remember what it must have been like to own or drive a car. Neither of these perspectives will help us on the ground, in our cities, over the next 15 years.

In reality, robotic mobility will be channelled and crunched among hundreds of confusing, exciting and competing changes overlapping all at the same time. There will be more social change than technology change. More policy lag than user reticence.

The AV of 2030 will more likely be printed and assembled locally than manufactured and delivered via a dealership. It will be upgraded over the air like a smartphone is now (and Teslas are already). Because the technology for robotic mobility will evolve so quickly household vehicle life-spans will plummet and their ability to retain resale value after purchase will be abysmal. These vehicles will not have 12-15 year life spans. Three years would be a more likely average. If you are on your second or third smartphone, you will understand. A motor vehicle that is used only 5% of its calendar life and have little re-sale value after 3 years will be a very poor investment, indeed. Who would provide a five-year lease or loan on a three-year asset? Robotic vehicles will become obsolete far too quickly for ownership.

Perhaps all the opportunity will inure to planners of urban transit.

The transit AV will be designed for quick maintenance, daily cleaning, a 24-month life span and made from 95% recyclable materials, rather than bristling with features to attract household buyers. Minibuses for autonomous transit can be 100% electric opening the door to 100% renewables transit.

To ask the question: Should we start planning special lanes for AVs? means you are seeing the AV as only the little Google granny car that gets rear-ended for over-cautious rule-following at intersections . By the time such lanes would be planned and built with sufficient coverage to matter to the household AV (untold billions of dollars), robotic technology would be several generations beyond any such need to be coddled. Besides, one need only consider the effort, expense and political capital expended to wrest a few bicycle lanes from under the wheels of our fossil fleet to walk away from the task of building Google granny lanes.

To ask the question: should we be prepared to repaint the lines on our highways and turn four lanes into six? means you are thinking like a driver instead of a murmuration of starlings. Schools of fish, too, do not need painted lines to avoid sideswiping each other.

To ask the question: how can we proceed to exclude non-autonomous vehicles, roadway-by-roadway, area-by-area? would be a more interesting question. Clearly this cannot be done all at once, but it will be done, just as horses are excluded from many roadways today.

Nothing should be built or torn down just yet. Rather cities should focus on disrupting transit by deploying autonomous minibuses and vans in constrained applications to handle first and last mile problems already begging for solution or short local loops that cannot be sensibly served by 40-passenger buses. Such deployments would add jobs, and familiarize transit staff as well as users with robotic TaaS. If deployed strategically, such growing transit applications would spread to ever-larger areas, further encouraging travelers to abandon household ownership completely. The only way for transit unions to protect transit jobs is to grow transit ridership and boardings. At the discretion of our cities and their transit agencies, autonomous vehicles could destroy transit utterly or could leverage a ten-fold increase in transit use.

The AV will not leave transit alone. The time to be thinking about this is now.