When autonomous cars are here

In 2009 many people, especially investors in silicon valley, told me that there is no sense in statistical solutions to parking because sensors will be everywhere within a few years. In particular, they said, parking will be obviously solved by having sensors everywhere that would tell you if parking is available or not.

I am a technologist for all of my life, so I saw such promises before, but being a scientist I knew that this promise cannot be done with current technologies and understandings.

While the claim that sensors would be everywhere is true it is also irrelevant. Yes, IoT (Internet of Things) sensors, cell phones, sensors on cars, sensors in the home, sensors in the sky, are indeed all there. And yet, sensors in the way that my investor friends envisioned were not (and are not going to be anytime soon) possible in a cost-feasible way. Nobody would install sensors at even $1 per sensor per year per parking spot in the street, where streets have free parking because there is today no way of reclaiming that cost.

My investor friends were confused by marketing that said sensors would be everywhere (propagated by those who invested in sensors), not noticing that it did not imply the conclusion that they had.

Today many people, especially investors in silicon valley and Israel, tell me that when autonomous cars are here two things will be true: (a) sharing economy — everyone would use an Uber-like service for their mobility, and (b) cars would not need parking. In my opinion, they confuse marketing saying these things (propagated by those who invested in Uber, electric cars, and autonomous cars) with the conclusions from those statements.

I believe neither (a) nor (b) would be true. In a future of 10 or 20 or 30 years from now, when autonomous cars are possibly going to dominate the roads, sharing would enable transportation, but would not dominate it. Most trips would still be with owned cars (or other owned devices), because people would still need to get to work at 8am and come back home to their kids at 5pm. At 8am and 5pm there would be peaks of demand, with drops in demand at midnight and mid-day.

Those cars that supply the demand would need to be somewhere, such as in a parking garage or parking on the street, or otherwise would be spending valuable energy (electrical or fossil) for hours per day. Assuming the “no free lunch” fact of life, someone would need to pay for those hours of cruising or parking, and market forces would dictate a stable point where price of cruising would be slightly less than the price of parking (because cruising costs also contribute to needed maintenance of cars).

Investors are smart and knowledgeable, and have taken their Macro Economics 101, so they know that the above analysis is true. (In a separate article I would make this precise, showing how the same models for house prices versus distance from desirable areas are those that apply here as well for parking prices versus the price of cruising.) The place where their predictions do not reflect their understanding of efficient-market solution is where market trends (fueled by marketing) are projected in a linear way to their limits.

Our world is not a linear projection of trends, it is a combinatorial non-convex labyrinth of trends intersecting with other trends. Linear projections of trends simplify and explain short-term movements and evolutions, but they rarely can predict beyond 2–3 years. To see where the future leads, we should consider the intersection of those trends.