Self-driving cars aren’t cars
Recently, David Silver wrote about a kind of epiphany he had when someone pointed out that the rise of self-driving technology will probably make any human presence unnecessary on many cars that carry goods.
His post points to two things that I think are often forgotten and therefore worth reminding ourselves of.
The first is that self driving cars aren’t cars.
What are they then?
I don’t know. I don’t think anybody does. I’m pretty certain they won’t be cars, though. They won’t be trucks or buses, either. I suspect they’ll be closer to messenger pigeons (they have something to deliver and a general sense of direction, but handle navigation pretty much autonomously), but even that isn’t certain.
It makes sense that we tend to think of self-driving cars as cars, we have no other frame of reference for them. We don’t even really have another name for them. Karl Benz built what would be considered the first car in 1886, and cars as we know them today became widely available in the early 20th century. Practically everyone alive today has only known a world of ubiquitous cars and this has conditioned us to frame our thinking about personal and product transportation in terms of automobiles.
But self-driving cars won’t be cars anymore than trucks are locomotives or cars are horses.
It’s interesting to consider the history of transformative technologies in the past. When the second industrial revolution began with the advent of cheap, readily available electricity, many expected a rise in productivity and living standards similar to what steam had brought about in the first industrial revolution. Spurred on by this optimism, a lot of factories were converted to use electricity early on.
The problem is that productivity gains were initially very modest. It seemed as though electricity was just hype after all.
Only, as we know now, it wasn’t. What kept early adopters from reaping the benefits of electricity is that they replaced steam driven machines with electric machines. But electricity isn’t steam. The early adopters didn’t understand what they were working with.
Until that point, the layout of factory floors had been dictated by the constraints of steam-powered production. You needed boilers that had to be accessible so you could feed them coal and water. You needed pipes to route steam so that it could power your machinery. These pipes had to be thick to insulate the team, they couldn’t be too long or the steam would cool.
Electricity made new, more efficient factory layouts possible. Cables to carry electrical current could be long and thin. It took quite some time for this to sink in, as people who were proficient in steam-centric technology struggled to “unlearn” and free themselves of steam’s heritage. (You can read more about this in an excellent NBER paper: http://www.nber.org/papers/w8676)
We struggle to unlearn the lessons of the car age to such a degree that we refer to self-driving cars as cars, and companies that work in this space are called car companies. We tend to imagine self driving cars as having the size of current cars, and driving on similar roads.
But autonomous vehicles could be bigger or smaller, even much smaller. They could carry much smaller loads (groceries? packages? letters? a single pill?), they might get purpose-built infrastructure that might go through buildings or under cities (do they need light? do they need oxygen? would you trust them to be able to let themselves into your home?). They will inherit a world built around cars, but they won’t be cars anymore than your computer is a typewriter or a filing cabinet.
A second important point is the underlying technologies that will make self-driving cars possible are being built by people who grew up in a car-dominated world and who often describe themselves as car nuts. In some ways, these people will stand in their own way before it becomes really apparent what it is they’re building. These people are steeped in a technology stack that’s about to become legacy, but no one can really see around the corner to see what the next one will look like. In some ways, proficiency in current car-tech exacts a price from people in that it primes them to think in car terms.
I expect some interesting cross-pollination as self-driving tech is adopted and transformed in other sectors that aren’t as invested in the notion of cars (Zipline is probably on the right track).
So what are self-driving cars? I don’t know, but I’m pretty sure they’re not cars.