Natalie Behring / Reuters

Subscribe to Crazy/Genius: Apple Podcasts | Spotify | Stitcher | Google Play

I have seen the future of the car. It looks like a minivan.

Earlier this summer, I went to Mountain View, California, to visit the headquarters of Waymo, the autonomous vehicle company that spun off from Google, and get driven around in one of its cutting-edge cars. The model waiting for me in the parking lot was deceptively fuddy-duddy: a white Chrysler Pacifica, outwardly distinguished from the typical minivan by a small black dome on the roof, which houses a family of cameras, sensors, radars, and lasers.

Rarely is an experience made more boring by the presence of lasers, but this is that rare experience: The ride itself was astonishingly tranquil. Making use of technology that’s a bit like a multi-sensory echolocation, the car turned smoothly out of the parking lot. It navigated through several wide intersections. It decelerated calmly. It stopped politely for pedestrians. When I got out of the car, it seemed almost tedious to point out that a minivan drove itself pretty much perfectly. It was only later, upon reflection, that the sentence returned to me with proper emphasis: The minivan … drove itself … pretty much perfectly. And that means it’s time to start thinking about what happens when cars just like it can drive millions of other people perfectly, too.

In the latest episode of Crazy/Genius, produced by Kasia Mychajlowycz and Patricia Yacob, we talk to experts about how autonomous vehicles might change the world. People tend to predict the future by analogy, and so self-driving cars’ potential has been compared to a smartphone, an elevator, and an autopilot. But after speaking to several technologists, urban planning experts, and historians for this episode, I’ve come to the conclusion that the best metaphor for the self-driving car is a 20th century technology so subtle, almost invisible, you probably forgot that it had to be invented in the first place.

The self-driving car is a highway.


From a distance, car culture looks like a cavalcade of waste.

First, it requires an extraordinary sacrifice of life. Approximately 1.2 million people die every year due to automobile accidents around the world. In the U.S. alone more than 30,000 people die from cars, making it the leading cause of death for people under 35. “This is insane,” Dmitri Dolgov, Waymo’s head of technology, told me. “Our cars, they don't get distracted, they don't get tired, they don't fall asleep, and they don't drive while drunk.”

Second, consider the wasted space in cities. Before the 1920s, most urban residents commuted to work using public transportation. Today, eight in 10 Americans drive to work; just 5 percent still take public transit. The vast majority of commuters drive alone, despite occupying a big metal box designed to hold a medium-sized family. In the space taken up by a typical car with one driver on a highway, rail transit can fit 25 people.

Finally, there is the wasted investment: Americans spend tens of thousands of dollars on a car—more than any asset besides a home, on average—but leave the machine idle for 95 percent of its lifespan. As the automobile is the mechanical equivalent of a sloth, it only follows that it would need ample room to nap. In cities like Philadelphia and Seattle, there are four parking spaces for every household, requiring gaping cavities of concrete in some of the most valuable real estate in the country.

“The automobile has been pretty much the same ever since it was invented in the late 1880s by Karl Benz,” Lawrence Burns, a former GM executive and author of the new book Autonomy, told me. “It is really messed up! But that's the way it's been for over 130 years.”

Burns has a bold vision of the future, where just about everything about the car’s relationship to the consumer will change. Instead of buying cars to keep in their garage, more riders will choose to subscribe to trusty autonomous fleets (like a taxi service or Uber, but cheap, electric, and driverless). For more than 100 years, car companies built driver-centric machines. But self-driving cars will be designed around passengers—consumers—and that means they don’t have to be what we consider “cars” at all. Imagine a coffee shop on wheels, that drives you to your doctor’s office. Or a comfy hotel bed that drives you to your in-laws for Thanksgiving. Or an Amazon store carrying household staples that roams a downtown area, waiting to be hailed to an apartment.

If you think of driverless cars as nothing more than cars without drivers, Burns says, you’re not seeing the full picture. These will be rooms with wheels. And that means their implications extend far beyond transportation—into retail, commerce, and even an expansive re-imagination of where Americans should live. Commuting an hour to work from the far suburbs isn’t such a drag when the autonomous pod comes with Netflix, email, and Wi-Fi.

But not everybody is so enamored by this vision. “I see our future of autonomous vehicles going either towards a heaven scenario or a hell scenario,” says Robin Chase, the co-founder of Zipcar. If self-driving cars are as useful as their boosters say, everybody will want one—every family, every car company and city trying to build a municipal fleet, and every retailer trying to put their business on the cheapest, most-valuable real estate available: public roads. Without smart urban planning, the result will be infernal congestion, choking every city and requiring local governments to lay ever-more pavement down to service American automania.

In sum, self-driving cars have the potential to improve existing transportation technology in unambiguous ways, to expand the suburbs, and to create new economic opportunity for a variety of industries, from hotels to restaurants. But they might also change the character of our cities for the worse and strangle roads with cars in a way that ruins the urban experience for millions of people. What does this sound like? It sounds like the legacy of highways in America.

“The concept of the highway was all about the idea of unfettered mobility,” Eric Avila, a historian at UCLA, told me. “You might think of the early highway system kind of the way that we thought about the internet 20 years ago. People had access, autonomous access, in a way that they had not had before.”

Between the 1950s and 1970s, the U.S. laced its landscape with 48,000 miles of highways. The infrastructure stimulated interstate trade, swelled the suburbs, and made road trips a thing. But the construction of urban highways—like those in New York City and Los Angeles—gutted neighborhoods and allowed the richest families to glide back to leafy suburbs on “ribbons of asphalt” taking tax revenue with them. “Mass transit systems [and] streetcar systems were systematically dismantled,” Avila said. “There was national protest in diverse urban neighborhoods throughout the entire nation.”

Just as it is too simple to say that the highway was a force for good or ill, it’s similarly reductive to predict that self-driving cars will be inherently good or bad. The highway was a tool with the capacity to both connect people and divide them. Autonomous cars aren’t just cars. They’re infrastructure. If the U.S. wants to avoid its past mistakes, American cities have to start thinking about how to use autonomous cars to make their downtown areas more efficient for humans, not just for machines. That could mean aggressive congestion pricing, limiting the number of AVs in an area, prohibiting certain companies from operating such cars, or even capping ownership. Today these may seem like radical steps. But compared to a future where every family and company tries to operate an autonomous car in the same metro area, it may feel like pure old-fashioned conservatism.

There is a future where a handful of efficient, self-driving cars operated by a responsible fleet manager could return American cities to a superior age, where people mixed promiscuously around the streets, rather than be confined to narrow sidewalks to maximize space for cars. A few years ago, my colleague Alexis Madrigal went on a quest to hunt down America’s first “city soundscape”—an audio recording of an urban area. He found this; the recording dates back somewhere between the 1890s and early 1900s. (It’s technically a vaudeville group performing the sound of a turn-of-the-century town.)

If you listen closely, beyond the crunchy whir of the brown wax cylinder, you can make out the urban scene. Newsboys yell “Extra! Extra!,” bells chime, horses clomp down roads that would be shared with streetcars. Then, unmistakably, the sound of a shiny new invention pierces the hubbub. A car horn honks and then fades. It’s the sound of a machine that doesn’t dominate a city. The car is simply a part of the crowd.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.