The Long Road Ahead

Picture a world full of cars zipping down roads without a driver behind the wheel. You catch a ride to school from a robot that uses artificial intelligence and high-tech cameras and sensors to deliver you safely from door to door.

After Google unveiled its autonomous car prototype in 2014, tech experts promised fully functional self-driving cars were just around the corner. In 2015, Tesla CEO Elon Musk predicted they were only two years away.

Instead, what has resulted from the billions of dollars poured into developing autonomous vehicles are some clear advances, along with some high-profile injuries and deaths, and many unresolved questions. Some experts now say self-driving cars are still years from entering the mainstream.

“This is one of the biggest technical challenges of our generation,” says Dave Ferguson, an early engineer on the Google team who’s now president of Nuro, a company focused on delivering food and other goods using autonomous vehicles. He says many believed self-driving technology would improve like an internet service or a smartphone app. But robotics is a lot more challenging.

Picture a world full of cars zipping down roads without a driver behind the wheel. You catch a ride to school from a robot. It uses artificial intelligence and high-tech cameras and sensors to deliver you safely from door to door.

Google revealed its autonomous car prototype in 2014. Afterward, tech experts promised fully functional self-driving cars were just around the corner. In 2015, Tesla CEO Elon Musk predicted they were only two years away.

Billions of dollars have been poured into developing autonomous vehicles. The investment has led to some clear advances. But it’s resulted in some high-profile injuries and deaths and many unresolved questions as well. Some experts now say self-driving cars are still years from entering the mainstream.

“This is one of the biggest technical challenges of our generation,” says Dave Ferguson, an early engineer on the Google team who’s now president of Nuro, a company focused on delivering food and other goods using autonomous vehicles. He says many believed self-driving technology would improve like an internet service or a smartphone app. But robotics is a lot more challenging.

In 2021, after it became clear that humans wouldn’t regularly be commuting to work in self-driving cars anytime soon, Uber and Lyft sold their autonomous vehicle units as the industry consolidated.

One major appeal of self-driving cars is that they have the potential to be safer than having a human behind the wheel. After all, robots can’t drink and drive and don’t get drowsy on the roads. Yet confidence in autonomous vehicles began sinking in 2018 after one of Uber’s self-driving test cars hit and killed a Tempe, Arizona, woman walking a bicycle across the street.

In 2021, it became clear that humans wouldn’t regularly be jetting to work in self-driving cars anytime soon. The industry also began to shrink. In turn, Uber and Lyft sold their autonomous vehicle units.

One major appeal of self-driving cars is that they could be safer than having a human behind the wheel. After all, robots can’t drink and drive. They also don’t get drowsy on the roads. Yet faith in autonomous vehicles began sinking in 2018. That year, one of Uber’s self-driving test cars hit and killed a woman in Tempe, Arizona. At the time, she was walking a bicycle across the street.

Autonomous vehicles use cameras, artificial intelligence, and sensors to detect a person crossing the road and other obstacles. But it’s difficult to prepare self-driving cars for the uncertainty of the real world, such as encountering someone crossing against traffic or other cars making illegal turns. According to an estimate by RAND Corporation, autonomous vehicles will need at least 11 billion miles of road training before they can match the capabilities of human drivers.

“You see all kinds of crazy things on the road,” says Bryan Salesky, CEO of Argo AI, a company that makes autonomous driving technology. “With radar and high-resolution cameras and all the computing power we have, we can detect and identify the objects on a street. The hard part is anticipating what they’re going to do next.”

Autonomous vehicles use cameras, artificial intelligence, and sensors to detect a person crossing the road and other obstacles. But it’s difficult to prepare self-driving cars for how uncertain and fast things are in the real world. That includes instances like someone crossing against traffic or other cars making illegal turns. Autonomous vehicles will need at least 11 billion miles of road training before they can match the skills of human drivers, according to an estimate by RAND Corporation.

“You see all kinds of crazy things on the road,” says Bryan Salesky, CEO of Argo AI, a company that makes autonomous driving technology. “With radar and high-resolution cameras and all the computing power we have, we can detect and identify the objects on a street. The hard part is anticipating what they’re going to do next.”

Autonomous vehicles have seen some progress over the past decade. In October 2020, Waymo—the self-driving car unit of Google’s parent company, Alphabet—started the world’s first fully autonomous taxi service in the suburbs of Phoenix. Now anyone in Waymo’s service area there can ride in a car with no driver behind the wheel. Today, driverless taxis also snake through the streets of San Francisco, where both Waymo and Cruise—the self-driving unit of General Motors—received permits to begin operating their autonomous fleets. In essence, the companies are using public roadways to test and improve their vehicles as they operate.

“It’s a little bit jarring knowing that nobody else is in the car,” says 21-year-old Arizona State University student, Harsha Lakamsani, who took a trip to a mall in Chandler, Arizona, in a Waymo Chrysler Pacifica in May. “It’s almost like a ghost operating it.”

Autonomous vehicles have seen some progress over the past decade. Waymo is the self-driving car unit of Google’s parent company, Alphabet. In October 2020, it started the world’s first fully autonomous taxi service in the suburbs of Phoenix. Now anyone in Waymo’s service area there can ride in a car with no driver behind the wheel. Today, driverless taxis also snake through the streets of San Francisco. Waymo operates there, and so does Cruise, the self-driving unit of General Motors. Both received permits to begin operating their autonomous fleets in the city. In a sense, the companies are using public roadways to test and improve their vehicles.

“It’s a little bit jarring knowing that nobody else is in the car,” says 21-year-old Arizona State University student Harsha Lakamsani, who took a trip to a mall in Chandler, Arizona, in a Waymo Chrysler Pacifica in May. “It’s almost like a ghost operating it.”

Companies including Aurora and Waymo are also working to develop and test self-driving trucks. With long-haul trucking, the routes are simpler. According to Chris Urmson, Aurora’s CEO, once you master one stretch of highway, it’s easier to master another. Urmson and others argue that autonomous trucks will make it to market faster than anything designed to transport regular travelers.

But as self-driving technology drags in its development, optimism has turned into realism. The possibility of a future where cars, but no drivers, fill roads everywhere may be further away than experts predicted. But that doesn’t mean it isn’t coming.

“Part of me feels that self-driving is impossible,” says John Leonard, a robotics expert at the Massachusetts Institute of Technology, “but things that I think are impossible are happening today.”

Companies including Aurora and Waymo are also working to develop and test self-driving trucks. With long-haul trucking, the routes are simpler. According to Chris Urmson, Aurora’s CEO, once you master one stretch of highway, it’s easier to master another. Urmson and others argue that autonomous trucks will make it to market faster than anything designed to transport regular travelers.

But self-driving technology continues to drag in its development. That’s turned hopefulness into realism. The idea of a future where cars without drivers fill roads everywhere may be further away than experts thought. But that doesn’t mean it isn’t coming.

“Part of me feels that self-driving is impossible,” says John Leonard, a robotics expert at the Massachusetts Institute of Technology, “but things that I think are impossible are happening today.”