Čvc 10, 2019
It’s difficult to unnerve drivers in China, where on any given day roads are filled with innumerable obstacles. Pedestrians walk haphazardly on highways and food vendors push carts loaded with snacks through the streets. Flocks of electric scooters whizz around cars, taxis swerve into oncoming lanes to circumvent traffic jams, and bicycle rickshaws stop abruptly to pick up passengers. Motorists use horns with abandon and speed through intersections, disregarding red lights.
Unsurprisingly amid this chaos, drivers in the sprawling southern Chinese city of Guangzhou adapted quickly to seeing self-driving cars on their roads when tech startup Pony.ai started testing the vehicles there in late 2017.
“We used to see people taking out their cell phones to take a video or photos,” James Peng, the co-founder and CEO of Pony.ai, told TIME. “Now we get people trying to test our vehicles by aggressively cutting in front of them.”
Peng, who led development of the autonomous driving division at the Chinese search-engine company Baidu, founded Pony.ai in 2016 with another ex-Baidu employee, Lou Tiancheng, a renowned coder. Things have moved quickly in the past three years.
A few months ago the company launched an Uber-like app called PonyPilot—and naturally the car that picks users up doesn’t have a driver. The app is currently open to the company’s employees (who use it to commute to work and to get to their lunch appointments), government officials, and invited individuals. But Peng says they plan to open the service to the general public as soon as regulation allows for it.
Pony.ai also operates in the U.S (where companies like Alphabet’s Waymo and Tesla are furiously competing to get an edge in the autonomous driving space) and was just granted a license to test a robotaxi service in California.
Peng talked with TIME about self-driving cars and how China’s autonomous vehicle startups are faring against their U.S. competitors.
How is PonyPilot going so far?
It’s going great. We’re continuously scaling up the size of the fleet, the hours of operation and the operational areas. We’re now covering 50-60 square kilometers. Because of the scale, because of the convenience, we’ve started to have much higher ridership.
Pony.ai is testing autonomous technology in both the U.S. and China. How does the environment for autonomous driving differ in China?
They’re very different driving environments. The traffic behavior and patterns are quite different. China’s roads generally are populated with much more diverse road agents: bicycles, tricycles, trucks of all different sizes, scooters, and pedestrians everywhere. You name it, it’s on the road.
Read More: Waymo and Jaguar Are Making Self-Driving Electric SUVs Together
Can you tell us about your first experience in a self-driving car?
Calling it life changing may be a little bit exaggerated, but it’s to that degree. To see a vehicle drive by itself without human intervention was like science fiction to me, at that point.
How often do you ride in autonomous vehicles?
Several times a week. In the earlier days, I was more or less a de-bugger. These days, I use it just for a commute, so that shows the maturity of the technology.
Why is the advancement of autonomous driving important?
We still see more than a million people die from traffic accidents a year and millions of people get injured a year. I think something like 90% of traffic accidents are caused by human error. People get tired and distracted. By using computers to drive a vehicle it can make the safety much, much better, especially as technology gets better.
You have headquarters in Guangzhou, China and Silicon Valley? Why is it important to have operations in both regions?
A big reason is talent. In terms of AI, the two biggest talent pools are the U.S. and China. We go where the talent is.
Autonomous vehicles have been tested on public roads in the U.S. for years. In China, the government just started issuing permits for testing last year. Has this given companies in the U.S. an advantage?
I think the U.S. companies have first mover advantage, but not necessarily competitive advantage in the long run. Government support in China can foster much faster development. Currently, in terms of technology, the U.S. still leads the way, but China is catching up quickly.
When do you think fully autonomous cars being will be available?
I think within 3 to 5 years there will be thousands or tens of thousands of vehicles on the road.
—This interview has been edited for length and clarity.
