The Robots Are Coming!
Page 26
8
THEY’RE COMING FOR FACTORY WORKERS!
THE FUTURE OF TRANSPORTATION AND MANUFACTURING
MOUNTAIN VIEW, TOKYO, SEOUL
Elon Musk, the founder of the Tesla motor company and one of the world’s most innovative billionaires, wasn’t kidding around when he said that in less than twenty years, driving a car will be as old-fashioned as riding a horse. “Any cars that are being made that don’t have full autonomy will have negative value. It will be like owning a horse. You will only be owning it for sentimental reasons,” he said. Musk has earned an international reputation for making bold predictions and investing hundreds of millions of dollars of his money in daring projects. In 2002, he founded the Space Exploration Technologies Corporation, better known as SpaceX, to produce reusable spacecraft, which he says—seriously—that he wants to use to establish a colony of 80,000 people on Mars. Since then, SpaceX, with roughly 6,000 full-time employees, has become a leading force in the aerospace industry and one of NASA’s primary partners in space research.
And while it remains to be seen whether Musk will achieve his goal of establishing a human settlement on Mars, he’s already well on his way to making Tesla one of the automotive industry leaders in self-driving cars. Unlike General Motors, Ford, and other automakers, Tesla has started from day one to produce electric cars and to experiment with self-driving technology. In 2016, the company was already selling electric cars with automatic driving, braking, and parking. For legal reasons, a person had to be behind the wheel, but instead of driving the car, the driver could be resting or checking emails. And while a fatal 2016 accident in Florida involving a Tesla S made headlines around the world, the company rebounded from that public relations fiasco when an investigation later revealed that the car had warned the driver several times to take over manual control, and the driver had failed to do so.
Experts have long concluded—and investors seem to agree with them—that self-driving cars will be much safer than human-driven ones. According to Google, whose subsidiary, Waymo, has been producing sensors for automated cars for years now, autopilot technology will reduce the number of deaths from car accidents by more than 90 percent. The reason is that the vast majority of today’s accidents are caused by human error, whether it’s drunk driving, texting, or falling asleep at the wheel. As Brad Templeton, a member of the Google team working on self-driving cars, once told me, “Since robots generally don’t drink or fall asleep at the wheel, self-driving cars are much safer than the ones we’re using today.” Indeed, according to the World Health Organization, every year more than 1.2 million people around the world die in car accidents. If Google’s estimate is correct, cars driven by robots could save over a million human lives a year.
And many investors have bet that Musk will be a major player in the self-driven car industry. So many that in 2017, Tesla exceeded the market value of General Motors, even though Tesla produced just 76,000 cars per year compared to GM’s 10 million vehicles. Obviously, investors and their financial advisers were taking Musk seriously when he said that driving a car would soon be a thing of the past, just like riding a horse.
IN CALIFORNIA, SEMIAUTONOMOUS CARS ARE ALREADY ON THE ROAD
It didn’t take much investigating to conclude that self-driving cars will be hitting the streets very soon. I have already seen many of them wandering along the roads of Mountain View, California, and Miami, Florida, as part of test runs by Waymo—Alphabet’s self-driving car subsidiary—as well as Uber, GM, Ford, Volvo, Audi, and several other car and tech firms. Semiautomated Teslas were everywhere in 2018, and several other companies are planning to produce fully automated cars commercially by 2020, if not sooner.
Driving along Route 101 from San Francisco to Mountain View for several interviews at Google and other companies working on self-driving cars, I was struck by how many semiautomatic Teslas were on the road. I could spot them easily: the ones ahead of me kept their course with almost absolute precision, without drifting an inch to either side. Obviously, they were on autopilot. Their human drivers were behind the wheel, but they were reading or texting on their cell phones, quite confident in their vehicles’ automatic brakes.
When I visited my friend Vivek Wadhwa, the Mountain View–based professor at Carnegie Mellon and Harvard and author of several books on technological innovations, he told me that driverless cars are already a reality in Silicon Valley. He said he had bought a Tesla the previous year, and that for all practical purposes, it was already a self-driving vehicle. “I’m behind the wheel because legally I have to do that,” he said, “but the car drives itself just about all the time.”
Wadhwa, who had just published a book titled The Driver in the Driverless Car, told me that one of the main reasons autopilot technology is progressing so rapidly is that all autonomous vehicles on the road are constantly learning from one another. Collectively, they have already traveled millions of miles, and they share the information they pick up on the road each and every day.
SELF-DRIVING CARS WILL BE MUCH CHEAPER
“Few people seem to fully grasp the profound improvements in our lives that driverless cars will bring,” Wadhwa wrote in his book. “Their adoption will slash accident and fatality rates, saving millions of lives. As well, it will remove one-third to one-half of all vehicles from city streets.” Fully automated cars in New York, San Francisco, or other big cities will no longer have to be on the streets at the same time looking for a parking place, because they won’t need to park. They will be able to continuously circulate, picking up and dropping passengers, or parking themselves outside the cities, he said.
In addition, growing numbers of people will stop buying cars. Instead, they will use Uber, Lyft, and other car-sharing services that offer cheaper and more efficient transportation in cities. And these companies’ services will become increasingly cheaper. Today, about 60 percent of what you pay when you take a ride with Uber, Lyft, or another private taxi company goes to pay for driver-related costs. Once Uber, Lyft, and other ride-sharing companies start using robotic cars without human drivers, their prices will plummet, and even more people will use them.
According to a UBS study, the number of private car owners will drop by 70 percent by the year 2050. Plus, private taxi companies will be able to reduce costs even more because self-driving cars will cost much less to produce than traditional cars: they don’t require steering wheels or brake or gas pedals, nor many of their other current features. As accidents are gradually reduced to near zero, autonomous cars may not need steel-reinforced doors and fenders, airbags, or even seat belts. And since almost all self-driving cars will be electric, the savings in terms of gas will be tremendous.
Why own a private car if it spends 95 percent of its time parked in your driveway or garage? Growing numbers of people will conclude that it makes more sense for them to use automated taxis. Robotaxis will be responsible for a quarter of the total number of miles driven on U.S. roads by as early as 2030 and will reduce the number of cars in cities by a whopping 60 percent by then, according to BCG consulting firm projections.
“My grandchildren will ask me to tell them what it was like to drive a car in an old city. I’ll tell them it was scary, dangerous, and wasteful, and that they are lucky to have a better way of living,” Wadhwa wrote. He added that he will tell his grandchildren that the few accidents that did happen during the transition from human-driven to self-driven cars “were because of the pesky, ill-mannered, and dangerous humans that they had to share the road with.”
MY EXPERIENCE WITH A SELF-DRIVING CAR: BORING!
Lauren Barriere, a spokeswoman for Waymo, had arranged for me to take a ride in one of its autonomous vehicles during my visit to Google’s headquarters in Mountain View. As we had agreed several weeks before, a Google engineer came to pick us up in one of the company’s driverless cars. It was a white Lexus sedan equipped with a radar dish on the roof and
small cameras with sensors poking out from all sides.
After the standard introductions and presentations, the engineer sat at the wheel, just as he would in any conventional vehicle. He was required to do so by law, he explained, even if he didn’t so much as turn on the radio. And we had to wear our seat belts just as we would in any other car. Inside, there was nothing special, save for a laptop connected to the dashboard that was positioned next to the engineer and which he would use to drive—or rather, control the car’s movements. When all of us were strapped in, the engineer hit a key on the laptop, and the car started to move.
What followed was the most boring ride of my life. The Lexus left the Waymo offices at a snail’s pace. At first I assumed it was just a precaution to avoid any accidents there in the parking lot. But when we got out onto the road, the car was still crawling along at less than twenty miles per hour, despite the fact that the streets were almost empty. Barriere and the engineer explained that their self-driving cars are programmed to never exceed the speed limit on any street. And in that part of the city, the speed limit was twenty miles per hour.
Soon enough, the ride went from boring to excruciating. Not only were we moving at an annoyingly slow pace, but the driverless car wouldn’t pass any other vehicle. Even when the car in front of us was going much slower than we were, ours hung back obediently, waiting patiently for it to get out of the way or for everyone to turn onto a wider road where it could change lanes. And to make matters worse, every time we came to a stop sign, the car came to a complete halt—as is the law—before waiting three never-ending seconds before moving forward. It was a special precaution to avoid any potential accidents caused by a careless human driver who might be coming in from one side or another, the engineer explained.
About fifteen minutes into our ride, I looked pleadingly at Barriere, hoping that she could put an end to this torture. Ever the PR manager, she explained that a self-driving car’s primary objective was precisely that: to avoid accidents. “The biggest compliment you can give to our engineers is that the trip was boring,” she said, obviously repeating a line she had delivered to other guests before. I nodded as politely as I could. But at the same time, I was thinking to myself that transitioning from traditional cars to self-driving models won’t be all that simple. People stuck behind an autonomous car moving at a snail’s pace on an empty road will get impatient, honk, and look for any opportunity to pass it. It won’t be easy to get used to it!
AUTONOMOUS CARS ARE ALMOST READY TO GO
Although the autonomous car revolution probably won’t shake up the world economy until the early or mid-2020s, these cars have been built and tested for decades. The reason that they are attracting so much attention now is because they have become much more reliable in recent years thanks to new sensor technology that allows them to react most of the time to unforeseen situations, like a dog running across the street on a foggy night. With a few exceptions, like the 2018 accident in which an Uber driverless car killed a woman who was crossing the street outside the designated path during the night in Tempe, Arizona, these problems have been overcome. At the time of this writing, technicians are working feverishly to get automated cars to detect increasingly complex signals, like the hand signs of a police officer directing traffic around an accident and other nonverbal communications, such as a cyclist trying to make eye contact with a driver in order to get permission to pass.
Just about every major car maker or tech company experimenting with robotic cars is on the verge of removing the last few obstacles remaining in the path of the fully autonomous car. These companies have hundreds of these cars on the streets in several cities, with engineers behind the wheel ready to take control if some unforeseen problem arises, but such cases are becoming less and less frequent. In fact, as Waymo’s Barriere told me, the number of emergency cases where engineers had to step in and drive fell from 0.5 percent per 1,600 kilometers driven in 2015 to 0.2 percent in 2016. And everything indicates they will be hitting the final goal of 0 percent in the very near future. “The car is now ready to drive,” she said, “and very soon you’ll be seeing them on the road in places where it’s relatively easy to drive, like where there’s no snow. Bit by bit, they’ll be operating in more complicated conditions.”
THE DRIVERLESS CAR ACCIDENT THAT KILLED A PEDESTRIAN
The 2018 accident involving an Uber driverless car that killed a woman in Tempe, Arizona, drew worldwide attention because it was believed to be the first pedestrian death caused by an autonomous vehicle. Shortly after the crash, Uber announced it was temporarily suspending its experimental program in Arizona, fueling speculation in the media that it could be a serious setback for all autonomous car companies. But shortly after the accident, Tempe police chief Sylvia Moir told The Arizona Republic that the initial police investigation and review of the car’s video had revealed that the accident was most likely “unavoidable.” In other words, it would have been difficult for a human driver to see a woman crossing that street on a dark night at ten P.M. outside any crosswalk.
Months later, I had a chance to interview Uber CEO Dara Khosrowshahi, the forty-eight-year-old Iranian American tech executive who had previously been CEO of Expedia, and was a member of the board of directors of Hotels.com and The New York Times. He is a media-shy tech executive—his press handlers insisted that we didn’t take video of the interview, because they said he was not yet feeling comfortable with doing TV interviews—who had taken over Uber in 2017 after a tumultuous year that had led to the resignation of the company’s founder, Travis Kalanick.
I asked Dara—as he is referred to by virtually everybody—to what extent the Tempe crash would delay or halt the development of autonomous cars. Dara said the crash was “a tragedy” and “a step back for our company,” but added that “ultimately self-driving will improve safety across the board.” He told me that “these machines will get better and better on an infinite basis. A self-driving robot learns how to drive, and then keeps getting optimized. The equipment and the sensor packages are going to continue to improve. So, at maturity, I can confidently say that this accident would not have happened in a mature self-driving environment.”
He added that “this mature driving ecosystem will take a lot of work, but I’m 100 percent confident that at the end it’s going to be much safer than where we are now.” When I asked him when he expects that to happen, he responded, “In the next eighteen to twenty-four months, in 2019 or 2020.” In other words, the fatal accident in Tempe, Arizona, was not likely to stop the development of autonomous cars, much like the early fatalities in the first flights didn’t stop the development of commercial aviation.
TRUCKERS WILL BE AMONG THE FIRST TO FEEL THE IMPACT
One of the first places where we will see commercial driverless vehicles will be the highways. An increasing number of fully automated big-rig trucks like the ones produced by Otto, a company founded by former Google engineers and recently acquired by Uber for $700 million, are already being used on an experimental basis in California. According to Matthias Kässer, a transportation industry analyst at McKinsey, more than a third of the trucks on American highways will be fully autonomous by 2025.
That will affect the jobs of large numbers of truckers. While most of us who live in big cities don’t see them very often, there are tens of millions of truck drivers in the world. The American Trucking Association estimates that there are 3.5 million professional truck drivers in the United States, not including several million more who do related jobs—from maintaining engines to handling paperwork—and are not sitting directly behind the steering wheel. It’s a fairly well-paying job in most countries, especially for those who are both truck owners and drivers. But what will happen to them when their vehicles become fully automated? In the short term, human drivers will still be needed to get their trucks in and out of congested cities, but it won’t be long before the autopilots can handle even those tricky s
tretches.
Lior Ron, one of the cofounders of Otto, is convinced that the new technology will make life a lot easier for truckers, especially those who own their vehicles. He reasons that a new tractor-trailer costs between $160,000 and $200,000, and can be driven by a human for only nine hours at a stretch. But with self-driving technology, the driver can take as many naps as he wants while the truck drives itself. That doubles or triples the driver’s performance. “The truck is always productive,” Ron said. “They’re making more money, because they can use it more. They’re seeing their families more often, because they can finish their long-haul routes faster. And most important, they’re safer.”
“IT WOULD MAKE THE JOB NOT WORTH DOING”
But what Ron didn’t say was that many truckers who don’t own their own vehicle may lose their jobs. Wade Dowden, a thirty-two-year-old truck driver, told The New York Times that he doesn’t see much of a future in his line of work in the age of autonomous vehicles, because the pay will get progressively worse. The idea that automated trucks will drive themselves while human drivers sleep won’t work for most truckers, he said. “It would make the job not worth doing. Once you’re only paying a guy to drive the final miles into a city, we’re certainly not going to get a raise for that,” Dowden said.
In many countries, truckers’ unions will be able to hold off layoffs of salaried drivers for some time. And in some of them, governments—facing the threat of truckers’ strikes—may ban the use of autonomous trucks altogether. But how long will they be able to do that? If a country tries to outlaw autonomous trucks, how will it compete with other countries whose exports will become increasingly competitive thanks to lower transportation costs? And conversely, how will domestic industries in countries that prohibit autonomous trucks compete with cheaper imports from countries where transportation costs are much lower? Sooner or later, it will be all but impossible to keep this technology from taking over the trucking industry, let alone the streets.