The Robots Are Coming!

by Andres Oppenheimer

  THE SELF-DRIVING PIZZA DELIVERY CARS

  In 2017, Domino’s Pizza began using experimental self-driving cars to deliver pizzas in Ann Arbor, Michigan. And in 2018, I saw these same cars on the streets of Miami. Through an arrangement with Ford, Domino’s began using several dozen self-driven Fusion sedans to gauge the public’s reaction to automated pizza deliveries, in hopes of starting that service commercially in 2021. The cars—with big signs that read “Experiment” and “Self-Driving Delivery Vehicle”—carried the pizzas in the back seat.

  It worked like this: You ordered your pizza through a cell phone app, received a code via text message, and when the delivery arrived, you punched the code into a tablet attached next to the car’s rear window. Then a speaker on the roof of the car confirmed your order, the back window rolled down, and the pizza was there—hot and ready—sitting in a box waiting for you to grab it. Why would people want their pizzas to be delivered that way? Among other things, because they won’t need to leave a tip, saving at least 10 percent of their order’s price.

  “It’s going to be a real learning experience,” Dennis Maloney, Domino’s chief digital officer, said. “No one really knows what’s going to happen when customers walk out to the car. They’re faced with a car. There’s no human interaction. What happens if they approach the car from the wrong direction? Will people mind coming out of their house? We want to understand all that.”

  In the beginning, all of Domino’s self-driving delivery cars had an engineer on board to keep an eye on everything, but the idea is that soon these cars will be on their own, a Ford spokesman told me. The experiment could pave the way for a possible new niche for big car companies such as Ford or GM, which could help them offset their expected losses once autonomous taxis become the norm and fewer people buy their own cars. Interestingly enough, the title of the New York Times article on the new trend in pizza delivery wasn’t focused so much on the future of Domino’s, but on the future of Ford. It read, “There’s a Pizza Delivery in Ford’s Future, by Driverless Car.”

  WHAT DOES THE FUTURE HOLD FOR CABDRIVERS?

  Up to this point, most headlines about self-driving cars have not focused on their potential for distributing packages or delivering food, but on their likely use as taxis. What will happen to cabdrivers once Uber, Lyft, and other ride-sharing services start using driverless cars? The taxi industry is one of the most labor intensive in big cities across the globe. There are no fewer than 38,400 registered cabs in Buenos Aires, 35,000 in Tokyo, 24,000 in Mexico City, 23,000 in Seoul, 15,000 in Paris, 13,400 in New York, and 7,300 in Washington, D.C. And that doesn’t include Uber and Lyft’s privately owned taxis.

  Already, even before the commercial introduction of driverless taxis, traditional cabdrivers have protested—sometimes violently—against human-driven Uber and Lyft cars. In some places, like Buenos Aires, traditional taxi drivers have managed to stave off the legal use of Uber for several years. But that’s an exception to the norm. In most large cities, the protests failed to stop the wave of private taxis. Sooner or later, market forces prevail over government regulations, and growing numbers of people will choose to moonlight as private cabdrivers to bring in some extra income.

  Traditional taxi drivers may be a species on the verge of extinction. “Uber and Lyft Have Devastated L.A.’s Taxi Industry, City Records Show,” read a 2016 headline in the Los Angeles Times. Since the appearance of Uber and Lyft three years earlier, the number of traditional taxi rides had plummeted by nearly 30 percent, the article said. In San Francisco, that drop was an even steeper 66 percent. In many states, traditional taxi drivers—who have to pay as much as $600,000 for a medallion—can’t compete with these new private taxi start-ups that don’t need to pay these fees and thus can offer much cheaper rides. The impact of ride-sharing platforms on the taxi industry has been “brutal,” the story said.

  THE FUTURE OF UBER AND LYFT LIES IN SELF-DRIVING TAXIS

  Some studies are less ominous for traditional taxi drivers. Carl Benedikt Frey—the Oxford economist who coauthored the 2013 study predicting that 47 percent of jobs would disappear over the next two decades—raised eyebrows in 2017 with a new study arguing that Uber’s impact on the transportation industry hadn’t been as significant as previously thought. The study examined rides by traditional taxi drivers in several U.S. cities between 2009 and 2015, and concluded that Uber’s net impact on wages was not dramatic: it caused a 10 percent reduction in traditional taxi drivers’ income and a 10 percent increase in Uber drivers’ income. As for actual jobs, the study found that Uber’s arrival hadn’t significantly reduced the number of traditional taxi drivers.

  But the study was retrospective and conducted during the early years of Uber and other ride-sharing services. In addition, it didn’t take into consideration the fact that Uber and Lyft are focusing on the development of self-driving cars, with which they may soon replace many—if not all—of their human-driven taxis. When I asked Uber CEO Dara Khosrowshahi what will become of his company’s drivers once autonomous taxis start hitting the streets, he told me, “I don’t believe that 100 percent of our demand is going to be able to be served by robots. There are always going to be exception cases, where people will want people to drive them.” And considering that Uber plans to grow exponentially as more people stop buying cars and use private taxis—the company currently accounts for 0.5 percent of all miles driven in the United States and projects its share to grow to 30 percent—the pie will grow for both automated and human-driven taxis, he said. In the end, the boom in ride-sharing services could even have “a net labor positive” impact, he added.

  But the fact is that Uber and Lyft are betting much of their future on self-driven cars, and that human drivers will most likely be—to use Dara’s word—“exception cases.” In 2014, Uber cofounder Travis Kalanick traveled to Pittsburgh and, in one fell swoop, hired almost all members of the robotics team at Carnegie Mellon University, one of the most advanced in the world. Overnight, Kalanick snatched about fifty Carnegie Mellon engineers. And in 2016, Uber announced with great fanfare its first fleet of experimental, self-driving taxis in Pittsburgh. At the same time, Lyft sold a chunk of its shares to General Motors in exchange for a $500 million investment to develop autonomous taxis. Most likely, human-driven taxis will begin to disappear in the early 2020s.

  THE FLYING TAXIS OF DUBAI

  In addition to Google, Apple, General Motors, BMW, Ford, Mercedes-Benz, Uber, Lyft, and several other companies that are developing self-driving vehicles, the Chinese corporation Ehang is already producing autonomous flying taxis. In 2017, Dubai bought several of these drones, which are powered by eight propeller engines and can carry a 220-pound person with a small suitcase, and began using them experimentally on predetermined routes.

  “This isn’t just a test,” Rohan Roberts, director of the Dubai Science Festival and host of the 2017 Global Innovation Summit, told me in an interview. “It’s a real service, one we’ll be putting into operation soon. People are going to be taking rides in flying cabs, and what’s even more amazing is that these electric drones will be able to carry a person as far as thirty miles at a top speed of nearly one hundred miles per hour.”

  But have these drones been thoroughly tested? I asked him, wondering to myself whether I would want to take a ride in an autonomous flying taxi. Roberts responded that Dubai had already tested these drones in more than a hundred flights with real passengers, without any problems. One of the main reasons why the flying taxis will be safe is that they will leave no room for human error, he said. “It’s a completely autonomous vehicle: the passenger sits inside the vehicle and has no control over it. The drones will be monitored remotely from a central command station, and at the first signs of any trouble, they are programmed to land immediately,” he explained.

  What if the passenger is drunk or falls asleep? I asked. “Again, the passenger won’t have any control w
hatsoever over the drone,” Roberts said. Asked about the possibility that parts of the drones will break loose and fall, hitting pedestrians on the streets, he replied, “Well, there really aren’t many parts that can just fall off. These vehicles fly at an altitude of a little under a thousand feet, and of course, as with any new technology, there are certain risks, so the important thing is to keep the drones well maintained. But the same could be said about airplanes: thousands of them are flying across the skies at any given time, and you could certainly make the argument that a piece of metal could fall from the sky. But they don’t, because they have good maintenance and are serviced frequently. The same goes for the drones.”

  When I asked him if flying taxis wouldn’t create havoc over city skies, Roberts told me that “at the moment, they are only going to be used for certain, predetermined routes, so they won’t be flying all over the city. We are starting on a small scale, and later will expand these routes to make the drones available to a larger number of passengers. As with any large city, Dubai has occasional traffic jams. That could make for an interesting experiment: to see how flying drones could help solve some of our traffic problems.”

  DRONES WILL DELIVER PACKAGES TO YOUR FRONT DOOR

  For several years now, Amazon, FedEx, UPS, and DHL, which account for a significant amount of the world’s package transportation business, have been developing drones able to deliver products to your front door. The skies above our cities most likely won’t become a whirlwind of unmanned flying vehicles carrying packages in all directions. Rather, shipping companies will use self-driving trucks with drones mounted on their roofs, from which the drones will take off to deliver packages to addresses far away from their standard driving routes.

  In fact, UPS estimates that if the use of drones could cut just one mile per driver per day, the company and its fleet of 66,000 trucks would save around $50 million a year. FedEx, on the other hand, has said that using autonomous trucks will be much easier than using drones, in part because automated trucks won’t have to comply with regulations that prohibit air traffic over densely populated areas. Either way, drones delivering packages will become increasingly common across the world. Walmart, America’s largest retailer, announced it will begin challenging Amazon in delivering packages with drones. But it won’t be using trucks. Instead, it will use blimps.

  In its patent application for blimp-style floating warehouses that would also serve as drone launching platforms, presented in 2017, Walmart said these airships would fly at heights between 500 and 1,000 feet, and their movements would be determined by which areas have the greatest demand for deliveries. “Movable warehouses are a really nice idea,” says Brandon Fletcher, an analyst at Sanford C. Bernstein, “because any flexible part of a logistics system allows it to be more efficient when demand varies widely. The e-commerce world suffers from highly variable demand, and more creative solutions are needed.”

  ROBOTS ARE TAKING OVER FACTORIES

  After several decades of failed forecasts that robots would soon be everywhere, these predictions are finally beginning to materialize. According to the International Federation of Robotics, annual sales of industrial robots tripled from 81,000 units in 2003 to 245,000 units in 2015. And annual industrial robot sales will triple again in the near future, reaching some 900,000 units by 2025, according to the tech consulting firm ABI Research.

  Why such a sudden explosion in industrial robots? It is mainly because robots are getting increasingly cheaper and smarter, and because the population in industrialized countries such as Japan and Germany is aging rapidly, which is driving up the demand for robots to replace their shrinking labor forces. Also, the steep increase in wages in China is forcing companies there to buy robots in order to remain competitive. In part because of its economic success in recent decades, China is no longer a cheap labor country, and Chinese businesses are now buying huge numbers of industrial robots to replace workers they cannot afford.

  Back in 2011, Terry Gou, founder of Foxconn, the company that produces iPhones and many other electronic products in China, announced that he would be buying a million robots over the next three years to supplement the million people he had working in his factories. Not only were Chinese workers too expensive, they also brought with them too many problems, he said. Gou was quoted by the Chinese news agency Xinhua as saying that “as human beings are also animals, to manage one million animals gives me a headache.”

  In Japan, industrial robot manufacturers are thriving. When I visited the Japan Robot Association, I was told that more than 70 percent of the robots produced in Japan are for export and that China is their number one client. Hiroshi Fujiwara, executive director of the association, told me that while China had 189,000 robots in operation in 2014, the association was projecting the number for 2019 to reach 726,000.

  It is believed that the word robot was first used by the Czech writer Karel Čapek in a 1920s play set in a factory where an android did the work of two humans at a much lower cost. Čapek initially called his work machine a labori, the Latin word for labor, but later changed it to robot, very similar to the word robota, which means work in several Slavic languages. In Čapek’s play, titled Rossum’s Universal Robots, one of the characters says, “Robots are not people. Mechanically they are more perfect than we are, they have an enormously developed intelligence, but they have no soul.” Čapek’s play ends—let us hope mistakenly—with the destruction of the world at the hands of the robots when the intelligent machines discover love.

  Today, in part because of skyrocketing robot purchases in Asia, robot prices are plummeting. According to a Bain & Company consulting firm study, while in 2010 the payback period in China for replacing workers was about 5.3 years, by 2016 it had fallen to 1.5 years. Industrial robots are also becoming increasingly efficient thanks to technological advances like sensors that let them “see” better and perform the more meticulous tasks that they weren’t able to do before. And many multinationals, encouraged by the declining costs of industrial robots, are moving their manufacturing plants from China to their home countries or to places closer to their markets, to reduce transportation costs and delivery time. Chinese companies and those in other major manufacturing nations don’t have many options: they either robotize their factories to remain competitive or go out of business.

  IN SOME FACTORIES, THERE ARE 20 WORKERS TO 400 ROBOTS

  It’s no coincidence that South Korea, Japan, Germany, and the United States are the countries with the largest numbers of industrial robots per capita in the world. According to the International Federation of Robotics (IFR), in 2015, South Korea had an average of 531 robots per 10,000 manufacturing workers in all industries, followed by Singapore with 398, Japan with 305, Germany with 301, the United States with 176, Spain with 150, France with 127, Slovenia with 110, the Czech Republic with 93, China with 49, Mexico with 33, Argentina with 16, and Brazil with 11.

  However, in certain specific fields, like the automotive industry, the percentage of robots is much higher. The IFR estimates the overall density of industrial robots in Japan’s automotive industry at 1,276 robots for every 10,000 workers. But according to what a couple of executives at Yaskawa—one of the world’s largest manufacturers of industrial robots—told me in an interview in Tokyo, there are some car plants in Japan that have only 20 human workers and 400 robots. “Right now, robots in car factories mainly perform welding and painting tasks,” Kei Shimizu, one of the sales managers at the Yaskawa robotics facility, told me. “Human workers are totally separate from the robots for safety reasons, and are mostly involved in engine and cable assembly, which robots aren’t so good at. But that’s changing very quickly. We figure that at least part of the work being done by people will be automated within the next five years.”

  Shimizu also told me that over the next decade or so, the biggest expansion of the use of robots will take place in the food industry. In factori
es, supermarkets, and restaurants, “the cost of labor is still low, but sooner or later we’ll reach a point where robots will be even cheaper,” he said. “In Japan, for example, we have an old tradition of having lunch with the food inside a wooden container known as a bento box. Today that box is filled by human workers, because for now the robots don’t quite have the dexterity of human fingers. But that’s changing fast. It won’t be long before robots take over the job of filling the bento boxes with food.”

  THE FUTURE OF ROBOTS IS NOW

  I doubt Shimizu was exaggerating, because robots are already everywhere in Japanese factories. A few years ago, in what you might call an exercise in science fiction, McKinsey & Company painted a picture of the factories of the future:

  Imagine you are a manager in a manufacturing plant in 2035. At your plant, injuries are virtually unheard-of. In fact, there are few people on the floor: a small group of highly skilled specialists oversee thousands of robots, interacting naturally with the robot workforce to produce goods with unprecedented speed and precision, 24 hours a day, 365 days a year. When a new product or design improvement is introduced, factory workers train robots to follow new routines, using simple touch-screen interfaces, demonstration, and even verbal commands. Most of your day is spent optimizing processes and flows and even assisting with product designs based on what you see on the factory floor and the data that your robots generate.