SMM News: in Intel's senior vice president and Mobileye CEO Amnon Shashua (Amnon Shashua), for Tesla, Uber and Lyft, the battle for the driverless taxi market is a matter of life and death.
If you have to listen to an executive at a taxi-hailing company talk about what the future of human drivers will look like with the rise of driverless taxis, Travis Kalanick (Travis Kalanick), the controversial former chief executive of Uber, would be a good choice. He currently holds management positions at two ride-hailing giants, Uber and Lyft, which have recently been cautious about the timetable for artificial intelligence to replace humans and the steering wheel. Or suggest that one day human drivers may be completely eliminated.
The drivers went on strike on the eve of Uber's initial public offering of (IPO), and Lyft's share price has continued to fall since it went public, making the uncertainty of drivers' income and working conditions a major social problem. But as early as 2014, Kalanick identified one thing, and without fear, that the future of driverless cars was coming and drivers would lose their jobs.
"look, this is what is happening in this world." At the 2014 Code conference, Kalanick explained to Uber drivers who could lose their jobs in the future, "if Uber doesn't go in that direction, it won't exist." The world is not always beautiful. "
Lyft announced last week that it had reached a partnership agreement with Waymo, Alphabet's self-driving car business, as it released its first financial report, which saw investors continue to sell its shares. Waymo had been suing Uber for years.
As Tesla struggled to find a truly profitable car roadmap, Elon Musk (Elon Musk) began promoting his company's potential to be a winner in a market valued at $500 billion. His argument is that Tesla cars will be transformed into driverless taxis by their owners. Musk said his electric car company should have 1 million cars next year that could serve as driverless taxis, from which owners should be able to make tens of thousands of dollars a year (an ambitious plan, but it is well known. Mr Musk's predictions often fail to come true on time.
The public tends to pay attention to the risks of self-driving-many people are uneasy about the deaths of Tesla and Uber test vehicles. But the economy of driverless car technology will largely determine the winner of driverless future. The race has already begun, and according to Intel's head of autopilot, it will focus on significantly reducing the economic costs associated with today's human driving.
Over the past six months, Mobileye has been conducting "very in-depth research" on what its leader, Shashuya, calls the economics of mobile travel-as-a-service (MaaS). Mobileye, an Israeli driverless company, was accepted by Intel in 2017 for $15 billion. Mobile Travel as a Service, similar to Software as a Service, Uber declared to investors during the IPO roadshow that it was a whole new industry. But the conclusions of the Mobileye study help bring mobile travel as a service back to reality: the success or failure of competition will be determined by a factor that has always dominated business competition: significant cost savings based on the current economic situation.
"the way to really change the mobile travel market is from taxi-hailing services based on human drivers. Into a driverless taxi service. " "currently, drivers account for 80 per cent of the cost of the economy," Shashuya said in a recent interview with CNBC reporter Jon Ford (Jon Fortt). Once you get rid of the driver and replace it with capital expenditure-the cost of the car, the technical cost can reach tens of thousands of dollars. As far as the discounts you can offer to your existing taxi-hailing business, this will revolutionize the entire market. Even with a 40-50 per cent discount on existing taxi-hailing services, it is still a viable business; it can achieve high profit margins. "
Shashua claims there is no breakthrough in the economy. "you can see what Uber and Lyft are doing. They made it clear that the next stage of their evolution was to have driverless taxis. Because they also know that such cars can change the entire market. This is not to say that we have suddenly discovered that it has the effect of changing the market. "
Two stages of cost change
Several of Lyft's founders have said they are chasing a market of $1.2 trillion and that their profit path is moving consumers to find it more cost-effective to take a taxi than to buy a taxi. "if you believe that Americans spend $9000 a year to own and operate cars that use only 5 per cent of the day, we can help you use your car more efficiently and save a lot of money, then taking a taxi is good for you." John Zimmer (John Zimmer), founder of Lyft, told CNBC on the same day as IPO in March.
The Mobileye co-founder said the cost of the autopilot system is currently estimated at tens of thousands of dollars, a reasonable capital expenditure investment for a company, but it is clear that it cannot be transferred to consumers of passenger cars. Shashuya points out that the premise of real cost reduction is a sharp drop in the autopilot cost curve, which will have two stages of change.
The first phase of change "dramatically" reduced the cost per mile, reaching the cost of owning a car per mile. "this is not a small change."
A second phase of change will come when the cost of autopilot falls tenfold to a few thousand dollars.
"passenger cars can't invest tens of thousands of dollars in technology because they need to be profitable."
Intel has its own market problems. The chip giant recently slashed its earnings forecasts, such as a weaker three-year outlook on an investor day on Thursday. The slowdown in the traditional PC processor business is one of the reasons Intel is increasingly interested in growing markets such as data centers and potential markets such as autopilot.
Build an autopilot system
Mobileye's system uses cameras, lasers and radars to guide vehicles, and its core component is the camera system.
"Auxiliary driving is a large-scale industry." Shashuya pointed out, "this involves tens of millions of cars." In a large-scale industry, you can't imagine the cost of a car imaging system exceeding a few hundred dollars. You can give a car a surround camera and a high-performance computing device at a cost of no more than a few hundred dollars. "
Of course, competition in the driverless car market will test more than just cost.
"in addition to building an autopilot system, you have to make mobile travel a service. What you need to do is route optimization and customer needs analysis-put driverless taxis in a strategic position so that when customers ask for service within a minute, the car will arrive at its destination on time. You also need to deal with peak hours, mixed fleet scheduling and remote control. " Shashuya said.
Humans will still have work to do. "at some point, the car may throw, so you need manual help. Either call the paramedics or do something else. " "at Intel, we will work together on building autopilot systems to develop and acquire all the essential elements to create mobile travel services," he said. We are already developing an autopilot system. For each component, it is either acquired, developed on its own, or implemented through collaboration. "
Mr Shashuya believes that industry consolidation will eventually leave only a small number of participants in the market, leaving only "single-digit participants" who can not only build technology, but also test it and work with regulators. Get these cars on the road.
"when your computer crashes, no one's life is at risk. I think this is the big difference between the artificial intelligence of autopilot and the artificial intelligence of advertising recommendation algorithms on social networks. "
Mobileye has released a formal security model and is working with regulators around the world. The company also plans to launch a joint venture with Volkswagen to launch a driverless taxi service, which will be launched this year in Tel Aviv, Israel, and will be commercially operational in 2022. The company has tested self-driving cars in Jerusalem.
"it's a very, very important sandbox because this launch includes a complete mobile travel service. This is a launch service, a commercial service. It's not a test. No driver will sit behind the steering wheel and we need to provide a full range of services. It will be a very important sandbox for us to set up this service. "
Shashua also pointed out that Israel provides an important alertness test for self-driving cars.
"We are testing those vehicles under very challenging conditions. Driving in Israel is challenging, and driving in Jerusalem is even more challenging because of the need to drive when many people do not obey the traffic rules. The streets are narrow and roundabout is not uncommon. "
The Intel executive realized the importance of security issues.
"if your smartphone crashes, it's no big deal. Just restart it. For a car, you can't afford to run the risk of collapse. You can't make mistakes; especially when you drive a self-driving car. " "for consumer products, your artificial intelligence needs to be more accurate than ever before," says Shashuya. You can't make a mistake. You have to prevent problems, because this is a matter of life and death. When your computer crashes, no one's life is at risk. I think this is the big difference between the artificial intelligence of autopilot and the artificial intelligence of advertising recommendation algorithms on social networks. "
"autopilot is a wonderful thing because it is really a very natural case study of artificial intelligence. It involves a variety of factors, including ethics; the problems you need to solve when you talk about artificial intelligence. It involves perception. You need to understand the world around you, understand the various pattern recognition, although we humans can do those pattern recognition effortlessly, but it is very difficult for computers. It involves decision-making ability; integrating into traffic is a decision, right? When you are negotiating with other road users, you are making a decision. Should I make way for that car? Should I go that way? It's this traditional decision-making ability. The advantage of autopilot is that it is a huge artificial intelligence problem, but there are huge business opportunities. It's a huge transformative business. "