Even if autonomous driving is fully popularized, humans still need not worry about unemployment.
Editor’s note: This article is from the micro-channel public number “geek park” (ID: geekpark), Author: Zhao Zi Xiao.
Even if autonomous driving is fully popularized, humans do not have to worry about unemployment.
In June 2020, Didi’s self-driving online car-hailing opened a trial ride experience for users in Jiading, Shanghai. Didi has added a touch of “technology” to its travel landscape.
Although autonomous driving sounds like a cool black technology, Meng Xing, COO of Didi Autonomous Driving Company, has a clearer understanding of this technology. From the perspective of technology or safety, autonomous driving will not become a scene in science fiction movies in the short term. He prefers to look at autonomous driving from a safe and comfortable perspective, rather than prioritizing how advanced his own technology is.
The style of Didi’s autonomous driving is the pursuit of safety and comfort｜Geek Park
As for autopilot technology that has approached “old driver” technology, will real drivers be unemployed? As the closest connection to the driver, Didi does not think this will happen. At the “Geek Park x Douyin | Innovation Conference 2021” site, Meng Xing mentioned that drivers and autopilot safety officers will still exist in the future for a relatively long time. Some scenarios may always require humans to drive, and autopilot development In the process, new occupations based on autonomous driving will continue to emerge, making machines smarter.
As has happened time and time again in history, the purpose of machines is not to replace people, but to make things more valuable.
The following is a transcript of Meng Xing’s speech, edited by Geek Park:
Hello everyone, I am Meng Xing of Didi Autonomous Driving. I am very happy to come to the Innovation Conference 2021 today to discuss the future development of autonomous driving dialectically with you through online live broadcast.
First of all, “the road-traveling technology deeply transmitted by old drivers” is an interesting interpretation of Didi. The term “old driver” also has a lot of meaning in the current network environment. I believe I don’t need to explain too much. The netizens who are watching the live broadcast all understand themselves.
What style is Didi’s autonomous driving?
In 2020, Didi’s biggest breakthrough in autonomous driving is to run a pilot route for autonomous driving in Shanghai. So for Didi, the key to the interpretation of the term “old driver” lies in the experience we ultimately hope to give passengers.
Let me give you an example. Today, when many companies are doing autonomous driving, they show that the advanced way is to quickly jump in the queue. But in turn, let’s think about what kind of drivers will do this. ? If I, as a boss, hire a driver, I would never ask him to go to Gasser to jump in the queue to squeeze others in order to grab those few seconds, but I hope he can have an accident within a few years of serving me Don’t even go out, so I can have a coffee and read a file on the road. Therefore, we hope to provide users with a more comfortable riding experience, so that users feel safe, calm, and not surprised.
At Didi, safety first is the highest rule we adhere to in our work, and it is also our starting point for autonomous driving. We have some rigid requirements for our self-driving vehicles. First of all, we should reduce accidents from the rules. The car must abide by the traffic rules 100%; and in the case of others not complying with the traffic rules, the car can still use the way to comply with the traffic rules. Deal with other people’s non-compliance with traffic rules.
Secondly, from the reduction of user experience, don’t let users feel the rush or frustration when driving. On the premise of protecting user safety, try to let users perceive all the dangers of this car, and deal with them. Some of the dangerous behaviors can be represented by the self-driving coping behavior of the car itself, and some can be represented by some interactive displays in our car. This is what our current technology has achieved.
And this sense of peace of mind and safe experience is “controlled” by our operating monitoring center. First of all, we can see the complete map of the operating area. Each car will be displayed in real time on the map. The car will have its starting and ending points, order data, how much electricity is left in each car, whether there is an alarm, whether there is an abnormality in the sensor, and then when the car encounters a special situation, the monitoring security officer can check it through the front camera of the car And to help the car solve the situation encountered; at the same time, you can see the data returned by the vehicle-road coordination device we deployed on the map, and see all the cars passing by each intersection from the perspective of God, including the traffic in each lane Situation, the average speed of this line.
There is also the supply-demand balance diagram we mentioned earlier. On the map, we divide this area into many small hexagons. The color of the hexagon representsIn view of the supply and demand situation at this time and in the next 15 minutes, if it is red, it means that the supply is less than the demand, and the green supply exceeds the demand. This will change from time to time. It is these one-by-one databases that constitute our “magic weapon” for providing passengers with better services.
Safety is the starting point of Didi’s autonomous driving｜Geek Park
But the goal of Didi’s autonomous driving is more than that. We have a longer-term and grand vision. One step further, in addition to driving safety in autonomous vehicles, in the event of a traffic accident, turning passive safety into active safety is our goal in the future. For example: if I brake suddenly and I will inevitably have to rear-end, I can predict this event earlier than a human driver, my airbag can react early, and the seat belt can be tightened in advance to prevent it from appearing. The kind of situation where the airbag has not come out after hitting the airbag or the safety belt is still loose after hitting half of it.
These may be the imagination that our so-called old drivers can bring us. So when we talk about old drivers, the dimension we are talking about has become deeper and deeper. It turned out that it was just that the car drives well and drove steadily. Now we hope that the old drivers can also take on the role of emergency ambulance, bodyguard, The assistant waits for these tasks. This is the experience we hope to deliver to passengers in the end.
At this time, some netizens may ask: Is there a possibility of a customized driver style function in the future? You can choose a calm old driver or a “fast” old driver. My answer is, I can guarantee In the case of safety, the possibility of customized driving experience will not be ruled out in the future, but I think the most urgent task now is to use one of the most universal and reliable “old driver” experience methods to go online to ensure that more people can experience it. The convenience of autonomous driving only then developed to the stage of personalization. As to whether this service can be provided, it ultimately depends on the technology at that time and everyone’s preferences.
After technology is popularized, people will become more valuable
In the future, when autonomous driving is fully popularized, everyone will inevitably have concerns about drivers. How to solve the employment problem of drivers in the future? Where will they go? First of all, what I want to say is that when everyone thinks about the profession of a driver, they often objectify the driver and turn him into a certain driving behavior. The driver is not actually a machine that controls the car. The driver can actually do a lot of things.If he can think, he will make different decisions based on changes in the environment.
I used to talk about a real case that happened on the platform. A driver bumped into a pregnant woman and took a taxi to the hospital. The driver discovered that she might be born immediately. This is not a standard pick-up behavior. First of all, you You should consider driving him to the hospital. At the same time, you may drive more urgently than normal, and you may even do some forgivable violations in it, and you may have to notify the hospital in advance. Inform the staff that this matter is particularly urgent. After you arrive at the hospital, you still need to take the pregnant woman out and give it to the doctor. This matter is not within the scope of the automatic driving that we are considering today, so What we humans provide is not only a gesture of hand waving, but also the thinking and temperature behind this gesture.
The second is that autonomous driving itself will bring some new professions. The machine is essentially a bunch of children who have never driven a car. He needs a teacher, a trainer, and a class teacher to regulate and restrain. They will become their head teachers, and humans will gradually hand over their own experience and abilities to it. At the same time, different people may have different behaviors and experiences, so we will generate many new professionals, such as labelers and safe driving testers. , Remote monitoring service personnel, etc.
In addition, the function of the online car-hailing driver is not just to drive. When the car is out of power or dirty, autonomous driving technology cannot solve the problem of car cleaning. At this time, the driver needs to take on the responsibility of the cleaner Therefore, drivers only changed their professional status. With the popularization of autonomous driving, many other new occupations that need people to complete will be derived in the future.
As autonomous driving technology matures, human drivers will get more value instead｜Geek Park
Third, even if the autonomous driving technology is very mature, I think some scenarios may never be completely solved by the autonomous driving technology. At least today and tomorrow are definitely not, and I think there will be some scenes that will always be opened by humans, because the evolution of technology today is not completely imitating the evolution of the human brain. There is no way to say for sure from the first principles. What the machine can do is only through the improvement of computing power and algorithm, it will definitely reach the same standard as humans, and deal with any random situation in the world. We can only say that in our clearly defined scenariosAchieve safe driving.
If this is done, the final result is that the machine will open some scenes, and the human will open some scenes, but when the scenes that people open, the scenes that the machine cannot open, people become extremely valuable, and humans serve The premium of those scenes will also become very high, and the income that humans can get will also increase.
The fourth is an interesting little discovery of my own. Since the first industrial revolution 250 years ago, mankind has experienced countless breakthroughs in productivity brought about by major technology in history. A large number of machines have entered human world. However, the proportion of human employment has basically not changed. The only thing that has changed is that the jobs that humans are engaged in have become more and more decent and more valuable. So when we are facing a technological revolution, what machines do is not to replace people, but to make these things more valuable.
For example, like an annotator, our car sees many scenes every day, some are head scenes, and some are long tail scenes. The car is recognized by sensors, but the sensor feedback to the car is some pixels or some point clouds , These things can be understood by people but not by cars. The car needs people to tell him that the combination of these pixels is a car or a person. Only when the machine gets enough labeled data, the data enters the training platform. Only machines can learn to iterate, upgrade their identification capabilities, and deal with more and more complex situations. Therefore, people need to use their experience to label and verify each other. This requires a large knowledge system. This knowledge The system depends on experienced people to see these data and turn it into structured knowledge that they can understand. The so-called, only old drivers can bring out old drivers.
For example, like a safety officer, in the future, we think that a safety officer will always exist. Maybe he will walk from the inside of the car to the outside of the car, to our operation monitoring center, and let our old drivers move towards The position of a security officer in the cloud. He can turn on the camera or sensor to see what is happening in reality, tell you what to do, and even handle some emergency situations remotely. These will be new jobs in the future. And positions, their existence will make machines become smarter and smarter.
Technology creates beauty. As the theme of this conference, it is a pleasure for us to bring changes to everyone’s lives through technology. But our goal has never been one step to the end. For the “impatient” netizens, I also want to state here that it takes time to settle before we go out to be able to hit the autopilot and complete the necessary stages. , First of all, from partial popularization to realization of carrying passengers; security personnel gradually walked from the car to our operation monitoring center, and then to our operation monitoring center. From one person to manage one car to multiple vehicles, this is In a gradual iterative process, Didi is moving forward steadily at every step, striving to make this beautiful vision come true as soon as possible and become a reality for everyone.
Thank you, I’m working hard for everyoneMeng Xing who created the “old driver” riding experience.