After the general trend of self-driving cars was established, the new travel revolution has been launched and Ni Kai, who is also known as China's first self-driving pilot, has been involved in a large group of start-up companies that have actively lit the revolutionary fire. Multi-technology just completed ten-million-dollar Angel Wheel financing in mid-October, but he believes that autopilot financing is very hot, but probably about 2-3 years later, you have to see each skill, there is bound to be a wave in 2019 Industry eliminated, but he is not too worried, because Wo do more data companies, automatic driving is just the starting point.
Wo multi-technology founder Ni Kai will be transformed into a car business enthusiasm
After leaving LeTV, Ni Kai renovated and tested his car in a garage in Beijing. He started his 'Garage' business venture, and He + Duo was the dismantling of the word 'shift', pointing to the core of this autonomous driving revolution .
Why is not L3 or L4? He explains that 'we will definitely be L4, but L4 technology is still too far from mass production so it is not the first step to pursue.' According to the definition of SAE International, basically above L4, it can be said that full-automatic driving is realized, and human drivers do not need to pay attention to the external conditions and are handed over to the car for control, but because of technical difficulties and complicated implementation, Optimistic estimate is about 5 years after the road, the industry's objective view even takes 7 to 10 years. 'Too far things will have problems, because there is no business model before landing, capital and partners can not do it completely for the dream to do This incident, burned for too long will certainly be a problem, 'Ni Kai is very straightforward.
Why Level 4 does not have a business model at this stage, mainly because the technology is not stable, and key components such as laser radar still need to be certified.
Not long ago, Google's Waymo announced the official launch of a completely human-free taxi service at Chandler, a suburb of the US Sun City, becoming the first of many self-driving companies to launch the service. However, why not choose the oldest tested Silicon Valley? In fact, it is not hard to guess because Chandler's road conditions are simpler than those in Silicon Valley. 'Level 4 is now costly, as a demonstration, but the business model does not yet exist.'
L3 wins the key: than the first mass production
The biggest difference between L4 and L3 is the takeover of emergencies. When the emergency situation develops, the L3 car will give control back to human beings, and the time will come to a quick realization. For example, Audi announced that A8 By the end of 2017. Therefore, Ni Kai choose a pragmatic road, cut into the L3, but must add some material to make the difference with competitors, the second key is to put it to the car earlier than anyone.
Different from L3, what is 0.5 more? Has the function that the traditional L3 does not have.For example, the traditional L3 is when the car encounters the emergency, will return the control to the humanity, but Ni Kai thinks also needs to add the extra Functionality: In the event of an emergency, the car must have some emergency handling capability, despite the inability of the person to take over.
Another key to staying ahead of competitors is faster than mass production, with most of them cutting into L4, like Waymo, Uber and Nike, who think these companies are 'Party A' concepts and do their own operational services , L3 is from the 'B' to do, such as Wo multi or Tier 1 companies such as Bosch, Continental, etc. is based on service OEM car factory.For Google, Uber this type of business model, is to see who first landing , Low cost conditions, adaptability; and L3 is aimed at OEMs, compared with the first volume production.
Inevitably, start-ups play a role in accelerating industrial change in this auto-driving wave, but Ni Kai pointed out that auto financing was a fire during the year but, like the previous AR / VR, drones, and fire time About two or three years, then see what you do.Therefore, automatic driving from 2015, 2016, the rapid rise is expected in 2018, 2019 will be a round of elimination, is the Ebb tide, when the entire industry began to integrate, you can Can not be integrated into the survival of the key.
Alliance data sharing will be a norm
There are two important links for companies involved in the study of autonomous driving. One is the development and continuous evolution of the algorithm. However, the challenge of the algorithm is relatively low. In particular, the flow of international talents and finding good people can solve some problems. At the same time, The algorithm is still changing, the latter can directly target the latest algorithm into the market.But another link: the data, it is not so simple, not spend a sum of money can have, but also by collecting, labeling, such as the accumulation of more time Importantly, the data can be further optimized algorithms and models.
But because autopilot is a fairly complex area that requires each and every character in the industry to work together, sharing data to become an autonomous driving ecosystem has to be done. 'Data is all about wanting to be AI's core asset, from the largest Business interests, we must share, the future of the automotive industry is a union, 'Ni Kai stressed.
Indeed, there are many alliances that have emerged, such as the early appearance of the European Union Trucking Union, which brings together the six truck concessions DAF, Daimler AG, IVECO, MAN, SCANIA and VOLVO to test cross-border autochanger transportation In addition, Delphi, Continental join the BMW, Intel, Mobileye alliance, the other there are Benz + Bosch, Volvo + NVIDIA + Autoliv, which is a 'depot + Tier 1 + software / key components' mode.
Traditional car depot in the past do not pay attention to the data, to do autopilot can not escape with others, the establishment of alliances of complementary benefits more reflected in the data sharing, Ni Kai further explained that software can not generate data, the need for hardware, vehicles Operations, all the roles in the industry chain can come in, including sensors and other components, so 'sharing of data will be a normal'.
At present Wo Wo strategy in addition to their own data collection vehicle modified road, but also with the new cooperation of investors NavInfo collection, soon Wo will announce a cooperation with a Chinese depot, Wo multi-developed system into the car, is expected to soon It will be really commercial.
He said: "The car is the conduit for collecting entrance data. Our vision is a data company." The starting point is the subsystem of the automotive industry that accumulates data, but in the long run, there are many business models 'I even think autopilot systems can be free, and the insurance industry and the entertainment industry are likely to demonstrate more business models.'
Targeted 2 years window period, eat the Chinese market
Recently, investment and acquisitions in autopilot have been frequent, and Japan's SoftBank has just reached an agreement with Uber to directly invest $ 1 billion in Uber. Not long ago, auto maker Ford acquired ArgoAI, a startup that develops virtual driver systems, for $ 1 billion Delphi, a Tier 1 company, bought NuTonomy, a UAV startup founded by the MIT team for $ 450 million, mainly because autopilot is a combination of two areas: AI and cars, and software is a very important start-up. The traditional car spared no efforts Strengthen the scientific and technological strength.Facing the possibility of acquisition, Ni Kai to 'not rule out, it is also possible' to express its open attitude, in October just got IDG, NavInfo new investment, but he also revealed that the new round of Fund-raising is in progress.
In addition, Wo choose to focus on the Chinese market, Ni Kai calculations are two things: the Chinese market is difficult enough, but also large enough to form a barrier, China and Europe and the United States of the traffic conditions are not the same, the road situation is more complicated, At the same time, the production and sales volume of China's automobiles is the highest in the world. More importantly, "the past two or two years have been the process of educating the depot. However, the efforts of the European and American car factories are now concentrated in Europe and the United States. This gives a very good chance for start - ups in China. 'You must grab this window now or I will not start a business!
My experience at LeTV made me really understand the automotive industry
At the Tsinghua University, when studying master, Ni Kai invested in the study of autopilot. 'The software and hardware used at that time were different. The laser radar was only 1-line, and the industrial cameras were also black and white.' 'Obtained a PhD from the Georgia Institute of Technology Degree after joining Microsoft, is doing is considered as a secret project HoloLens. After joining Baidu, his life is inseparable from the automatic driving. In addition, Wo company more English name is HoloMatic, is curious if with HoloLens? The answer is no, but Holo means complete and complete. Matics means 'control' in the car, and full and safe control of the car is what Ni Kai has to do.
'AI has a lot of things to do in this world, but I am passionate about the car.' 'At Baidu I wanted to be closer to mass production. They wanted to be L4, but the road was long.' This passage also echoed the above point of view, he believes that L3 is able to mass production of products as soon as possible, so chose to start .For the previous experience in Baidu and music, he is more direct: 'LeTV experience to help me Than in Baidu, Baidu is doing AI, or like doing IT, but the real understanding of the car is in music.
Nike is mainly engaged in Baidu's partner is BMW, but when he entered the music, began a large number of people interact with the automotive industry, 'learning a lot of things are not industry knowledge, but to know how they do things, language of doing things, As well as the automotive industry how to look at the IT industry. 'In addition, he is based in Baidu technology research, music began to manage,' This is very helpful for my current business, capital chain, financing, 'he said.
Although autopilot is a popular topic, at this stage many people still like the traditional car to give the sense of control, despite trying to get rid of the joy of human control, Ni Kai still like to have a sense of driving the car, like 'BMW, Tesi Pull. "However, when the autopilot started commercial, the traditional car probably would be like a mechanical watch becoming a minority collection or highlight the taste of niche products, but during this time we all know not too short, he believes that autopilot in the future 2 to 3 years is not a mature period, but it will be a time to integrate with the entire industry. He must set the technology in the next two years. Therefore, with the title of the first person in China to drive autopilot, what Ni Kai wants to do now is , Automatically driving the fastest landing first person.