"Around it?" Lin Youde didn't react for a while.

No matter which country's laws are very strict with traffic laws, it requires a lot of road tests before any car is launched.

As a representative of new high-tech, unmanned driving can also be applied for testing in China, but it is very difficult and the approval time is very long.

Zhou Yujie continued: "Since the road test is difficult to apply, we will not do it for now. We will simulate all the programs on the computer.

What I want to talk about next is a new logic for unmanned driving - pure visual unmanned driving!"

"Pure vision?" Zhong Guangwei asked: "Only rely on the optical video around it, not on various sensors?"

"Yes, except for a few necessary sensors involving safety, all other sensors used to assist unmanned driving will be thrown away." Zhou Yujie affirmed.

The core safety of driverless electric vehicles on the market today relies on a large number of signal sensors around the car, including microwaves, ultrasonics, lasers and optical sensors.

When these sensors are abnormal, the car will be quickly controlled to stop.

Theoretically, the more sensors there are, the higher the safety of the car, but the problem also comes. If there are too many sensors, the cost of the car is so high. For example, the most sensitive laser sensor among the sensors is RMB 5,000. If you install a few more, you can buy a Wuling magic car just by the cost of the sensor.

Cost has a great impact on the popularization of large items such as cars, and consumers' choices. No matter how good the quality is, but the cost is too high, they can only become one of the luxury cars and can only serve a few users.

If we only look at the car manufacturing thinking, this model is fine. The profit of each car is high enough, then the operational risks can be greatly reduced.

However, Internet giants have never built cars for a single profit, but for the whole people. What's more, with Tesla's existence, Tesla's desperate price cuts in the future will make countless other car companies crazy.

The strongest way for Tesla in later generations to reduce costs is to use sensors. Millions of Teslas are driving around the world, and every car is shooting videos. These videos will be uploaded to Tesla's servers for Tesla to improve its visual algorithms.

"Let the automotive chip simulate the human brain and analyze optical videos. If successful, it can indeed subvert unmanned driving. However, although the human brain is not big, its computing power and storage are much stronger than supercomputers. What do you plan to do about the chip? " Lin Youde asked.

"The chip is a difficult problem. We do not have the ability to manufacture, so we can only develop it ourselves and then outsource it to produce it." Zhou Yujie nodded and said, "But this chip is still different from a mobile phone. The size of the car is as big as that of a mobile phone chip. For temperature control, we can also make a cooling system specifically in the car.

However, it is still too far away to say that this algorithm is still a few years to make, even in the early version, and later testing and gradually becoming popular, let's decide this kind of problem until it depends on the situation in the future.

However, with DiDi, we can obtain a large amount of driving data in the early stage to carry out algorithm research and development. This is our advantage. Other companies may not be that easy to play this trick."

In addition to investing massive amounts of funds and a large number of talents, another important factor for pure vision algorithms is a large amount of driving video data. With this data, scientific researchers can slowly develop visual algorithms and then gradually improve them.

Tesla can do it because its sales will become larger and larger, but no matter how sales increase, it is difficult to compare with the scale of online car-hailing. Compared with Tesla, this should be Toutiao's only advantage.

"That's true, but we don't have the ability to develop chips at the moment." Lin Youde nodded and said.

Zhou Yujie said: "First set up a group, and then poach Chinese engineers from Tesla in the United States, and from Huawei and Samsung to poach chip R&D teams. I will take out $100 million separately to do the starting funds for this project. This year, we will build the team first."

The production cost of chips is so high. Without market support, a country cannot afford to burn this money. R&D is a little easier, but it basically starts with 10 billion or even 100 billion yuan.

For example, Huawei's Kirin chips have burned out more than 200 billion yuan in more than ten years.

After returning from rebirth, Zhou Yujie's headlines made by him competed with traditional Internet companies for the market, entered new industries such as takeout and online car-hailing. In fact, he valued that these industries made money very fast. By controlling traffic, he could generate amazing profits.

And these profits will enter real high-tech research and development in the future.

[American technology companies are studying the universe, biotechnology, new energy, artificial intelligence, while Chinese Internet companies are studying how to grab business with small vendors.]

Although this sentence is a joke, it is also a fact. Just talk about Baidu and Google. The former is studying how to sell advertisements, while the latter has been developing AlphaGo and trying to use high-tech algorithms to simulate and crack human protein genes.

Protein genes are an important part of humans' cracking of cancer and lifespan. Traditional methods can only crack thousands of groups in a year, and the total number of human protein genes is as high as 200 million. After Google started this project, it only took a few years to use supercomputer simulation to complete half of the task.

Other American technology companies are also developing their own technology. Even e-commerce companies like Amazon have begun to build rockets. Although they are following Musk, they have more or less contributed to technological innovation.

However, domestic giants are all studying how to seize user time, study user consumption habits, destroy intermediate suppliers, etc. These have certain benefits, but in fact, the core is the transfer of wealth. Whether it is takeaway, group buying, online car-hailing, etc., Internet giants are snatching the right to dominate the right to speak from offline industries and having certain innovations, but the essence is to grab cakes in a stock market.

Only technological innovation can create an incremental market. Like the physical industry, this should be the core of the development of high-tech enterprises.

Of course, we will not give up on traffic profits. After all, if you want to do high-tech research and development, you can’t do without money. The fastest way to make money is to subvert the mobile Internet to all industries.

Lin Youde said: "Mr. Zhou, if we recruit talents in this field on such a large scale, the news will not be hidden."

"This does not need to be concealed." Zhou Yujie said disdainfully: "Bat knows my purpose, but can't he dare to keep up? Besides, they can't know our true purpose!"
Chapter completed!