Yu Kai: Central computing architecture is the ultimate solution to chip shortages

From June 10 to 12, 2021, the 13th China Automotive Blue Book Forum was held in Hefei, with the theme of "First Move". This concept comes from Go, where taking the first move often means an advantage or even a winning position. For automotive competition, how to take the first move and have the first move mindset also determine victory or defeat.

　　
Behind the booming new forces, there is also a group of suppliers who have emerged with the help of the transformation of the automotive industry, covering various fields such as chips, batteries , and radars. Yu Kai, founder and CEO of Horizon Robotics, brought his own thoughts from the perspective of automotive parts supplier and chip R&D manufacturer.

In the era of full intelligence, the value of chips to smart cars is like the power of fuel cars. Data is the energy of the new era. Converting the energy of the new era into accurate decisions for users is computing power. In the face of competition, China's independent brand intelligence strategy faces three paths, namely relying on full-stack intelligent giants, vertical self-research, and ecological win-win.

The first path, if you rely entirely on intelligent giants, will eventually lead to mediocrity, loss of initiative, and being controlled by others. The second path is vertical self-research, which belongs to "unrepeatable excellence". The risk of self-research is huge. If you can succeed, you will become "Apple", but once a decision is wrong, all the previous efforts may be wasted. In contrast, the third path of ecological win-win is a success with rules to follow. In addition, Yu Kai also said that autonomous driving should be people-centered, not machine-centered. Therefore, human-computer interaction and autonomous driving should be inseparable to form full-scene vehicle intelligence.

"We believe that autonomous driving must be people-centric rather than machine-centric. Our goal is not to make machines more powerful, but to make people the masters of the car. The real value creation here is that human-machine interaction and autonomous driving must be inseparable and integrated."

Corresponding to this is the evolution of computing architecture, from software 1.0 to software 2.0. In the software 1.0 era, programmers traditionally programmed and used logic and processes to complete these functions, while in the software 2.0 era, data is driven.

In the future, both cockpit and intelligent driving will move towards the era of software 2.0 represented by neural network computing and data-driven, and the final outcome will be a central computer. The whole vehicle will have an AI middle platform, which is not only about intelligent driving, but actually calculates all the data of the vehicle body, road environment data, and human-computer interaction data in an AI middle platform.