On April 25, Huawei's recruitment WeChat public account showed that Huawei once again recruited talented teenagers from all over the world. There is no limit to academic qualifications, no limit to school. It is required to have special achievements in mathematics, computers, physics, materials, chips, intelligent manufacturing, chemistry and other related fields and aspire to become a technical leader. Huawei said it will provide a mentor, a global vision, platform and resources, and a 5+ times salary.
On April 26, at Huawei's 19th Global Analyst Conference, Zhou Hong, president of Huawei's Strategic Research Institute, made a very brain-burning speech, especially announcing 10 issues and challenges for the future.
(Zhou Hong, President of Huawei Strategic Research Institute)
In my opinion, this information all points to the same conclusion: Huawei's talent strategy is actively innovating in the new situation, attracting world-class talents with world-class problems, and coping with a better and more challenging future digital world.
World-class challenges attract world-class talent
Hu Houkun, Huawei's rotating chairman, explained Huawei's talent strategy this way: "At Huawei, when it comes to innovation, the first thing we think of is top talent. We hope to attract world-class talents with world-class problems to meet challenges together and promote scientific and technological progress. ”
(Huawei Rotating Chairman Hu Houkun)
He further explained Huawei's move to recruit "talented teenagers": "Huawei provides the world's most difficult topics, a powerful platform, and enough resources to support geniuses to explore. ”
Dr. Zhou Hong visualized world-class problems and proposed 2 basic science problems and 8 cutting-edge technology challenges at the Huawei Analyst Conference (see figure below).
Why is that? Because under the human yearning for a better intelligent world, digital technology will grow at a rate of more than a hundred times in more than ten years, which poses a great challenge to basic theories and sciences, and requires more world-class talents to overcome.
In Huawei's previously released "Intelligent World 2030", it has been predicted that 2030: by 2030, the total number of global connections will exceed 200 billion, at the same time, enterprise network access, home broadband access, and personal wireless access will exceed 10 Gigabit, ushering in an era of 10 Gigabit connectivity; by 2030, human beings will enter the YB data era, the global universal computing power will reach 3.3ZFLOPS (FP32), and the AI computing power will exceed 105 ZFLOPS (FP16). ), a 500-fold increase; by 2030, 390 robots per 10,000 manufacturing employees; electric vehicles account for 50% of all cars sold; and intelligent vehicle networking (C-V2X) penetration rates reach 60%.
Hu Houkun also pointed out at the Huawei Analyst Conference that continuing to strengthen innovation capabilities and firmly grasping the two major opportunities of digitalization, intelligent development and low-carbon development of human society in thousands of industries is the key to Huawei's future.
Such predictions and initiatives are correct and beautiful, but the reality is very bone-chilling, and the challenge is that most of the existing theories and technologies were proposed decades or even more than a hundred years ago, and cannot adapt to the future. Moreover, to achieve this, we can no longer rely on engineering innovation or innovation based on customer needs as in the past, and the whole logic of innovation needs to be reconstructed.
In a discussion with Huawei Academia Sinica last year, Ren Zhengfei, founder of Huawei, said that as a scientist who studies cutting-edge science and technology, there are two paths for you to choose in the future: one is to take the path of scientists, do theoretical research on the endless frontiers of science, and create new knowledge in the company's vision and hypothetical direction; the other is to take the road of experts and participate in our "pig killing" and "coal digging" with a "scalpel"... The commercialization battle.
Ren also mentioned that curiosity-driven basic research and applied research driven by business value may also be combined to create both scientific knowledge and business value. This is the "Pasteur Quadrant" innovation advocated by Professor Stokes of Princeton University in the 1990s, and the reason why last year's new Infinity Frontier Act proposed to transform the American Science Foundation into a science and technology foundation.
Both Ren Zhengfei's speech and Dr. Zhou Hong's speech mentioned a term: "Pasteur model", which represents Huawei's view on the integration of basic and applied research.
The Pasteur model represents the general trend of upgrading the scientific and technological innovation model. In 1945, Bush of the U.S. Bureau of Scientific Research and Development proposed the classic one-way transformation model of scientific and technological innovation, that is, scientific research is divided into basic research and applied research, with the goal of pure knowledge innovation and practical application, respectively. At the end of the 20th century, Stokes of Princeton University in the United States proposed in "Pasteur Quadrant: Basic Science and Technological Innovation" that basic and applied research are not opposed to each other, but have a certain degree of integration.
In discussions with Shanghai Jiao Tong University, Fudan University, Southeast University, and Nanjing University, Ren Zhengfei once gave many examples of basic acting skills to promote the application: a mathematical paper by Professor Arikan of Turkey, which became a "raging fire" of 5G ten years later; a paper published by Soviet scientist Peter Uffimtsev in the early 1960s that "diamond cuts can scatter radio waves", and 20 years later, the United States built a stealthY F22 fighter jet; in the 1950s, Professor Wu Zhonghua of the Chinese Academy of Sciences' ternary flow theory calculates the isoentropic section of jet engines, laying the foundation for today's aero engine industry.
Yes, in order to support the ten-year growth of digital technology, breakthroughs in basic research are needed, and to achieve this, co-creation is needed, which is why Huawei proposed to "attract world-class talents with world-class problems".
In the past few years, Ren Zhengfei has visited many universities and research institutes, which is to promote school-enterprise cooperation and let more talents join the breakthrough of basic research. "The unknowability of the future technological world is like a lighthouse in the darkness. The responsibility for lighting the beacon of the future undoubtedly falls on the university... We hope that in ten or twenty years, the universities on the mainland will shoulder the burden of catching up with the theoretical center of the world. ”
These are the backgrounds of Huawei's proposal to "attract world-class talents with world-class problems".
How can world-class talents overcome the world's problems?
How can world-class talents overcome world-class problems? Dr. Zhou Hong proposed four scientific hypotheses and business visions, namely to expand the boundaries of cognition; to expand the boundaries of perception; to explore new computational models and implementation methods; and to break through the assumptions of Shannon's law.
These four breakthroughs have one thing in common, that is, the logic of innovation needs to be changed, becoming based on assumptions and vision-driven, and becoming breakthroughs in theory and basic technology. For example, in the new innovation logic presented by Dr. Zhou Hong, there are five major links from assumptions and visions to theoretical, technical and business innovation.
Breakthroughs in cognitive boundaries require breakthroughs in physics, chemistry, biology and other fields. For example, compared with today's optical fiber communication, quantum communication has the advantages of absolute security, large channel capacity, zero latency, and strong anti-interference ability, and to achieve quantum communication requires a breakthrough in the basic science of physics. Dr. Zhou Hong gave an example, in order to better support quantum communication, it is necessary to achieve quantum storage that stores photons and quanta, which was unimaginable before.
I noticed that in March this year, researchers at the University of Geneva in Switzerland stored a qubit inside a crystal for up to 20 milliseconds, which laid an important foundation for the development of long-distance quantum communication networks. The team of academician Guo Guangcan of the University of Science and Technology of China also realized the basic link of quantum relay based on absorbent memory, and demonstrated the communication acceleration effect of multi-mode quantum relay. In the future, it is even possible to develop a quantum U disk similar to a classical U disk to help achieve remote quantum communication.
The breakthrough of the perceptual boundary needs to go from approaching human perception to transcending human perception, from alternative perception to extending and creating perception, from human perception to machine perception. It is also necessary to better perceive and control the human body itself.
Among them, biomimicry is a very important direction. Taking the frog's eye as an example, single photon detection can be achieved, compared to the human eye can only respond to about 10 photons. Inspired by this, it is possible to create a single-photon detection camera to see the dark forest with five fingers out of sight. The frog's eyes, another feature, can only see moving objects, and are blind to the still background picture. In response to this feature, the researchers designed an event camera that only records "changing" pixels, rather than recording all the information like traditional cameras, which gives event cameras the advantages of high frame rate, low power consumption, and high dynamic range. In the field of autonomous driving, event cameras have a huge advantage over traditional cameras.
The exploration of new computing models and implementation methods is to better understand the world, solve problems, and create value. "Can communication be extended from the traditional linear Fourier transform to the nonlinear inverse scattering transformation?" Can AI further develop mathematical logic calculation mode, geometric manifold calculation mode, game calculation mode, etc.? Can scientific computing develop adaptive and efficient computing models to create new architectures and components without being constrained by traditional computability theories and von Neumann architectures? Dr. Zhou Hong's question is profound.
The assumption of breaking shannon's law is even more significant for the future development of communication. Born more than seventy years ago, Shannon's law, which defines the limits of entropy, channel capacity, and compression under distortion, is the basic theory of communication. However, the rapid development of 5G technology has reached the limit of The Law of Fragrant, and many communication experts in the industry even say that "there is no G after 5G", and have approached the capacity limit in terms of point-to-point channel capacity. In the future development of communication technology, all that remains is the expansion of the number of communication channels. In his speech, Dr. Zhou Hong mentioned some possibilities for breaking Shannon's law, such as theoretically assuming that the world has prior knowledge and memory; in engineering, if high-gravity attosecond lasers can be made, millions of T of traffic can be generated... "If these technologies are grafted into the wireless and optical fields, can they improve communication performance thousands of times? ”
We also see the industry exploring new ways of information transmission and processing to break through the bottleneck of Shannon's law: the direction of computing and communication evolution extends in the direction of quantum computing, DNA storage, and semantic communication, using semantic communication to solve problems such as information transmission storage and efficiency, and using the characteristics of quantum computing to perform huge computing tasks in parallel.
Publish world-class challenges, recruit world-class talent... All of this makes us not pessimistic about future innovation, when world-class talents come together, they will form the greatest synergy and launch a charge towards world-class problems such as basic theory and science. Similarly, when the logic of innovation is reconstructed, starting from scientific assumptions and business visions, daring to break boundaries, anything is possible.
The clarion call to world-class problems has been sounded, and mankind's exploration of the "no man's land" of the digital world will also go deeper.