On October 25, the Xinhua News Agency reported on the "14th Five-Year Plan," mentioning that Chinese modernization relies on the support of technological modernization. Strengthening original innovation and key core technology breakthroughs, promoting the deep integration of scientific and technological innovation with industrial innovation, advancing the development of education, science and technology, and talent in a coordinated manner, and deeply advancing the construction of a digital China... From this, we can clearly perceive the methods and paths to accelerate high-level technological self-reliance and strength.
In today's world, a new round of scientific and technological revolution and industrial transformation is accelerating. Artificial intelligence is undoubtedly an important driving force. Self-developed high-performance chips and operating systems, artificial intelligence large models that empower hundreds of industries, robots that greatly improve production efficiency... On this new track, Chinese enterprises are attracting global attention with outstanding performance.
At the time of the release of the "14th Five-Year Plan," Guanchazhe.com specially invited Sun Xi, deputy director of the Department of Enterprise Management at Capital University of Economics and Business and associate professor, to share his views. His research focuses on industrial upgrading and technological innovation. This time, he discussed issues such as China's industrial revolution, the development of cutting-edge technologies, and innovation pathways.
Guanchazhe.com: In this "14th Five-Year Plan," there is an elaboration on the "bottleneck" issues. You previously believed that enterprises are the main body of scientific and technological innovation, and you also affirmed the vertical integration strategy of enterprises. Currently, facing the "bottleneck" fields with high interdisciplinary complexity, such as high-end chips and industrial software, is the vertical integration strategy still applicable? How should we understand it? What attitude should the government take to help and guide enterprises to tap into their potential?
Sun Xi: In recent years, due to drastic changes in the international situation, our country's innovation system has made some adaptive adjustments during this process. The two most prominent adjustments are: one is increasingly emphasizing the leading role of enterprises in scientific and technological innovation, and the other is emphasizing the construction of national strategic scientific and technological forces to serve national strategic needs. Combining these two aspects is an important hub, which is through national major scientific and technological projects and other forms, transforming specific "bottleneck" problems encountered by enterprises and industrial transformation into national scientific and technological tasks and projects.
For example, a specific research institution has been invested in massively to tackle photolithography resins and photolithography machines. Another example is the establishment of industry-specific institutions to tackle common industry technologies, helping most enterprises, especially small and medium-sized enterprises, to make up for technical upgrade shortfalls.
This kind of inertia in recent years may lead people to believe that an important form of the government helping and guiding enterprises to tap into their innovative potential is to turn the difficulties encountered by enterprises and industries into specific projects when they face difficulties.
This form has its role, but this role is mainly remedial in nature. The so-called remedial meaning is that through this mechanism, an inevitable time lag is created: from the time when enterprises truly encounter bottleneck problems, to the time when these problems are transformed into national projects (project applications), and then to the time when project results are generated (project acceptance and delivery), there must be or delay several years (because our national science and technology projects are planned and implemented on an annual basis).
In the long term, over-reliance on this form will inevitably lead to some problems, such as misjudging the supporting capacity of existing industrial foundations for solving bottleneck problems. Many times, we do not lack technological breakthroughs, but due to certain non-technical factors - such as communication barriers between supply and demand sides and other social reasons - these technological breakthroughs are not better utilized at a specific period. A typical historical lesson is that the Wuxi Oil Pump and Injector Research Institute once developed a high-pressure common rail system for diesel engines, but Chinese diesel engine factories did not use it.
Because at that time, they trusted the German Bosch and Japanese Denso technologies more. This phenomenon of local innovation trust bottlenecks was called "later disadvantage" by Professor Gao Xudong from Tsinghua University. If we ignore this issue, we are likely to overestimate the role of government science and technology projects in innovation and transformation, or lead to a generalized approach.
In other words, when we discuss today the combination of a capable government and an effective market to promote industrial innovation and transformation, we should first try our best to design and regulate the national unified market well, fully utilize the market mechanism, and fully mobilize the technological accumulation in various fields of the existing industrial foundation and all kinds of social forces. This is an important aspect of the government's "capable" role; moreover, in the long run, this is more important than simply pushing science and technology breakthrough projects.
Guanchazhe.com: You believe that enterprises should be the main body of innovation because only enterprises can focus more on the market and understand the needs, but there may also be a gap between real market demands and real social needs, how do you view this issue?
Sun Xi: To understand the above question clearly, we can better understand the effective ways for the government to help and guide enterprises to tap into their innovative potential. Looking far ahead, when we overcome some key "hurdles" now, the government should spend more time and energy with top domestic enterprises, top engineers, and strategic scientists to jointly discuss and understand the possible directions of industrial development.
Through this communication process, the government not only builds its own industrial understanding ability, but also "self-discovers" key technological nodes and technological system development paths that have significant implications for the transformation and upgrading of the industrial system and the realization of national long-term strategic goals, thus defining a series of significant future products and future engineering (such as the Xinjiang photovoltaic and Yarlung Zangbo Hydropower projects in recent years are typical examples of this way of thinking), rather than relying on the previous way of following Western definitions of "bottleneck" areas. Because history and practice have proven that this way of thinking will eventually lead to being beaten.
Based on this "self-discovery," the government further plays the role of a market rule designer (including standard designer) and a key buyer, using a more institutionalized way to continuously raise industry entry barriers and promote enterprise survival of the fittest; at the same time, using this widely announced way to create a common long-term vision for all industry participants, breaking down information barriers between industries, and guiding and mobilizing more national enterprises to participate. (The large aircraft project had such a wide announcement effect: the national strategic declaration attracted participation from many industries such as machinery, steel, textiles, electronics, and software.)
The measures mentioned above can effectively bridge the gap between the real market demands faced by enterprises and the real social needs required by the country, i.e., the so-called "temperature difference" problem. (In other words, the essence of the current temperature difference problem is the insufficient ability of national governance and market design.)
Guanchazhe.com: Under the context of the release of the "14th Five-Year Plan" and the emphasis on technological self-reliance and strength, how does the advantage of corporate vertical integration manifest? Are there any technical breakthrough cases you can share?
Sun Xi: More importantly, further clarify the relationship between industry-university-research cooperation in the above-mentioned process. Let those enterprises who want to participate have the qualifications to become the "question setters" of scientific and technological innovation, let more innovation subjects become better "question setters," thereby enabling enterprises to find suitable university and research institute partners to solve the difficult problems encountered in the innovation process in a timely manner, instead of going through national projects in circles.
This is the fundamental way to achieve diversified scientific and technological investment, and also an important means to break the recent alienation of national research projects. In other words, in this path, national projects should no longer be a role of "snow in a cold winter," but should play a role of "adding flowers to brocade" more often.
Take an example, in recent years, President Xi Jinping has mentioned in multiple occasions that we have made significant breakthroughs in the field of high-end medical imaging equipment. How did these breakthroughs happen? The source was actually in 2008, when Mindray Medical found the Shenzhen Institute of Advanced Technology of the Chinese Academy of Sciences, hoping to develop a locally produced high-end color ultrasound device. At that time, Philips, General Electric, and Siemens monopolized the global high-end color ultrasound technology, and high-end color ultrasound almost became a life-or-death battle for the entire Chinese medical equipment industry. At that time, young researcher Zheng Hairong from the Advanced Institute, who is now vice president of Nanjing University, participated in this industry-university-research cooperation project, and later became one of the most important leaders in this field. Mindray Medical also made a very successful comeback in the market.
Mindray Medical related technical breakthroughs
Later, when developing domestically produced magnetic resonance equipment, United Imaging Healthcare was also in Shenzhen, and established a cooperative relationship with the Shenzhen Institute of Advanced Technology during this period (although later United Imaging chose Shanghai over Shenzhen). In this process, the national projects undertaken by Zheng Hairong and the Shenzhen Institute of Advanced Technology related to high-end medical imaging equipment and instruments were almost all delayed compared to the industry-university-research cooperation projects. This is the embodiment of the main position of enterprise scientific and technological innovation, and this is the role of national science and technology projects as "adding flowers to brocade."
Guanchazhe.com: Then, can the current "bottleneck" situations in the high-tech field all be alleviated through the strategy of vertical integration?
Sun Xi: Specifically, for the so-called high interdisciplinary complexity "bottleneck" fields such as high-end chips and industrial software, there are still some industry differences. Relatively speaking, the chips used in the industrial field can be innovated in different degrees by the users. Hisense entering TV chips, Gree entering air conditioning chips are reflecting this logic, they are precisely because there are no suitable local suppliers, so they have to solve these internal needs through integrated ways.
Compared to this, the general computing chip industry is a larger market size but a higher degree of industrial division. The relationship between pure design companies (such as HiSilicon) and pure foundries (such as SMIC) has already moved beyond the past purely transactional relationships, and is increasingly regulated by external factors (including institutional and policy factors). Therefore, this field has gone beyond the scope of vertical integration issues.
By comparison, the industrial software industry is more like industrial chips, with higher fragmentation, smaller demand, and these software carry the key technical experience of leading industrial enterprises. Therefore, we can see that in recent years, the development of China's industrial software industry has shown two very obvious paths: one is represented by China Petroleum East Geophysical Prospecting, through vertical integration to solve key technology softwareization. The second is the recent industrial software industry's supply and demand side matching and transaction mediation promoted by the National Defense Science and Technology Industry Commission and the military industry association.
In short, these two paths have moved away from purely market transaction relationships, and instead solve technical barriers and trust bottlenecks through more organized or social ways.
Guanchazhe.com: In the current technological revolution, the competition is the fiercest and the impact may be the deepest, which is the artificial intelligence technology revolution. Combining the relevant statements in the recently released "14th Five-Year Plan," from the current perspective, what do you think China and Chinese enterprises should do to participate in this transformation of the AI landscape?
Sun Xi: A few days ago, I talked with the comrades from Zhejiang, and they said that Zhejiang is currently actively promoting the widespread application of artificial intelligence in the industrial field. After comparing, we found that the best in this field is Zheda Zhongkong. Why? Because since Mr. Chu Jian started his business thirty years ago, Zhongkong has accumulated very rich industry experience in industrial automation, and today it is almost the absolute leader in the automation control of China's process industry.
This allows Zhongkong to maximize the use of "exhaustive method" (training artificial intelligence with relatively orderly experience knowledge from industrial scenarios, ensuring that it establishes the basic judgment of "if A, then B", rather than allowing large models to automatically reason) to improve its artificial intelligence solutions and industrial operating systems.
More importantly, according to the current level of development of artificial intelligence technology, exhaustive method is still the most important way to use artificial intelligence technology to manage industrial scenarios, without any doubt. Compared to this, some enterprises from the internet and IT background doing artificial intelligence industrial applications are not as skilled as Zhongkong. Because they don't have the industrial experience base to implement the "exhaustive method."
This case actually points out some basic principles for understanding how Chinese enterprises participate in the AI transformation.
The first principle is that we need to clearly define that the application methods of artificial intelligence technology revolution are completely different in different industry fields. In those 2G, 2C industries, the requirements for safety and consistency are relatively low, and the variables in the scenario are relatively fewer. At this time, we can use large model technology (after appropriate migration) to solve some industry application problems. This is also the field where internet and IT background enterprises can deeply participate and achieve certain success. Cars, home appliances, personal computers can be the first to introduce large models, which is this reason.
By contrast, in 2B industries, the requirements for consistency and safety, and the diversity and complexity of variables in the scenario are much higher than in consumer goods and public utility scenarios. At this time, it is extremely dependent on the existing experience accumulation of industrial enterprises to reduce and decompose the scenario (such as forming a set of highly specialized mechanistic models), decomposing it to a level that artificial intelligence can understand and operate. Therefore, the industrial experience accumulated by Chinese enterprises over the past 30 years participating in global division of labor is absolutely not a burden, but the "golden key" to our participation in the artificial intelligence technology revolution. Without this "golden key," we cannot open the door to artificial intelligence-driven industrial development (design/development/production/marketing).
The subsequent second principle is how to view the relationship between traditional industrial enterprises and artificial intelligence enterprises. The operation scenario and experience data come from the accumulation of traditional industrial enterprises, and the role of artificial intelligence enterprises (or some researchers in universities who are good at artificial intelligence R&D) is to help traditional industrial enterprises convert these experience data into mechanistic models and AI. If this master-slave relationship is handled improperly, it will cause many problems, such as the communication barriers between supply and demand sides, and in serious cases, it will result in "dead end".
At this point, the state may need to do some things: for example, by setting up demonstration projects in the field of industrial artificial intelligence, creating suitable opportunities and places for collaborative innovation between supply and demand sides, and solving the trust bottleneck problem. Ultimately, the process of using artificial intelligence to develop new quality productivity should focus on handling the relationship between "new" and "old".
Guanchazhe.com: You have analyzed the "blue sky research" from 0 to 1, believing that only by taking task orientation and responding to demand as the driving force, and taking technological scientific research as the "hub," can engineering and technological development and natural scientific research achieve complete temporal and spatial unification, and the coordinated development can truly take root. How do you view the relevant expressions in the "14th Five-Year Plan" about basic research?
Sun Xi: First, we need to clarify a basic judgment, that is, as practice develops, the central committee's understanding of basic research is becoming increasingly profound and practical. We can look at the book "On Scientific and Technological Self-reliance and Strength" published in 2023, which includes speeches by President Xi Jinping at the two academies' meetings in 2014, 2018, and 2021, as well as his speech at the scientist symposium in 2020. In these speeches, President Xi Jinping's discussions on basic research (and its goal orientation, demand orientation), as well as the relationship between basic research and applied research, fully prove this point.
From the situation in the past two years, whether it is President Xi Jinping's speech on the issue of basic research on February 21, 2023, or his speech at the two academies' meeting on June 24, 2023, including this "14th Five-Year Plan," all reflect this pragmatic character.
In the discussion on the topic of "improving the overall effectiveness of the national innovation system" in the past two years, the central committee has made a very important change in its discourse, which is increasingly emphasizing the deep integration of scientific and technological innovation and industrial innovation. The two academies' meeting this year also changed the term "technology transfer and application" that has been used for decades to "technology transfer and application." All these changes are efforts within our traditional policy language system to find alignment with the new policy paradigm of "the main body of enterprise scientific and technological innovation." This is of great significance for us to understand how to solve the problem of disconnection between basic research and industrial needs in the new era.
For example, the "14th Five-Year Plan" proposed the requirement to "strengthen the strategic, forward-looking, and systematic layout of basic research," which is essentially a continuation of the spirit of the "221 speech" in 2023. An important spirit of the "221 speech" is to emphasize that basic research should adhere to both goal-oriented and free exploration "two legs walking." And today's "strategic, forward-looking, systematic" requirements are a further deepening of the goal-oriented approach.
Guanchazhe.com: Then, how to avoid the disconnection between basic research and industrial needs, so that universities and research institutions can really serve the breakthrough of technical bottlenecks with demand?
Sun Xi: Further, only when officials, industries, and academia can sit together to jointly understand the development direction of various industries, and even the entire modern industrial system, can the effect of this strategic, forward-looking, and systematic layout be maximized: at this time, the nation that leads the direction of social and economic development plays the role of the "question setter" of basic research, and only when the nation has a clear understanding and grasp of the "direction of development," this "question setting" by the nation is operable and is most useful. However, from the current assessment methods for "organized research" in universities, it is evident that there is still a need for further deepening of this understanding.
Certainly, the innovation process is highly uncertain, so it is difficult to "hit" all major basic research topics 10-20 years in advance, and the nation cannot be the only "question setter." At this time, we need to better understand the main role of enterprise scientific and technological innovation. For example, I previously mentioned the case of the Shenzhen Institute of Advanced Technology of the Chinese Academy of Sciences. The institute is a typical representative of the new research institutions established in the early 21st century in China: its development positioning, the strength of the national team's scientific research, and more importantly, its rooting in the industrial soil of Shenzhen and the Pearl River Delta.
My observation of the institute began in 2018, and I have a summary of its practice in promoting the deep integration of scientific and technological innovation and industrial innovation, which I call the "tree model": scientific and technological innovation and basic research in the institute are the roots, industrial development is the canopy, and the new products made by enterprises are the fruits, while the deep integration is supported by the trunk. The function of the trunk is not only to transport water and minerals absorbed by the roots to the canopy, but more importantly, to transport organic matter absorbed by the canopy back to the roots, thus giving the roots the energy for development.
And in the institute, the reason why the researchers are willing to accept the energy (demand knowledge and funds) transmitted by the canopy (industry side) is that the evaluation mechanism for the researchers includes the part of industrialization. You not only need to publish papers, but also understand industrial language, and be able to do horizontal projects, and be able to transform industrial needs into scientific questions to study "useful science."
When goal-oriented basic research can simultaneously manifest as two pathways: "top-down" and "bottom-up," we need to carefully evaluate some new trends that have emerged in the science and technology system reform in recent years. For example, scientific research should "serve the major national strategic needs," and the scale and evaluation of "major" should be placed on what scale? When the trend of pursuing English journal publications has not yet completely faded, the national strategic scientific and technological forces indeed need to focus on their main responsibilities and roles, and concentrate their efforts to strike.
However, if the Chinese Academy of Sciences gives the definition of "main responsibilities and roles" entirely to its sub-institutes or specialized institutes, the narrow understanding of the main responsibilities and roles may conflict with the disciplinary structures formed by each specialized institute historically, even resulting in the problem of "you cannot plant beans in a potato field (specific specialized institutes' main responsibilities and roles are limited to field A, but specific project groups cannot plant beans in field B to serve the national strategic needs)."
Moreover, for example, the collaboration between scientific research institutions and industry leaders in the frontiers to layout seed projects, the amount of the project can accurately reflect the magnitude and strategic importance of the project's long-term value? If scientific research institutions pursue large projects and big deals in the short term, even with the intention of "achieving performance," and are unwilling to do small-scale projects with long-term significance, they may miss the "0 to 1" from scratch in the process of technology and industrial evolution. These problems exposed in practice urgently need policy designers to invest more effort to solve.
Compared to this, influenced by the discipline evaluation mechanism, the university system in the past 20 years has constantly (or at least once) distanced itself from industrial needs, constantly devaluing horizontal projects, narrowly and increasingly compartmentalizing the understanding of research and basic research. Therefore, the urgent task to solve the problem of disconnection between basic research and industrial needs is to reshape the research system, especially the universities' understanding of demand orientation and imagination, remove shackles, and broaden horizons.
Only when universities and research institutions treat every penny from vertical projects and horizontal projects equally can they rebuild the enthusiasm to serve industrial innovation needs. In the long run, we need to re-examine and establish a public opinion system about basic research. For example, emphasizing the "useless use" characteristics of basic research too much, even denying the possibility of "useful science," or when emphasizing the usefulness and demand orientation of basic research, accusing it of "utilitarianism." (Is the utilitarianism of the science and technology system not caused by the long-term science and technology evaluation orientation and the excessive reliance on research reward-based salary systems?) And for a long time, was the science and technology evaluation orientation demand-oriented and goal-oriented? Actually, the statement of "two legs walking" in the "221 speech" in 2023 has already profoundly answered this question. Reconstructing the public opinion system about basic research is also conducive to reshaping the relationship between scientific and technological innovation and industrial innovation at the social level. Often, many people in society still have a rose-colored illusion about scientific and technological innovation.
I once heard a Hebei comrade self-criticize at a Beijing-Tianjin-Hebei industrial coordination development meeting, saying that the reason for the failure of science and technology transfer is that the local industrial foundation is too weak, making it difficult for Beijing's scientific and technological achievements to find a landing scenario locally. At that time, I gave the Hebei comrade some ideological work, saying, "You are too humble! In fact, Beijing's scientific and technological achievements are also hard to apply in Beijing..." But from a systemic perspective, we need to deeply realize that this "humility" is absolutely not the attitude and way of thinking that should be adopted to promote deep integration. From this perspective, we need to demystify basic research.
Guanchazhe.com: Currently, the reshaping of the industrial ecology by the iteration of frontier technologies is also receiving attention. In your previous papers and articles, you have conveyed an idea that scientific innovation is not entirely a linear development path of "science → technology → industry," but also involves nonlinear processes where science-technology, technology-economy, and technology-society interact. Can you elaborate on this?
Sun Xi: This is a very abstract question. To help everyone better understand, I will tell two stories.
The first story is one I particularly like to tell. It's the story of Chinese breast milk research. There is a large part of the content in breast milk research that is about the composition of breast milk and the functions of these components. China's breast milk research started early, but for a long time it was scattered and unsystematic. China truly began systematic and long-term breast milk research because a group of dairy product enterprises, represented by Sanyuan, Feihe, and Yili, began independent innovation.
For example, Sanyuan built a database of millions of data points on breast milk during this process, and tracked the study of breast milk components, breast milk microorganisms (probiotics), and the functions of various components and their preparation processes for a long time. This led to Sanyuan's independently developed Chinese-style formula milk powder and various innovative fermented yogurts. As a result, Sanyuan's chief scientist Chen Lijun is called "the grandfather of Chinese babies" (because Old Chen does indeed have a daughter, and he is indeed a grandfather), and more importantly, to date, Chen Lijun is the most productive scholar in the field of breast milk research in China. This is a very typical example of industrial development driving technological development, which in turn drives scientific research.
Chen Lijun, National Animal Health and Food Safety Innovation Alliance
The second story comes from Anhui, in the agricultural sector. The involved company is Geyi. Due to non-technical reasons, Geyi's situation is quite difficult, but they have made a very interesting product. Geyi produces a solution during the deep processing of straw, and unexpectedly discovered that this solution can increase crop yields and restore land fertility (rehabilitate saline-alkali land). Further research revealed that the effective components in this solution can chelate heavy metal ions in the soil in large quantities, thus effectively reducing the salt stress on plant seeds and roots; and provide sufficient nutrients for the proliferation of beneficial bacteria, thereby enhancing the stress resistance of crops (enabling them to better cope with adverse growth conditions).
As Geyi themselves put it, "This principle is known to old farmers, but the material is only ours." In other words, the scientific principle has always been there, but no one knew how to do it or what to use. Why? Because Geyi's deep processing technology for straw is self-developed, so this unexpected solution is original. Even more interestingly, during the promotion of this revolutionary product, Geyi encountered an awkward limitation: our country's regulations require fertilizer products to contain nitrogen, phosphorus, and potassium, but their product is an organic solution that does not contain these traditional fertilizer elements. Therefore, they must add a certain amount of nitrogen, phosphorus, and potassium components according to national laws to obtain the "access permit" for the fertilizer market.
Understanding the nonlinear relationship between science, technology, and industry from these two examples, we can easily see that the core characteristic of innovation is practicality. Simply put, innovation is done, not thought up or calculated. Wherever practical development goes, innovation can go, and "useful science" can go wherever it is needed. Therefore, "useful science" has national boundaries and is strictly defined and shaped by a country's industrial system.
Specifically to China, due to our having an extremely rich and complex industrial system that has never appeared in world history, the practical materials for industrial innovation are extremely abundant, which will greatly reduce the threshold for major innovations (the threshold for making and commercializing). This means that respecting the pioneering spirit of the people, letting all industries actively carry out innovation practices, and until the entire national industrial system is revitalized, this process itself can nurture more and better major innovations. The "14th Five-Year Plan" placing "building a modern industrial system" before "high-level technological self-reliance and strength" and as the first of the twelve key work areas during the "14th Five-Year Plan" fully reflects this logic.
But this requires us to carefully handle the issue of "strategic scientists": if we return to the aforementioned policy paradigm of collaboration among government, industry, and academia, we will not find it difficult to see that the strategic decision-making to define the future direction of the industry is obviously not made by "scientists." The understanding of the national strategy, the insight into the competitive situation, and the understanding of the industrial foundation are more reliant on the strategic vision of politicians, entrepreneurs, and engineers.
When many people cite Qian Xuesen as an example to illustrate the extreme importance of "strategic scientists," they probably forgot that Qian's most important suggestion for the 12-year plan, "first missiles, then airplanes," was based on a deep understanding of China's industrial foundation. They probably also forgot that the purpose, content, and influence of Qian's academic thoughts, "Engineering Cybernetics," are deeply rooted in the engineering field.
Accurately grasping the practical characteristics of innovation can also help us understand why today Western countries, especially the United States, discuss original innovation so favorably. The reason is that deindustrialization has weakened their practical foundation for industrial innovation, so they can only rely more on new theoretical models that require less practical materials. But we should ask a question: a good chef, if he has enough ingredients and oil, salt, vinegar, and sauce, can cook a full banquet, would he do molecular gastronomy?
This article is a exclusive article from Guanchazhe.com. The content of the article is purely the personal opinion of the author and does not represent the platform's position. Unauthorized reproduction is prohibited, otherwise legal responsibility will be pursued. Follow Guanchazhe.com WeChat account guanchacn to read interesting articles every day.
Original: https://www.toutiao.com/article/7565781960498168320/
Statement: The article represents the personal views of the author. Please express your attitude by clicking the [Top/Down] button below.