Start planning for my 2024
On December 20, 2023, the 2023 Exploration Conference of "Digital Power, Exploring the Universe" hosted by the Internet Society of China, Weibo and Sina News kicked off in Beijing. Fan Gang, vice president of the China Society for Economic Reform, president of the China (Shenzhen) Comprehensive Development Research Institute, and director of the National Economic Research Institute, delivered a speech entitled "Managing Economic Fluctuations and Maintaining Long-term Growth".
The following is a transcript of Mr. Fan Gang's speech, which has been edited and slightly abridged
Congratulations on the convening of the "Exploration Conference", and especially thank you for inviting me to participate in this conference, I understand that all of you here are mainly engaged in scientific and technological research, including many scientific and technological enterprises, including some companies that are mainly engaged in scientific and technological investment, and may be the only one who is not engaged in science and technology, but engaged in economic research. Economic research has always been very concerned about the development of science and technology, because science and technology are very important elements in economic development, so today I am here to talk about some issues related to economic and social development in the development of science and technology, and to share my own views.
The first aspect: about the recent development of large models, the enlightenment of the development of AI to economics.
After ChatGPT came along, people in the tech community told me that the gap between the United States and China in AI research has widened again. In the past, the gap may have been one year, but now it seems to have become three to four years, and there is a problem of computing power and storage. Academician Ni Guangnan just said very well, the concept of generalized computing power, not only computing power has supercomputers, but also the requirements of supercomputing machines and ChatGPT's computing power are not quite the same, but also add transportation capacity, this gap will widen again. We'll have to chase it, but if we chase it, something new will appear. We may know what Q-Star is, but we don't know what Q-Star is all about, Q-Star has caused all kinds of reactions from all sides, and Q-Star may widen the distance between us.
What is the economic reflection that this has caused?It's about comparative advantage. Some scholars attach great importance to comparative advantage, saying that as long as we have a comparative advantage in backward countries, we will be able to catch up gradually, and they are very optimistic. But there is a fundamental problem that is not noticed: you have a comparative advantage, and the developed countries also have a comparative advantage, and your so-called comparative advantage is the primary factor: cheap labor, natural resources, these comparative advantages. The comparative advantage of developed countries, on the other hand, is that of high-quality factors: capital, human capital, knowledge, technology and, very importantly, the ability to innovate.
The ability to innovate is not only whether you have good scientists, but also a set of systems, various industries, and various cutting-edge fields of science and technology, all of which have reached a certain height and have a leading position, so that you can continue to innovate. In this sense, backward countries do not have the comparative advantage of high-quality factors, and more importantly, the comparative advantage of the primary factors of backward countries will decay as they walk. Why attenuation?You compare yourself with these developing countries of your kind, at the same level, after a period of time, your income increases, and you lose your comparative advantage, you lose it to Vietnam, you lose it to Bangladesh, you lose it to India, you lose it to Africa, you don't always have that comparative advantage. If you lose your comparative advantage, and you don't have the comparative advantage of high-quality factors, you fall into the middle-income trap. The original comparative advantage is gone, and you don't have the comparative advantage of those high-quality factors that the developed countries have, what should you do at this time?
So in my opinion, another element that is more important than comparative advantage is that of backward countries, which is the late-mover advantage. It is that you are lagging behind, but you can use relatively fast methods and a relatively short time to quickly learn the knowledge and technology that other advanced countries have created. Through open introduction, digestion, absorption, learning, etc., the knowledge and technology that have been created by developed countries will spill over into your country, and there is a word called knowledge spillover.
Our development in the past few decades, especially in the second half of the year, the development of the next 20 years, when our labor costs have increased, our development in the second half of the year largely depends on the late-mover advantage, and we have shortened the gap. The developed countries also see very clearly and know how you are developing, so they start to curb all the measures taken when you develop, and no one cares about the original comparative advantage. Labor-intensive industries are all about late-mover advantage, decoupling, cutting off supply, stopping the export of high-tech products to you, restricting scientific and technological exchanges, restricting international students from studying abroad, and so on, all of which are aimed at blocking the channels for your knowledge spillover, and are all aimed at your late-mover advantage. After we are clogged by it, we emphasize independent research and development, independent research and development is always our goal, and improving our innovation ability is always our goal.
We have to think that when the technology has not yet gone to the forefront, and the industry has not gone to the forefront, it is the best choice to give full play to the late-mover advantage. When you don't know how to do it, citing may be the best choice, and the growth rate will be faster, and the ability to gradually learn and innovate in the process of using knowledge spillover will be introduced. When you don't go to the forefront and don't have this ability, you have to do innovative things by yourself, and you have to overcome the problem of "stuck neck" by yourself, which is the next best choice. What is the performance of sub-optimal selection?Even if the cost is relatively high, the loss is relatively large, and the growth rate is low, there will also be such a thing as "unfinished buildings" in science and technology. But now there is no other way but to create ourselves, and at this time we are facing a relatively low growth rate.
Recently, we have lagged behind in the development of science and technology, and the gap has widened, which precisely reflects the comparative advantages of the developed countries. These problems are very clear, and it will take a lot of effort and time for us to catch up. One revelation is that we must continue to be open, you have to block the play of my late-mover advantage, we must be more open, we must be open to more people, we must be open to more industries, we must be open to more enterprises. Therefore, from this point of view, our industry and enterprises should continue to keep an open mind, we are in the world market, in the world's international competition, we cultivate our innovation ability, we learn to introduce, digest and absorb to cultivate our innovation ability, this is my research on AI, some reflections on economic theory.
The second aspect is some of the panic caused by the recent development of artificial intelligence (AI).
I participated in a forum in the legal community, and the legal community was quite panicked, the profession of legal assistant is now in jeopardy, and all legal assistants do is to find the relevant terms in the law books on the back wall of the law firm. Now it is said that no one else is needed at all, and the lawyer himself brushes it out, and the resulting document is perfect.
In the UK, I happened to visit an old factory and transformed it into a modern place like 798, and I happened to see American producers, American authors, and film writers on their mobile phones ** demonstrating that ChatGPT is not allowed to be used. Why is it so touching to see it there?Thinking of the incident of workers smashing machines 200 years ago, when the host was looking back on the scientific and technological revolution for decades, it was a similar situation for machines to replace people in the past, and at that time they simply smashed machines, which is a problem.
Then I heard about the internal conflict of OpenAI, and the creation of Q-Star made the people in their board of directors feel that this matter is very dangerous, and they want to stop this research and limit this research. At the same time, the United Kingdom has held the first artificial intelligence security conference, and countries have studied how to regulate and monitor together. This has created a trend to limit the slowing down of AI research, will it be successful?I don't know either, but logically the research doesn't stop. The curiosity that we talked about in that conversation just now, many people will continue to study it for the sake of curiosity, and human beings will continue to study the damage it has to human beings.
Therefore, we should not be affected by this aspect, just stop this aspect, on the contrary, we should continue to increase the research, the research will definitely push forward, as for how to use it, it is indeed worth discussing. The recent movie atomic bomb nuclear energy is the same, nuclear physics must promote research, but how to use nuclear physics, this has formed a world alliance later, to control nuclear **, to limit the development and utilization of nuclear ** and so on. Human beings want to regulate it, but you must first continue to study, you don't study others will study, in a certain sense, we also have to see the aspect of the machine, just said 200 years ago workers smashed the machine, but 200 years ago the machine has been developing, the machine has been replacing people, people are living better and better, people's income is getting higher and higher. Therefore, it is necessary to see the positive aspects, strive to promote, and to have proper supervision and control, of course, it is also needed. Another example is genetic engineering, which also needs to be regulated by humans. But development is the last word, it will definitely move forward, just said that our gap has widened again, now how do we close the gap, how do we keep up with the new research.
The third aspect: It is also related to the development of the economy, how to give full play to the role of large enterprises, encourage large enterprises to do big research, and promote the development of high and new technology.
This involves the question of how to look at big business. We used to have some large enterprises, some of which were ranked in the top few in the world, and now they have fallen out of the top ten. As soon as large enterprises develop, they often arouse some worries about monopoly, the power of capital, and so on, so they often adopt some restrictive measures. We desperately encourage the development of small and medium-sized enterprises, the development of start-ups, and there is nothing wrong with that, without the development of start-ups, nothing new will pop up. But we must see the role of big business, and the role that big business plays in scientific and technological progress.
After chatgpt and large models came out, we seem to be lagging behind again, and we also have several large models, but it seems that we only have one of the top ten in the world, and we also have three or four models with more than 100 billion parameters.
Usually the classical theory of economics always says that perfect competition is the best state. Perfect competition is really good, but fortunately, it puts the lowest pressure and the most affordable for consumers. But consumers now get the lowest ** it has a price, that is, because you are very low, profits are very thin, enterprises have no money to do research and development, so consumers have no benefits of innovation, can not get the benefits of innovation. Therefore, a theory of economic research is that "the best industrial structure may not be perfect competition, but oligopoly competition", a few enterprises, there is also competition, so it will not be too high and outrageous, not a monopoly, but its profits can be relatively thick, and large enterprises will have the money to do research, to do big research, so that there can be continuous innovation. Small enterprises can often only focus on a small technology for research and development, while large enterprises can carry out comprehensive research and development, and comprehensive research and development can drive some new research and development in various small fields, and drive small enterprises to research. How is our policy environment and institutional environment conducive to the development of China's large enterprises, do bigger research, do bigger investment, do continuous investment, our enterprises can really continue to develop, continue to invest.
From the current point of view, our technological progress, large enterprises play an important role, everyone can count these large enterprises. China's enterprises look at it as if it is quite big, it is actually very small in the world, and the competition in most fields is either domestic competition or international competition, and it does not encourage its development in the international arena, but makes other large enterprises have been strengthened, so how can our policy and institutional environment encourage large enterprises, especially private large enterprises. What is the concept of private ownership?The concept of taking risks with your own money, and the concept of taking risks with your own money to invest. State-owned enterprises use public money and public assets, so they have to be supervised, they have to prevent risks, etc., and the procedures are relatively slow, and objectively they will have more supervision. It is relatively flexible to use its own money and can seize market opportunities, especially large private enterprises. We should encourage it to play a big role in scientific and technological innovation, in a new round of scientific and technological innovation, and in the creation process of large models.
Fourth: the issue of supervision.
In two aspects, on the one hand, it is necessary to pay attention to the new problems in the new research and development field and supervise them in a timely manner. The second aspect is that our companies, our innovators, our innovative big enterprises have to embrace regulation, we have to participate in regulation, we have to try to help develop regulation.
These are two aspects that complement each other, and without innovation you don't know what to regulate. Finding problems that need to be regulated is what the original innovators discovered, and it is the vulnerabilities that are discovered. I remember in the late 90s and early 00s, when we had a meeting in Geneva, the luncheon speech of the conference invited the US ambassador to those international agencies in Geneva, a lady, to talk about the regulatory rules of e-commerce, at that time we did not have e-commerce, there was no online transaction. Because they are innovators, they know what the problem is, and we listened to it like a book from heaven, and we didn't know what the problem was, how did she know it?It is because they have technology companies, they have innovated in the front, have found problems, and then communicate with innovators and form their regulatory proposals, to implement joint construction and management around the world, because e-commerce also involves cross-border e-commerce.
This sees the relationship between the innovator and the first and the different roles, and the first must strengthen the intervention when new affairs arise, and supervise them, even if they are a little ahead at the beginning, they can do some pre-supervision. Problems should be raised first, and then specific measures should be gradually improved, which can be revised. Non-regulation is often a problem, we are slower in dangerous things, we are indeed more cautious about new drugs and genes, we do not supervise some companies, and the last business is goneDelay in supervising a financial format, and finally there are a lot of Ponzi's ** things, the company has disappeared, and the business format has disappeared, so it is necessary to find problems in time and supervise them in time. I think the International Conference on Artificial Intelligence Security is quite meaningful, and China has also actively participated in it.
Our enterprise wants to play a role, we innovate first to find problems, find problems and communicate with the first, we develop standards, we participate in the formulation of standards, and participate in the development process of the industry to solve some problems that may be faced. Therefore, when a new thing happens, we should consider all aspects of it, consider the economic and social impact it may bring, new technology can develop, our business can make money, and our investment can have a return. These are very important, we have made scientific and technological progress, and we can continue to develop internationally, which is very important. However, how to avoid some problems in the process does not mean that excessive caution and excessive avoidance hinder the development of science and technology, which is difficult to find a balance. The board of directors of OpenAI has considered it quite seriously and taken quite drastic measures, all of which have risen to the conscience and conscience of the society, and have risen to such a level, which I believe is also the common pursuit of our scientific and technological people. At that time, when nuclear technology was born, these scientists, from Einstein to Oppenheimer, all had this kind of conscience thinking torture. Therefore, these are also some unavoidable problems in the progress of science and technology, which deserve our timely attention.