Artificial intelligence competes, first on the table before there is a future

01

Recently, at the 2023 Zhongguancun Forum held in Beijing, Robin Li delivered a speech entitled "Big Model Changes the World". Li Yanhong believes that artificial intelligence has once again become the focus of human innovation, and the new era of artificial intelligence with large models as the core, the fourth industrial revolution is coming.

When talking about preventing AI from getting out of control, Li Yanhong said that to prevent loss of control, countries with advanced AI technology need to work together to formulate rules from the perspective of a community with a shared future for mankind. To participate in the formulation of the rules, you must first go to the card table before you can have a voice and have a ticket to the global competition.

Li Yanhong's "card table" here refers to participating in market competition in the era of artificial intelligence, using artificial intelligence for products and entering the market.

First of all, entering the market is the traditional model of technological development.

In the traditional economic and technological development model, the business model of technological upgrading is: research and development, sales, making money, and investing in research and development, technological upgrading and sales again.

This slow technology upgrade model, on the one hand, capital in order to maximize the profit of research and development, on the other hand, also need to accumulate profits to invest in research and development. It's like Intel's CPU, which has slowly developed to its current level of technology after years of this cycle. If intel holds its breath, burns a lot of money, and directly launches "black technology", it can speed up the upgrade of the CPU. But no company, no investor, can endure losses for more than a decade.

Therefore, entering the market first and supporting research and development through profits is the way to survive and develop scientific and technological products. Therefore, first on the card table and first into the market, in order to support research and development.

Second, in the era of the Internet, especially the mobile Internet, entering the market has become more important.

At this stage, the explosive development of technology has brought about a new "financing-listing" model, which is extremely dependent on market share. New technologies get VC capital, quickly turn the most basic business model into reality, then attract traffic and customers, then scale up through several rounds of funding and finally go public.

In this model, the first few players take all, and market share becomes crucial, even the only thing that matters beyond profitability. Therefore, it is very important to get to the table first, enter the market first, attract customers, and become the top player in the world.

However, whether it is the profitability of the traditional model or the market share required for "financing-listing" under the new economic model, it is inseparable from participation in the market. Therefore, participating in the game at the card table and developing oneself in business is already an iron law of technological development.

Third, in the era of artificial intelligence, entering the market has become an intrinsic demand for technological development.

Algorithm training, deep learning, etc. require a large amount of data support, and data mining is a prerequisite for artificial intelligence to do training samples. For example, unmanned vehicles need a large amount of test data, constantly test the algorithm, and form a self-improving feedback system in business scenarios.

The best data is to obtain real, first-hand, and most close to the actual scenario data in the market operation. Therefore, in the era of artificial intelligence, the market has become more important, and the market not only provides funds, but also provides technical elements.

Fourth, the table is actually a window after which there is almost no chance.

In November 2022, the European Parliament passed the first resolution related to the development of the video game industry, calling on member states to increase support and investment in game research and development. Although the total annual output value of the European game industry is as high as 23.3 billion euros, from a global point of view, the European game industry is not strong, in 2022, the Asia-Pacific region and North America accounted for 49% and 26% respectively, and the European market share was only 17%.

Therefore, the EU is of course looking at the economic value and technical value of the game industry and wants to participate in the formulation of industry rules, rather than only fines American companies. To put it bluntly, I just want to be at the table.

However, Europe is only waking up now, and it is too late. In contrast, China is already a big player at the global table. With revenues of 44 billion US dollars and 666 million players, China's game industry has become the largest country in the global game industry and is already a big player in the global market.

At present, games have become an important way to drive technology and economic development. According to the report "Game Technology - New Technology Cluster in the Process of Digital and Real Integration" released by the Chinese Academy of Sciences in July 2022, game technology is contributing to the rapid development of 5G, chip industry and artificial intelligence industry, and game technology has contributed 14.9%, 46.3% and 71.6% to scientific and technological progress in the chip industry, high-speed communication network industry and AR/VR industry, respectively.

The development of CPUs and GPUs driven by the game industry has in turn promoted the development of artificial intelligence and the metaverse. Therefore, participating in the previous hand often means a ticket to the next hand.

So, China is already a player in the global video game market. This is a good thing for China's technological development. The development of China's game industry, especially the mobile game industry, has been given a rare opportunity. Now it is already difficult for Europe to catch up. In contrast, large models are the head player takes all, and this time window is more important.

02

It is already a consensus that large-model artificial intelligence will reshape the business landscape. Artificial intelligence can fulfill the multi-dimensional needs of humans. You can book, navigate, and shop through conversations, just like having your own personal assistant; You can also set up routers, mobile phones, home appliances through conversations; You can also use artificial intelligence to assist yourself in design, documents and PPT.

Therefore, with the development of AI, from office, entertainment, to life, broader and deeper needs will also be discovered and met. Generative AI will greatly improve work efficiency. Some research institutions believe that in the next 10 years, the efficiency of knowledge workers can be increased by 4 times. This will change the picture of production, commerce and life in the future global society.

Behind the big model changing artificial intelligence, the IT technology stack has also undergone a very fundamental change. The IT technology stack will change from the chip layer, operating system layer, and application layer to the chip layer, framework layer, model layer, and application layer.

While the U.S. still leads the world in this area, Chinese companies already have the ability to "hit the table." Only when you are on the card table, have a ticket to the global competition, and have the right to speak, can you participate in the formulation of rules, and then have a say in AI safety, artificial intelligence and social and national rule-making.

Therefore, the rapid entry of artificial intelligence products into practical use is very important for the current development and even the long-term vision of Chinese intelligence. This requires not only companies to innovate, but also a level of supervision.

The most important thing in the current regulatory policy and industrial policy is to let China's large model go online, face the market, and continuously polish and improve in practice, rather than just staying in the internal testing stage. This kind of internal testing is not sustainable for a long time, from the perspective of technological development, there is a lack of a large amount of data in the market to support large models, and from an economic point of view, the lack of profits to support R&D and development.

When new technology goes online, of course, there are risks. In fact, the so-called safety and risk is not only a financial risk, but also a variety of social risks, Tesla automatic driving will have accidents, new drugs will have unexpected side effects, even a concert will appear the risk of the big screen being blown down by the wind.

Therefore, a country's innovation is by no means something that entrepreneurs can achieve if they are willing to bear risks, but requires entrepreneurs, society, regulatory authorities, laws and regulations to jointly bear risks and resolve risks.

(Liu Yuanju is a contributor to Qin Shuo's circle of friends and a researcher at the Shanghai Institute of Finance and Law)