The Peking University AI Aircraft Carrier Fleet makes its maiden voyage to Zhengzhou Airport Economic Zone.
"We mobilized over 40 Peking University alumni entrepreneurs from Beijing, Shanghai and other places to Zhengzhou in less than 10 hours."
At 9:20 am on November 25, 2024, at the Zhengzhou Garden Expo Park Conference Center in Zhengzhou Airport Economic Zone, the place was crowded with Peking University alumni who are rooted in the AI field. Jiang Jun, as the initiator of the Peking University AI Alliance, sat in the front row, depicting to the leaders of the port area the "Future AI City" plan that the alliance is about to launch here, and on behalf of 44 Peking University enterprises, held a signing ceremony with the Management Committee for the project to be implemented locally.
44 AI enterprises collectively flocked to one city and planned to establish their entire organizational structure there. "Such a scale is the first in Chinese history, not to mention that they are a high-quality AI enterprise group like Peking University." A business promotion official told 36Kr.
On the morning of November 25, the 44 Peking University AI enterprises had a meeting with the leaders of the Zhengzhou Airport Economic Zone Management Committee.
This gathering is a pioneering action in the era of AI agents. Jiang Jun and his team aim to obtain strong support from the local government and change the pattern of China's artificial intelligence industry in one fell swoop.
The speed is indeed astonishing, almost like an "impulsive decision". Because just two days before, at 12:17 pm last Saturday, Jiang Jun received a formal invitation from the Management Committee of Zhengzhou Airport Economic Zone. By 10:10 pm that day, the 44 enterprises had completed the assembly. The next day, which was Sunday, almost all entrepreneurs had put aside their current affairs and rushed to Zhengzhou. The unity of Peking University alumni made all the entrepreneurs present use the word "unprecedented" to evaluate it.
But insiders know that this matter is actually "planned for a long time". In October this year, the Peking University AI Alliance was established at the National School of Development of Peking University. This time, Peking University AI enterprises reached a consensus on development in the era of AI agents. This departure is to define a definite direction for the era of AI agents.
"One is the local government's plan to invest 1 billion yuan to support the development of the AI industry, and the other is the government's commitment to open up urban AI application scenarios. This made me determined to lead the Peking University AI Alliance to gather in a city outside Beijing for the first time." Jiang Jun told 36Kr.
Jiang Jun's WeChat nickname is "General", and this name seems to be destined for him to become a commander. He wants to find a winning route for the Peking University "AI Aircraft Carrier" fleet in the future, and he also wants to tell the entire AI industry where the future of AI is.
Finding the Direction for AI 3.0
After graduating with a master's degree from the School of Software and Microelectronics of Peking University, Jiang Jun has been deeply engaged in the AI field for more than ten years. In his eyes, the AI industry is vast, but there is a very clear context. At different development nodes, Jiang Jun can always make his accurate predictions. "It's a pity that one person's voice is too weak." This Peking University graduate who sees the trend of AI does not want to be just a "whistleblower" in the AI era, but more like an evangelist and activist of AI.
Jiang Jun explains the concept of the AI future city to the leaders of the Zhengzhou Airport Economic Zone Management Committee.
What kind of "path" should China's AI development follow? Jiang Jun, who is in the midst of the wave, is more qualified to speak.
Jiang Jun has experienced the first wave of the artificial intelligence era, which is defined as the image recognition era of the AI 1.0 era, and China has given birth to the "Four Little Dragons" of AI. From 2016 to 2019, Jiang Jun was in one of the Four Little Dragons and once led the team to grow at a rate of 600% per year. In 2018, he "inopportunely" predicted the collapse of the AI market. In 2019, when the AI 1.0 bubble burst, he failed to change the inherent path dependence of the enterprise he was in with his own strength, leaving only a sigh.
In 2020, Jiang Jun joined Alibaba. He was ambitious and designed the "AI City Training Ground Plan", planning to invest 5 billion yuan. The reason for this almost crazy idea is "because I am aware of the bottlenecks encountered by AI technology, which requires one to two orders of magnitude of data, computing power and algorithm scale." He led Alibaba DAMO Academy to unite more than 50 AI companies to jointly promote the breakthrough of domestic AI large models and make China's artificial intelligence reach the world-leading level.
With the emergence of ChatGPT, in 2023, large models swept the world. Isn't this the goal that Jiang Jun's 5 billion yuan plan wanted to achieve? Thus, the AI 2.0 era in China began with a bang. Represented by the base large model, after the "Hundred Model Battle", the "Six Little Tigers" of AI were born. Relying on the power of capital, they wrote the "myth" of the new AI era.
In 2023, Jiang Jun embarked on his own entrepreneurial path. At the beginning of this journey, he once again "blew the whistle" and poured cold water on the AI 2.0 era. He predicted that the price of large models would collapse and advised start-up enterprises not to build base large models. He said: "China does not need so many people to build engines."
In the AI 1.0 era, the "Four Little Dragons" of AI completed the valuation from hundreds of millions to 20 billion in three years; in the AI 2.0 era, the "Six Little Tigers" of domestic AI large models completed the valuation transition from several hundred million to 20 billion in just one year. Jiang Jun said: "This means that the matter of 'building engines' has almost siphoned off the funds invested in AI. At this time, it has become very difficult for Peking University enterprises to raise funds, let alone others. This is actually very unfavorable to the development of China's artificial intelligence industry."
One year later, the price of large models has indeed collapsed, "dropping by a hundred times". This means the end of the AI 2.0 large model era. The era of repeatedly building engines has ended. Next, a question that all AI practitioners are concerned about is: In addition to the "Hundred Model Battle", where exactly should AI go?
The AI 3.0 era is simmering beneath the surface. Just like after the invention of the steam engine, those who made the industrial revolution were the ones who invented trains and automobiles. From the perspective of various industries, the next wave of AI is how to integrate into industry applications and use the tool of AI to lead the transformation of the entire industry.
Jiang Jun and his Peking University AI Alliance believe that AI 3.0 is the era of intelligent agents. They have formed a consensus on the era of AI intelligent agents, restarting from the form of individual combat to a joint aircraft carrier, opening up a new AI track.
Jiang Jun envisions connecting AI companies and industry enterprises through an intelligent agent platform to promote the application of intelligent agent technology in various industries. Based on this concept, Jiang Jun founded Heichuan Technology. "Through AI to achieve true technological equality, enabling AI to empower various industries like water and air, helping every ordinary person improve their living standards and work efficiency."
Jiang Jun decides to take the plunge, get involved, build the "AI Aircraft Carrier" of Peking University, and determine the course of the AI 3.0 era.
Reorganizing the Peking University AI System
According to Jiang Jun's judgment, AI core technology, like water and electricity, has become an infrastructure. After the sounds, vision, and text were matured in the AI 1.0 and AI 2.0 eras, in the AI 3.0 era, large-scale industry applications can be realized. Therefore, in the AI 3.0 era, the industry landing with Agent as the core is the real integration of artificial intelligence and the industry.
However, the real application scenarios of AI landing are not so simple. There is a gap between AI and scenarios, which must be broken through by people who have a particularly good understanding of the industry, rather than those who have a particularly good understanding of AI. From this perspective, Peking University has the greatest advantage. Because Peking University is a leader in every specialized industry. Once combined with AI technology and drawing on the experience of the previous two eras, more unicorns will be produced.
Has the opportunity for Peking University AI come in the AI 3.0 era?
In April this year, the Hurun Research Institute released the "2024 Global Unicorn List". The list shows that since the new century, Peking University has incubated a total of 8 unicorns with a total valuation of 143.8 billion yuan; while Tsinghua University has incubated 33 unicorns with a total valuation of 533 billion yuan, leading by a wide margin. In the AI 2.0 large model era, the AI of Tsinghua University rose rapidly. Three Tsinghua University-affiliated companies completed the unicorn valuation of 10 billion yuan in one year, and more than a dozen became quasi-unicorns.
What is lacking in Peking University compared to Tsinghua University? If Peking University wants to achieve a lead by changing tracks in the next step, how should it learn from Tsinghua University?
Some observers have seen that the Tsinghua University system has its success code. First, the entrepreneurial teams of Tsinghua University have a think tank composed of Tsinghua professors and academicians, who organize technology, contacts, and other social resources for entrepreneurs. Second, the Tsinghua University system is united and mutually supportive. After some AI companies grow and become stronger, they also help other Tsinghua University-affiliated enterprises develop through investment or cooperation, forming a benign collective advancement model. Third, the investment and media circles have a particular fondness for the Tsinghua University system. Fourth, the flywheel effect also makes it easier for them to obtain the full support of the local government.
The superposition of various forces makes them unstoppable, able to accurately capture the trend, and even create the trend, forming a siphon effect.
Jiang Jun first put forward the slogan of learning from Tsinghua University. Because the era of AI intelligent agents requires more joint operations. Someone must stand up, unite all the forces that can be united, and move forward together.
In October this year, Jiang Jun first united with Dragon TV and led outstanding Peking University AI enterprises into the "Predicting Unicorns" column. Jiang Jun is the vice president of the Peking University Alumni Entrepreneurship Federation. Through Peking University alumni resources, he reached a cooperation with the National School of Development of Peking University to organize roadshows for Peking University AI enterprises. At this time, Jiang Jun contacted Chen Xiangqun, the Party Secretary of his college, and found eight professors in the computer field of Peking University.
Among them, the National School of Development of Peking University is composed of well-known scholars such as Lin Yifu, focusing on macroeconomics and future industries. It is a national think tank, and the AI industry is also one of its research focuses. At the same time, the top investor and Peking University alumnus Yan Yan also specially returned to Peking University from Japan to participate in this event. With the participation of various experts and the convergence of core resources, this event successfully attracted 50 Peking University alumni enterprises scattered in various places in the AI field to sign up, and finally 7 enterprises appeared on the "Predicting Unicorns" column of Dragon TV. During this process, many local governments also threw out olive branches, expressing their willingness to cooperate.
It is precisely because of this large-scale gathering that the Peking University AI Alliance reached a consensus on the AI 3.0 intelligent agent era and achieved ideological unity. Therefore, there is the collective departure of everyone in the Zhengzhou Airport Economic Zone.
In Jiang Jun's plan, the launch of this Peking University "AI Aircraft Carrier" fleet will unite Peking University professors, academic institutions, investment and financing institutions, and the media, combine local governments with alumni enterprises, learn from the successful path of Tsinghua University in the 2.0 era, seize the opportunity in the AI 3.0 era, continue this model, and lead artificial intelligence to the right path. According to his prediction, in the AI 3.0 era, there will still be a growth of hundreds of billions of yuan, which is the best opportunity for the rise of Peking University AI.
Why Choose Zhengzhou as the First Stop?
Why choose Zhengzhou as the first stop?
This is the question that Jiang Jun constantly communicates with alliance entrepreneurs. He also expressed his core viewpoint to 36Kr:
First, in the AI 1.0 and 2.0 eras, the core of industrial development is in Beijing, Shanghai, and Guangzhou, because that is where top AI technology R & D personnel gather. In the AI 3.0 era, the application scenario is the core, which requires AI to sink to the local level and into the industry. Henan is a populous province with rich application scenarios, so it is a very good market.
Second, there were no top AI companies or top universities in Henan before. Dozens of Peking University enterprises coming to Henan is to introduce intelligence and empowerment to Henan, jointly build a future city experimental field, and form a bridgehead between Peking University and the airport, which can form an enclave effect.
Third, the leaders in charge of the Zhengzhou Airport Economic Zone understand AI better. Tian Haitao, the director of the Zhengzhou Airport Economic Zone Management Committee, graduated with a Ph.D. from the School of Information Science and Technology of the University of Science and Technology of China. He himself is a computer expert; the one who conducts specific business docking with the Peking University AI Alliance is Henan Airport Digital City Development and Construction Co., Ltd. The company's chairman, Wang Jingzhong, used to be from the Government-Enterprise Business Department of Henan Mobile and was also the former general manager of the Digital Government Platform Company of Henan Province. Therefore, this joint effort between the two sides is based on a professional judgment of the trend of the AI 3.0 era.
At the symposium with Peking University entrepreneurs on the same day, Wang Jingzhong, the chairman of Henan Airport Digital City Development and Construction Co., Ltd., also said that in the future, the smart city of the airport zone will be made into a "large model +" mode, + government affairs, + environmental protection, + urban management, + health care, etc., to open scenarios for landing enterprises. At the same time, through the issuance of special treasury bonds, they will jointly invest, build a platform, and create a model for AI applications to help landing enterprises promote and sell externally.
The rapid formation of this blueprint for the future AI city benefits from the "AI City Training Ground Plan" led by Jiang Jun at Alibaba. It is derived from a grand narrative. Combined with 44 Peking University AI enterprises that have already united and have unique R & D capabilities in various industries, the future AI city map of the Zhengzhou Airport Economic Zone has evolved.
The Peking University AI enterprises jointly layout the future AI city.
Different from the plan five years ago, this is not the construction of a basic large model, but the reconstruction based on the AI 3.0 intelligent agent era. It is to start a new chapter of the future AI smart city on a blank sheet of paper.
When AI technologies such as LLM and Agent are applied to smart cities, it may trigger fundamental changes in related intelligent capabilities because these technologies essentially redefine the paradig