Frontier Guide
According to authoritative domestic media news reports, Pu Muming, an internationally renowned scientist and academician of the Chinese Academy of Sciences, asked domestic scientific research institutions rhetorically in the "National Science Review": Why can't domestic research institutions develop such influential technological achievements as DeepSeek after receiving a large number of funding for AI projects, employing a large number of talents in related fields, and publishing more and more papers?
Technological achievements
At the same time of asking domestic scientific research institutions, Academician Pu Muming also discussed this issue with some young scientific researchers, and found that the vast majority of scientific researchers did not dare to develop in the field of this kind of technical project, one was afraid that the risk was too great to complete, and the other was that there were not enough resources to support technological development.
This kind of risk concern is indeed an objective factor. After all, the United States has more than once carried out technological suppression of China's scientific and technological development, prohibiting overseas companies from providing technology and equipment materials to Chinese companies, and not allowing them to export products to the Chinese market.
To develop cutting-edge AI technology, high-computing chips are needed as support. Chip products are a technical field that the United States strictly guards against the Chinese market, and no one can predict whether it can successfully break through the technological blockade and achieve profitability in the consumer market under the premise of investing a lot of resources in the AI field.
Once resources are invested and technical products are successfully developed, but there is no ability to make profits, this will not only make investors lose confidence in their investment, but also make relevant technical personnel suffer a certain degree of public opinion "attack". From many aspects, many of our enterprises and institutions do not have the ability to develop new technologies, but are afraid that they will not be able to make a difference after failure.
Academician Pu Muming also believes that many scientific researchers now follow the technological development experience of overseas countries too much, resulting in a clear understanding of what China's science and technology needs.
Researchers in Europe and the United States are mainly developing in their own fields of interest, and then becoming top experts in the industry step by step, with the aim of realizing their own value and making their technology the cornerstone of leading the development of the new international era.
For example, Jobs, Noyce, Bardeen and other internationally renowned business leaders and scientists, they are persistent in the pursuit of technology, and even give up profitability in the market in order to pursue innovation in technology and products.
In particular, Steve Jobs, who is regarded as a technological revolutionary in the new era, insisted on using brushed aluminum materials on the bezel of the iPhone 4 in order to polish his mobile phone products into an unprecedented craft, and insisted on using brushed aluminum materials on the bezel of the iPhone 4 in order to polish his mobile phone products into an unprecedented craft. Although this product has ushered in a new revolution in smartphones, its signal problems have also caused many consumers to complain.
The Chinese market is different, and China's traditional scientific research culture advocates pragmatism.
The birth of new technologies needs to promote the development of the country and the living standards of the people. Although your technology is very advanced and has a very high gold content, people can't use it, and people can't appreciate how powerful your technology really is. Even if your product is full of technology, but it can't solve the pain points of consumers, the product technology can be defined as a failure in the Chinese market.
Dare to challenge
After the popularity of DeepSeek, Academician Pu Muming believes that domestic research institutions should seriously think about how to create conditions for young researchers at the bottom of the research system to pursue their own ideas.
Liang Wenfeng, the founder of DeepSeek, was just one of many domestic college graduates before he became famous. Later, he started his own business and developed in the field of AI. The resources at his disposal and the environment around him are far from those of young scientists working in scientific research institutions.
Liang Wenfeng was very interested in the financial market, so he began to use the financial industry as a breakthrough to pioneer the technology of using machine learning to analyze market data and conduct fully automated quantitative trading. In the subsequent development, he co-founded High-Flyer Quant and began to conduct in-depth research in the field of artificial intelligence.
In 2023, Liang Wenfeng transferred some of the company's elite personnel to establish DeepSeek, which is based on AI large model technology. After a period of time, DeepSeek launched two high-level large models, V3 and R1, whose inference effect and computing performance are comparable to Chat-GPT developed by open AI in the United States.
In addition, Liang Wenfeng also announced the open source of DeepSeek's large model technology, so that enterprises in need can easily access the DeepSeek large model language and introduce the technology into their own products.
Liang Wenfeng is a typical example of daring to challenge new technologies, and he did not blindly learn the operating models of European and American companies, but adjusted accordingly according to the international situation and the development of China's science and technology.
After the United States restricted Nvidia from exporting high-performance AI computing chips to Chinese companies, DeepSeek was also impacted. Under the premise that it could not obtain the most advanced AI chips, DeepSeek decisively chose to adjust the technical training, which not only greatly reduced the training cost, but also was not weaker than GPT in the United States in information reasoning in various fields.
DeepSeek has successfully developed the AI large model technology that the United States is proud of in the civilian market, and has also received strong support from many domestic companies.
In addition, DeepSeek also provides enterprise-based AI technology customization services, which are operated on a fee-based basis to ensure DeepSeek's normal revenue. In addition, for small and medium-sized enterprises, DeepSeek also provides a computing power leasing service system to help related enterprises develop AI technology to the greatest extent.
For highly innovative enterprises and leaders like DeepSeek and Liang Wenfeng, it is a typical model enterprise that dares to carry out technological innovation, and makes specific technical adjustments according to the existing situation, and realizes the technological catch-up of overseas enterprises with less resources.
For an innovative person like Liang Wenfeng, Academician Pu Muming felt that he should be strongly supported, and called on young researchers to learn from him.
Academician Pu Muming believes that we do not only pay attention to academic performance, but also pay more attention to whether they have enough enthusiasm and courage for scientific research. Liang Wenfeng has made a good start, and hopes that in the future, there will be more young scientific researchers who can become technicians with innovative spirit and certain courage like Liang Wenfeng.