Recently, the topic of DeepSeek has been widely discussed, and we are preparing to approach this issue from a different angle. Does anyone remember a couple of years ago when the entire internet was discussing issues like “the digital economy is merely a virtual economy” and “hard technology is the real technology”?
In the past two years, such discussions have noticeably decreased. This is because more discerning individuals have realized that the boundaries of what many people thought were binary oppositions are not as clear-cut as they seem.
A landmark event in this regard is the emergence of Open AI, where people discovered that AI technology is the most powerful technology, and it embodies the integration of the virtual and the real, specifically the model + GPU approach. With the advent of Open AI, there was a sense of urgency as it became clear that the U.S. had developed something new in the tech field, and this new development was precisely the result of the “integration of the virtual and the real”. If we do not keep up, we will be left behind.
Recently, the reason why DS has attracted so much attention is that people see that China’s large models can finally be on the table in the field of AI, allowing us to synchronize with the world’s cutting-edge technologies.
In fact, during the wave of mobile internet, China was already synchronized with the world. Over a decade ago, we were discussing how China’s e-commerce, mobile payments, social media, and business services reached the forefront globally, and everyone was applauding China’s business innovation. However, after the trade and tech wars began in 2018, many people made a 180-degree turn, as if the innovations of Chinese companies were merely model innovations based on a large population base, which were not impressive. Isn’t it true that we are still being choked in chip technology? That is the real hard technology.
Everyone knows that Alibaba’s e-commerce, including Alipay’s payment and finance, originated in Hangzhou, and they were among the earliest innovators in the internet era. When the voices of “only hard technology matters” and “only hardware matters” emerged, many people criticized Hangzhou, claiming that what the city produced lacked technological content and would not have much future potential.
However, starting last year, people began to notice that Hangzhou was continuously producing many new things. For example, the game “Black Myth: Wukong” suddenly became popular, and the humanoid robot from Yushu Technology was impressive and even appeared on the Spring Festival Gala. The whirlwind stirred up by DS is even more noteworthy. Including these companies, the term “Six Little Dragons of Hangzhou” has emerged. These “six little dragons” encompass those working on systems, hardware, and games, but their underlying logic is all about the digital economy.
All these innovations are an integration of software and hardware, a combination of the virtual and the real. Take robots, for instance; they need to connect to cloud data in the background; they are not standalone entities. Therefore, we cannot maintain the original mindset of completely separating software and hardware. The future is the era of the digital economy, where all industrial activities are data-driven. Data is the blood, the systems that carry data are the veins, and the control center of data is AI. If you detach from data technology and only focus on hardware, it’s akin to a person having muscles without a nervous system.
Looking back, the accumulation of Hangzhou during the mobile internet era is indeed very valuable. E-commerce, payments, and finance were once viewed by some as “virtual economies”, but they are actually the most important application scenarios for digital technology. Technological advancement is fundamentally driven by scenarios. At that time, Hangzhou had the scenarios, which naturally laid the foundation in this field, establishing Alibaba Cloud with strong computing power and a substantial talent pool in digital technology. Thus, it is logical for AI companies to emerge in such a city.
Ultimately, AI is fundamentally about data processing. The strength of AI capabilities essentially reflects the strength of data processing capabilities. Since the internet era, China has accumulated a vast amount of data and application scenarios in data processing, and the technology reserve is also stronger. Therefore, China’s level in data processing is quite prominent. In the global context, it is like a high-speed race with a bunch of cars, where two are sports cars, one from China and one from the U.S., while the rest are mostly tractors.
This is also why the AI competition is essentially a Sino-U.S. competition, with little relation to other countries. It’s like organizing an F1 car race in a village, with only two wealthy participants while others merely watch, as their capabilities can’t even support the wheels of the racing cars.
From last year’s “Black Myth” to the recently popular “Nezha 2”, the achievements made by Chinese in gaming, film, and animation are evident. These typical “virtual economies” are actually backed by hard technology. Everyone knows that games consume graphics cards, but many may not realize that animation and film special effects also demand substantial graphics card resources. Graphics cards were originally born from gaming; without such “virtual” scenarios, humanity’s AI could not have reached its current state.
Now that we have DS, people feel less anxious about the model level. However, many still worry about whether graphics cards are still being choked, or if chips are still being choked. I am quite optimistic because China’s demand for data processing and the richness of application scenarios will inevitably promote the development of the upstream hardware sector, though this is a process.
The biggest confidence for China lies in its market scale. Therefore, we must reassess the so-called “virtual economy”; as China’s “virtual economy” in gaming, film, and animation becomes more developed, our demand for hard technology may increase, leading to faster breakthroughs.
More and more people are beginning to understand that Sino-U.S. competition is all-encompassing. This competition exists in economic, technological, and cultural dimensions.
We need to ensure that we do not fall behind in this competition. Where the U.S. is weak, such as in traditional manufacturing, we should be strong. However, in areas where the U.S. is strong, such as digital technology, film, animation, and cultural dissemination, we must also be strong. It is not about going against the so-called “virtual economy” in the U.S. by solely focusing on hardware and manufacturing.
Lastly, I want to add that with the emergence of DS, there should be a re-evaluation of finance and investment as well.
Everyone knows that the company behind DeepSeek, Huanfang Quantitative, is in the finance sector. The finance industry has faced significant criticism in recent years. Moreover, they are involved in quantitative finance, which many people view as the most disreputable aspect of finance, believing that quantitative approaches create no real value and are merely shameful arbitrage trades. However, quantitative finance is inherently highly related to AI, making finance itself an important scenario that drives the development of digital technology.
Liang Wenfeng mentioned in a previous interview with “Under Currents” that innovation is not designed at the top level but grows freely through trial and error by individuals. Innovation is not solely reliant on capital; it fundamentally stems from a belief system. Why is Silicon Valley known for its strong innovation capabilities? Because they dare to innovate.
I believe he has an unspoken point; those in Silicon Valley are indeed daring, but if they were in Zimbabwe, their courage to innovate would be of little use.
The key is that when a genius’s inspiration bursts forth, they can find a team in Silicon Valley to validate it, rent servers at low costs to build environments, seek venture capital, and consult top experts.
The breakthrough of DeepSeek initially stemmed from a young researcher’s burst of inspiration, followed by the company organizing a team to validate it over several months, ultimately confirming that the direction was feasible, leading to increased resource allocation. This shows that Hangzhou has already adopted a Silicon Valley-like style.
Many people are now saying that DS is not in a hurry for quick results, not rushing for commercialization, and not pursuing DAU, but instead adhering to long-termism in developing large models. However, long-termism requires financial support; without Huanfang’s investment, DS would not be able to sustain its long-term approach.
Many people’s understanding of the U.S. is quite one-sided. The U.S. has a developed financial sector and advanced technology, but when discussing finance, many people refer to it as a “virtual economy”, wanting to “extract the essence and discard the dross” from the U.S., seeking technology but not finance.
Of course, China and the U.S. have different national conditions. The U.S.’s excessive financialization has its issues, but their tech finance is still very effective. Silicon Valley became Silicon Valley because good technological innovation in the U.S. receives robust financial support. Despite being on the west coast, Silicon Valley and Wall Street on the east coast are actually intertwined organic entities in promoting technological innovation.
Therefore, in the context of Sino-U.S. competition, if we want to “learn from the strengths of others”, we should not only focus on technology itself but also on the interaction mechanisms between technology and finance, which are worth emulating.
I now believe that DeepSeek is just the beginning. As long as the soil is right, various fruits will naturally emerge, and most of the results are unpredictable.
In fact, the economy operates in a similar manner; no one knows what will happen next. However, there is a fundamental principle: innovation cannot be planned; only by being open to various industries and fields can we unleash immensely valuable groundbreaking innovations.
Why do I often discuss these matters? Because many people are very suited for roles in gaming, animation, software development, etc., but when they went to college and were influenced by the overall atmosphere, they chose other majors, leading to a lifetime of regret. Thus, such consensus is crucial; on a larger scale, it influences societal progress, and on a smaller scale, it affects individual destinies.
Cool Play Lab is authorized to reprint this.
Please contact the original author for reprinting.
China’s Large Models Can Finally Be On the Table!
