This is the headline article I published on the technology edition of The Paper on July 13.
Continue to track US encryption regulatory policies for everyone and contribute better content.
My Twitter: @curiousjoe5
"Sam (Sam Altman, CEO of OpenAI) came to China in 2015. At that time, the term Web 3.0 had not yet appeared. We discussed what the next generation of the Internet would look like, with the focus mainly on artificial intelligence and blockchain." Recently, Kong Jianping, founder of Nano Labs and director of Hong Kong Cyberport Management Company Limited, recalled in an exclusive interview with The Paper (www.thepaper.cn).
Today, "the influence of artificial intelligence needs no further explanation, especially with its rapid development in the past six months." In Kong Jianping's view, artificial intelligence plus blockchain is the real next-generation Internet, because both artificial intelligence and blockchain are hugely transformative technologies. At the same time, with the development of artificial intelligence productivity, entities in the real world can be infinitely expanded in the metaverse. However, at present, the metaverse faces two types of problems: the first is that there are too many shortcomings in the platform, which limits the growth of the number of users; the second is in terms of infrastructure, that is, there are challenges in both concurrency and usage costs.
In the interview, Kong Jianping has been emphasizing how to promote the development of the Metaverse and Web3.0 from the perspective of infrastructure. Over the past decade, he has been continuously starting businesses in the Web3.0 track. After graduating from university, he worked in the semiconductor industry. After being exposed to digital currency in 2012, he began to devote himself to this field. He first invested in Canaan Technology, a chip design company in the digital currency field, and served as the company's co-chairman. Later, he invested in "The 9th City" listed on the US stock market to help it transform into the blockchain industry. Last year, Zhejiang Nanomicro Technology Co., Ltd., which he founded, was listed on the US stock market.
From "barrel theory" to "long board lengthening"
The Paper: When people talk about the definition of Web3.0, they tend to interpret it in the order from Web1.0 to Web3.0, for example: Web1.0 is read-only, Web2.0 is read+write, and Web3.0 is read+write+own. You once defined it as "Web3.0 is artificial intelligence plus blockchain". Can you explain it from your perspective?
Kong Jianping: The origin of this topic can be traced back to 2015. At that time, OpenAI CEO Sam Altman was still the chairman of YC (Y Combinator). Sam came to China in 2015, and the term Web3.0 had not yet appeared. We discussed what the next generation of the Internet would look like, with the focus mainly on artificial intelligence and blockchain. For our industry, Web3.0 in a narrow sense may refer to blockchain, but I think more people will understand it as the next generation of the Internet.
I firmly believe that the next generation of the Internet will definitely build an artificial intelligence network based on blockchain. Artificial intelligence plus blockchain is the real next generation of the Internet, because artificial intelligence and blockchain are both hugely transformative technologies. Among them, the influence of artificial intelligence is self-evident, especially the rapid development in the past six months. This shows that it can greatly improve the productivity of users, so it can develop so rapidly.
Then, why do we mention blockchain? Because we realize that the development of the Web2.0 generation of Internet companies needs to comply with the principle of the "barrel theory", that is, if you want to start a company, you need to constantly make up for the shortcomings. Therefore, in order to make up for the shortcomings, the company needs more and more employees. Or it depends on the ecology of large companies and needs to integrate into the ecology of a large company in order to survive better.
However, for the next generation of the Internet, we believe that this is a process of lengthening the long board, that is, companies with a few people can also survive and develop. Companies with a few people are very common in the Web3.0 field. In the Web3.0 track, we can see that companies with only 10 people can also get valuations of hundreds of millions of dollars, and companies with 100 people can also reach valuations of tens of billions of dollars. This was almost impossible in the previous generation of the Internet era. Since the underlying infrastructure of the blockchain has been well built, you can develop your own business on all these public facilities. You only need to do your best to have a good development.
The Paper: What made you decide to go all in on Web3.0?
Kong Jianping: I saw it at that time. The innovation of Web3.0 in the underlying technology made it a significant improvement compared with the traditional model. At the same time, we have observed that the development speed of this ecological development method of long board extension may exceed the past model. Because it lowers the threshold for entrepreneurship, allows more people to start a business here, and releases greater entrepreneurial value and user network effects.
The Paper: Please give us a detailed introduction to the main areas of your investment.
Kong Jianping: Our main focus is on the infrastructure of Web3.0, and we believe there are still many opportunities in this area.
At first, Canaan Technology mainly made computing chips for Bitcoin, and did not get involved in computing chips for Ethereum at that time. In fact, I think this is a field that should be entered. And not only Ethereum, we also pay attention to the development of the entire computing network, which is a valuable direction because we believe that computing may undergo several major changes.
First, chips may move toward application-specific integrated circuits (ASICs). In fact, this has already been demonstrated in the blockchain field, and we expect similar changes to occur in the field of artificial intelligence.
Secondly, the entire cloud computing may undergo a transformation from CPU to GPU and then to application-specific integrated circuits (ASICs). The reason is that both artificial intelligence and the metaverse require GPUs for computing, so the entire cloud computing architecture will be rebuilt.
Combined with the underlying logic and methods of blockchain, we believe that there are many things that can be done and subverted in this field, so we chose to participate in this field. Our company will include chip hardware and computing network sectors, and release more technological products and platforms around the Web3.0 metaverse ecosystem.
More low-level work is needed
The Paper: What does the metaverse look like in your opinion?
Kong Jianping: Everyone is talking about the metaverse. In my opinion, its forms can be roughly divided into three types: one is digital twin, which is to copy the real world in the metaverse; the second is the combination of virtual and real, augmented reality (AR); the third is digital native, which is to independently generate some worlds in the metaverse that do not exist in the real world.
In the past one or two years, we have seen domestic and foreign large companies, as well as the government, jointly promote the development of the Metaverse, but truly disruptive and revolutionary products have not yet appeared. The reason is that the development of the Metaverse is still in the stage of the wooden barrel theory, and is constrained by the existence of some shortcomings, such as the number of VR (virtual reality) devices, the demand for computing power performance, and computing power networks. There are still bottlenecks in many aspects. These factors have prevented the number of users from growing rapidly and the entry threshold is too high.
Therefore, we believe that to truly realize the development of the Metaverse, we may need to do more work at the bottom layer, such as solving large-scale computing concurrency problems. Our goal is to allow a large number of users to enter the Metaverse without feeling any lag, while reducing the operating costs of the Metaverse. Otherwise, if startups consume too much computing resources, they may not be able to afford the cost. This is the part that needs to be changed at the bottom layer.
At the same time, the entrance also needs to change. I think we need to form the rigid demand and stickiness of users among young people, and then gradually expand the audience. Generally speaking, young people and new scenarios are more receptive to new technologies. For example, Alipay was initially used by young people in a new e-commerce environment, and Alipay did not have various offline payment scenarios at that time. The user group at that time was not large enough, and it was impossible to make payments in offline scenarios. In addition, the network infrastructure at that time was also a limiting factor. So I think the development direction of the metaverse should be to tap some potential points in new scenarios. From the underlying infrastructure to the user entrance, corresponding changes need to be made.
The Paper: Following up on what you said above, you once mentioned: "Nowadays, many people may try to move the traditional world to the metaverse. This is a stock mindset, but I think there will be more native metaverse innovations." Can you tell me in detail what the "native metaverse innovations" will be?
Kong Jianping: For example, migrating the real world to the metaverse is the same as our thinking logic when we predicted that only five computers were needed in the world when computers were invented. This thinking mode is subject to certain limitations. With the development of artificial intelligence productivity, entities in the real world can be infinitely expanded in the metaverse. For example, in the real world, each individual has only one self, but in the metaverse, each individual may have countless selves, existing in the form of Avatars.
There are now applications that allow AI to talk to each other, even without human control, which was unimaginable in the past. In this case, your avatar and my avatar can communicate in the metaverse. People need to sleep at night, but the avatars in the metaverse do not.
The density of human knowledge is limited, but artificial intelligence like ChatGPT has raised the density of human knowledge to a higher level. Ordinary people cannot have their knowledge density. If you combine the elements of the native world and ChatGPT in the metaverse, then what you create in the metaverse will definitely be far greater than you in the real world.
The Paper: When people talk about the Metaverse, they often mention the word “distributed computing power”. When Haowei Technology mentions iPollo (Haowei Technology’s Web3.0 brand), it always mentions distributed rendering computing power. So what is distributed rendering computing power and how is it different from distributed computing power?
Kong Jianping: Our positioning is two-fold. The first is the bottom layer, which is to solve the problem of how to use computing power at a lower cost and make better use of idle computing power, because both the Metaverse and artificial intelligence cannot do without computing power. Our goal is to enable people to use computing power at a lower cost worldwide and solve the problems of artificial intelligence computing and Metaverse computing.
The second goal is to use the logic of Web3.0's long board extension to make various technologies on the iPollo platform Layer 2, that is, infrastructure for each other. For example, when a young person with innovative thinking designs a 3D model, first, he can use OpenAI and Stable Diffusion technology to generate this model. Then he can directly connect to Web3.0 to give this model vitality or artificial intelligence capabilities. He can also connect to an AR camera to allow this model to perform augmented reality interaction in the real world. Similarly, this model can also enter the metaverse developed by a certain project and become a prop or protagonist in it. In this way, even a one-person team can start a business in the metaverse ecosystem, and the number of team members is not a necessary condition.
The Paper: iPollo Pineapple Metaverse (a flexible computing network of NanoTech that provides Metaverse rendering computing services) completed large-scale real-time rendering verification of 15,000 digital humans in December 2022, and also established the first Metaverse 3D printing center. Can you give a detailed introduction?
Kong Jianping: In the past, we found that many Metaverse projects had performance issues when facing dozens to hundreds of concurrent users. Another problem was that the hourly operating costs of these projects were very high. To solve these problems, we successfully solved the concurrent problems of 15,000 people and other concurrent problems from a technical perspective. iPollo's goal is to reduce operating costs by 100 times and achieve this goal through real-time computing.
Optimizing the scheduling of computing resources only reduces costs at the first level. However, if you want to reduce the computing cost of the entire network by 100 times, you may need to deconstruct the computing methods and approaches in a higher dimension. Then let users operate according to this new set of computing methods and approaches, which will cost less than directly using computing resources.
We decided to build a Metaverse 3D Printing Center mainly because we foresee that the number of Metaverse assets will increase in the future, which will bring a large number of printing needs. However, the existing printing equipment and technology in the world cannot meet these needs, so we chose to cooperate with printer manufacturers to jointly build this 3D printing center that transforms the virtual world into reality.
With the development of artificial intelligence technology, artificial intelligence can now assist in generating 3D models, and may generate more sophisticated 3D models in the future. In a few years, what people need may no longer be an ordinary figurine, because it is static. What you need will be a digital trendy toy that you like, which has various AI functions behind it and can interact with you, and it will become a part of your emotional life.
Therefore, we believe that these problems need to be solved at the infrastructure level.
The Paper: You mentioned some of these earlier, but I still hope you can summarize them. With AI being so popular today, the popularity of the Metaverse seems to have cooled down compared to a year ago. What problems do you think still exist in the current development of the Metaverse? What are the prospects?
Kong Jianping: In my opinion, the problems facing the Metaverse can be roughly divided into two categories. The first category is that the platform has too many shortcomings, which limits the growth of its user base; the second category of problems is in terms of infrastructure, that is, there are challenges in concurrency and usage costs. Nevertheless, I believe that we are actively solving these problems. Looking ahead to the next three years, I expect there will be major breakthroughs and user development.
The Four Phases of Web 3.0
The Paper: You mentioned the development stages of Web3.0 when you were giving lectures to students before. Can you elaborate on it here?
Kong Jianping: I think there are four stages in the development of the Web3.0 industry.
The first stage was the blockchain stage dominated by Bitcoin. At that time, there were probably only 5 to 10 fields that the industry could participate in.
The second stage is the era of financial applications led by Ethereum. This stage not only includes the previous fields of mining machines, mineral transactions, chips, etc., but also covers various new applications. Therefore, the fields that developers can participate in may expand to hundreds or thousands of categories. There may be hundreds of thousands of entrepreneurial teams in this large ecosystem, including related teams in a broad sense, and there may be tens of thousands of truly in-depth participation.
The third stage is the era of native Web3.0.
The fourth stage is the era of industrial Web3.0. I divide it into native Web3.0 and industrial Web3.0 because once a concept becomes popular, many companies with resources want to join and consider how to combine. However, native participants and industrial participants are different groups, and they often cannot reach a consensus. The industry believes that technology that is not combined with the industry has no value, while the native side believes that they have done a good job and the industry cannot be combined well. In fact, neither side is wrong, they just look at the problem from their own perspective. When the number of users is small, only native gameplay can be promoted. Only when the number of users reaches a certain scale, such as 1 billion or 2 billion, can the industry be more easily combined and promoted on a large scale. Therefore, I divided it into these 4 stages, the first two stages are more financial, and the last two stages are more technological.
The Paper: What stage are we in now?
Kong Jianping: I think the current situation is that the second stage is dominant, and native Web3.0 is gradually emerging. Some people try to integrate, but it is not easy to form large-scale application cases. Just like the previous NFT (non-fungible token) issuance, it can exist as a marketing material, but it may be relatively weak in terms of sustainability. And the overseas native NFT, despite some ups and downs, the entire track continues to exist.
The Paper: So many people are working to popularize Web3.0, and more people need to understand and participate in Web3.0.
Kong Jianping: This is a question of user conversion. User conversion requires certain prerequisites. We want to guide users to switch from Web2.0 to Web3.0. This is not determined by one person or one project, but by an industry dividend. This industry dividend may be in several situations.
One is the wealth effect. People invest in stocks and cryptocurrencies because they see others making money. This wealth effect attracts everyone to convert users. This is one aspect. For example, in the current bear market, a lot of publicity is useless. When market conditions are good, they will naturally participate. This is the conversion of the wealth effect.
The second category is to improve user efficiency. For example, ChatGPT can increase user efficiency by a hundred times and also promote user conversion.
The third category is to significantly reduce the cost of use, which can also promote user conversion. I think user conversion must meet at least one of these three conditions, rather than just because you have users can achieve conversion.
The Paper: Can you elaborate on the three laws of the metaverse that you mentioned earlier?
Kong Jianping: The "Three Laws of the Metaverse" I wrote in 2021 are just my own ideas, and I hope to communicate with you more. My idea is, for example, the Internet had three laws in the past. If we regard the Industrial Revolution era as a two-dimensional space, the Internet era as a three-dimensional space, and the future Metaverse era as a four-dimensional space, you will find a characteristic, that is, the higher the dimension, the faster its development speed.
In the era of the Industrial Revolution, it took a company decades to grow into a unicorn. In the Internet era, the development time was shortened to more than ten years, and in the mobile Internet era, it only took a few years. In this era of the Metaverse, it may only take a few months for a company to grow. Therefore, I think that with the increase in dimensions, the speed of development has changed dramatically.
The first point is that the value of the metaverse is proportional to the cube of the number of users. In the Internet era, the user value is proportional to the square of the number of users, but I think in a higher-dimensional metaverse, it is at least proportional to the cube of the number of users. This should be the network effect of the industry.
The second point is that the network value of the metaverse is the sum of the values of each layer. In the past Web2.0 era, companies were in a competitive monopoly relationship, and each party was relatively closed; while the Web3.0 metaverse is more of an open relationship, so at the metaverse level, the value should be the sum of the values of each layer of the network, rather than a competitive relationship.
The third point of view is that marginal users of the Metaverse can also obtain corresponding value. In the past, you may have been a Taobao user, you shopped on Taobao, or you purchased a .com domain name, but you could not enjoy the dividends of these companies’ growth.
However, now you buy a product like BAYC (an NFT brand called "Bored Ape"), and as one of the earliest digital e-commerce consumers, the project also airdrops you tokens. In this case, every user can enjoy the increase in network value, which did not exist in the past. Therefore, based on these views, I think there are the above three laws. Welcome to criticize.
(Author Curiousjoe is a cross-border researcher of international politics and cryptocurrency.)