Data intelligence industry innovation service media
——Focus on digital intelligence and change business
"Metaverse" is a concept proposed by American science fiction novelist Neo Stevenson in "Avalanche" in 1992. In the book, it is assumed that people in the real world have a clone in the online world. The world is the "metaverse". Today, with the development of virtual reality technology, the metaverse has entered reality from science fiction.
For a long time, people have tried to use the metaverse to construct a digital space composed of many 3D virtual worlds, and make these worlds interconnected and interact with the real world. After the concept of the metaverse exploded last year, some people believed that the metaverse would be the outlet of the Internet in the next decade. However, the current level of technology is far from sufficient to support many of our scenarios for the Metaverse. After all, the Metaverse involves multiple technical fields. Integrating these technologies into a stable and efficient platform to achieve real large-scale applications still faces many challenges and bottlenecks.
At what stage has the metaverse developed so far, what key issues still need to be solved urgently, and where is the direction of future breakthroughs, especially how to integrate the metaverse with the big model and AIGC? In order to clarify these issues, Data Ape organized a live broadcast of an online roundtable discussion, hosted by Zhang Yanfei, co-founder and editor-in-chief of Data Ape, and invited Liang Zhihui, vice president of 360 Group and head of N World, chief architect of Inspur Information, Yuan Universe Ye Yurui, joint secretary-general of the Industry Committee and author of "Top Ten Technologies in the Metaverse", Guo Kuangye, deputy general manager and technical director of Ali Yuanjing, and other industry experts jointly conducted in-depth exchanges on the theme of "finding the key force for the breakthrough of the Metaverse".
Here is the full live replay:
The foundation is not firm, and the computing power must be increased by at least "10,000 times"
The development of the customer service industry can be traced back to the 1950s, when the customer service center was mainly responsible for telemarketing and after-sales service. With the rise of e-commerce, the business scope of the customer service center has gradually expanded, covering various forms such as online customer service, email customer service, and social media customer service.
Zhang Yanfei: Whether it is based on the low cost and portability of mobile phone face capture technology, or the improvement of audio-visual immersion and interactive experience of VR equipment, the industry is still solving the "entrance" problem of the metaverse at this stage. The first step in the future of the Metaverse. At present, the metaverse involves many core technical foundations, such as near-eye display, rendering processing, sensory interaction, network transmission, content production, etc. Which field do you think is relatively more hindered?
Ye Yurui: If the metaverse is understood as the third generation of the Internet, we are still in the construction stage of the metaverse from 0 to 1. If you want to get closer to the metaverse, you must first develop core underlying technologies including 5G, cloud computing, and edge computing. In addition, you need to upgrade hardware technologies such as chips, interactive displays, and computing. This series of infrastructure construction will be a complex systematic project. The generation of various materials and the connection of various links in the metaverse world must have a huge computing power support on the end side.
I mentioned in the book "Top Ten Technologies of the Metaverse" that the basic technologies of the Metaverse are divided into several levels, and the bottom one is the five foundations. "The Metaverse is a new digital world that is continuously shared and multi-person concurrent. To continue to run this new world, it needs a steady stream of computing power from the real world or the physical world. This includes computing, Storage, network, system security and artificial intelligence. These are the foundation of the metaverse or the platform construction of the metaverse.” With the foundation and platform, it is necessary to start building the scene of the metaverse, that is, the five pillar technologies of the metaverse, Namely interaction and presentation, digital twins and digital natives, technologies for creating identity systems and economic systems, content creation technologies, and governance technologies.
The short board is divided into two levels, the first level is the global scope, and the second level is China. From a global perspective, we regard this metaverse as the ideal state described in "Avalanche", but in fact we are still far away. For example, on near-eye display, the current resolution of VR eyes is about 2k-4k. If we want a more immersive and realistic experience, it usually needs to be above 14k; the refresh rate on the market is currently about 90fps, but if we want to achieve the ideal state, we need to increase it by more than two times to 180-240fps in order to have a real feeling. This also requires an increase in computing power. I once wrote an article saying that "the computing power of the metaverse must be increased by at least 10,000 times." China has many shortcomings, and technologies such as chips, computing power, and 3D engines also have a lot of room for improvement.
AIGC will inevitably improve the quality and speed up the realization of the Metaverse
Zhang Yanfei: AIGC (AI-Generated Content) refers to the use of artificial intelligence technology to generate content. It is considered as a new content creation method after PGC and UGC. We know that the development speed of AIGC in 2022 is astonishing, and the iteration speed is exponentially explosive. From the perspective of content, has AIGC advanced the process of realizing the Metaverse?
Guo Kuangye: AIGC will definitely help. To realize such a huge and complex virtual world as Metaverse requires the support of various cutting-edge technologies, and it is a long-term and gradual process. In order to obtain a close-to-real experience in the Metaverse, only human players are far from enough, and an infinite number of virtual scenes, intelligent NPCs, story engines, etc. are required, which can only be achieved by AIGC.
In fact, starting from chatgpt and large model technology, I think it has reached a qualitative change. There is a saying that "the highly compressed large-scale model technology for human knowledge can very well accelerate the efficiency of creative work", I very much agree with this sentence. We are very cautious and very optimistic at this stage. The current capabilities of large models cannot replace all content production. Now in the industry, there is basically a consensus that AIGC has formed a relatively good supplementary role in the original painting and some simple things in the art, and can achieve more than 50% of this part of the work. Efficiency improvement. However, in some aspects of traditional work, such as field editing, 3D model construction, etc., the current AIGC has no way to achieve practical accuracy. Although AIGC is very hot now, it has brought a lot of changes. But we have to be cautious and see where the upper limit of its specific effects can be achieved. Maybe in a year or two, we'll find something that's still hard to do.
Zhang Yanfei: Regarding the topic of AIGC, 360 should have a say. We know that the stock price of 360 has also changed a lot recently, and the secondary market has paid attention to the performance of 360 in the near future. From the perspective of large models and AIGC, what aspects of empowerment can you do for Metaverse?
Liang Zhihui: Since the concept of metaverse was born, artificial intelligence has been included in its territory. Metaverse has always been regarded as the ultimate form of the Internet, a digital universe composed of countless virtual worlds, in which users can freely create, explore, socialize, entertain, learn, work, etc. Its vision is to free human beings from the limitations of physical space and realize the seamless integration of digitalization and reality.
I think the empowerment of AIGC has two levels. One is that with the capabilities of AIGC and Chat GPT, a professional worker can become more professional and efficient through these tools; the second level, as I have been talking about before The metaverse of human beings, but now we can understand that AIGC's metaverse is coming, it doesn't just mean how beautiful or realistic the scene is. The significance of upgrading to AIGC’s Metaverse is not just to allow machines to generate content, but to allow this Metaverse to have a “brain” to support the entire Metaverse. Now everyone is discussing: Why should I go to the video and talk in the Metaverse instead of watching it face to face. The core reason is that the previous Metaverse had no thought and no intelligence. Then if there is intelligence, the metaverse should not only focus on generating content, or replacing some low-level human activities. Instead, we need to think: in the space of metaverse, can we use the so-called AIGC ability to help people solve real problems, rather than let people create a space environment similar to escapism. This is what excites me the most about AIGC.
Breaking through the suffering of computing power, the next step in the development of the Metaverse
Zhang Yanfei: 3D modeling and rendering require a huge amount of computing resources. Are the current computing resources enough? Is GPU + cloud computing + edge computing the golden key to untie the computing power bottleneck of the Metaverse?
Ye Yurui: Computing power naturally faces certain difficulties. The source of this difficulty in computing power is related to rendering on the one hand. Like the real world, the virtual world also has an imaging process. Taking photos in the real world receives real light, and the light enters the camera for imaging. But the difficulty of the virtual world lies in the propagation of light, its ejection between objects and its reaction with materials. These physical processes need to be calculated, not natural. The final rendering process also requires calculations, which is only the implementation of rendering. In addition to rendering, how to make the scene move is also a problem. Characters may move by animation or physical simulation, which involves a very large amount of calculations. Not only that, but also the real-time status of many people needs to be synchronized in it, which also includes complex calculations.
Zhang Yanfei: Looking at it now, the Metaverse has fallen into a "cooling-off period", but during this cooling-off period, the "outline" of the Metaverse gradually became clear. In the current situation, when the market enthusiasm returns to rationality, the next one or two days How will the metaverse develop in 2020? Which point is the main point?
Guo Kuangye: I am very optimistic about AIGC. This should also be a consensus in the industry: everyone hopes to make some industrial innovations by taking advantage of the booming technological wave under the condition of the development of productivity. Specifically, there are many points that need to be improved and realized. For example, the hardware equipment of VR is gradually mature and its performance is constantly improving, but it is still limited by equipment cost and comfort. High-performance VR/AR equipment is still relatively expensive, and wearing it for a long time may cause discomfort; 3D modeling and rendering also face resource consumption problems; distributed computing and storage need to continue to increase the level of computing power; large-scale metaverse scenes Real-time rendering also remains challenging.
In the past few years, we have been mainly serving the scene of the game industry. One is because the game industry has a relatively large scale and volume, and the other is that it also has relatively clear and high performance requirements. Since last year, many entrepreneurs in the metaverse application direction will ask us to do cloud nativeization, such as real-time rendering on the cloud, which can eliminate the need for particularly complicated downloads. This year has also fulfilled a lot of this direction. This is our limited attempt to advance the metaverse. We maintain a point of view. No matter how fast or slow the pace of advancement is, we are very optimistic that it will continue to flourish. There are more possibilities in the Metaverse. For example, in the cultural tourism and cultural museum industries, some browsing explanations can be realized. In short, the current development of the Metaverse may only be 1%, and there are still many things that can be combined, and even form a linkage with offline, which are relatively promising aspects.
Zhang Yanfei: In the process of providing this service, not only the cost of computing power and broadband requirements are involved, but also the content production. Many manufacturers make a green screen and shoot in the studio. What do you think can be done in the future? to what extent?
Liang Zhihui: We position ourselves to be the Metaverse marketing cloud in the marketing direction. We think that the Metaverse must be the Metaverse of AIGC. In the past few years, due to the general environment, we have seen that people are willing to do some office work online, and the effect is actually very good. We also found that whether it is training or some corporate gatherings, everyone is willing to use the metaverse to increase interactivity and fun. In this context, we want to create a digital space that can accommodate unlimited people. This digital space itself serves as a venue for marketing activities, and it is now a very mature business model. To do a thing, it is not necessary to instruct the machine to do it, but to do it by itself.
I think the computing power consumption may continue to increase in the future. For example, when we communicated with many customers, they saw the conference center of the Metaverse, and they were very happy to ask if they could hold meetings in this space on a regular basis, and use his image and voice to hold meetings with different people at the same time. In the future, we may see many companies that use a very simple camera to drive a digital human with an extremely high degree of simulation. This digital human has both the image of "I" and the thought of "I" . I don't think this day will be too long . It may be a long time before we estimate it, but it may suddenly come true in two years.
Metaverse is not a castle in the air. Its future is closely related to human life. Future fields such as education, medical care, retail, finance, urban governance, remote office, industrial automation, and autonomous driving may all be created through Metaverse. A whole new reality. Whether it is the carrier of entertainment life or the productivity tool of work mode, Metaverse is gradually breaking the boundary between reality and virtual world, and effectively helping human beings improve the quality and level of life through more diverse ways.
Perhaps, the metaverse of the future will subtly penetrate into our daily life like water, electricity, and signals. And these technology companies engaged in "Creating Yuan" are trying to use technology as the key to open the door to this metaverse.
Text: Mu Qingyi / Data Ape
