Blockchain is a value network, and its core purpose is to enable the more efficient circulation of digital assets. On October 12, Professor Zeng Ming, Chief Strategy Officer of Alibaba, held a public lecture titled "Looking at Ten Years" at Zeng Ming Academy and Lakeside Research Center. In this public speech, which took place six years later, Professor Zeng raised a series of new thoughts on business transformation, such as "How does technological change drive changes in business paradigms? How should corporate strategy follow the changes of a ten-year vision? What are the fundamental judgments about business transformation in the next decade?"
Jianle from the Wànwù Island community took notes and organized the content of this informative and sincere learning session, with some minor edits that do not change the original meaning, sharing it with you in the era of intelligent innovation and entrepreneurship, worthy of repeated reading and collection.
Full transcript of the speech:
Good morning, everyone. Thank you very much for your trust and for taking the precious time to attend this class. In 2017, I suddenly had an impulse to give a strategic lecture like "Looking at Ten Years." At that time, I was stimulated by two main aspects. One was that I had been studying strategy since 1993, teaching strategy, and practicing strategy in companies. Especially during that period, it was a time of great upheaval in the internet and mobile internet, so I had some different insights to share about how to approach strategy. The second was that I had grown up with the internet since 1991, witnessing its development over more than twenty years, and I had many speculations about the future that I wanted to share. Thus, the public lecture in 2017 came about.
The strategic public lecture in 2017 had two themes. The first theme was to redefine strategy because, in an environment of rapid change, complexity, and high uncertainty, gaining momentum with the trends is the primary principle of strategy, which is very critical. When we talk about looking at ten years, "looking" means Visioning, and this process becomes very important. The more difficult the times, the more seriously we need to look and strive to see. We need both the determination to "look at ten years" and to gradually cultivate the ability to see ten years ahead. This vision determines your pattern and potential. Strategy is the repeated iteration of Vision and Action, a concept I have shared many times over the past five or six years, and today I will upgrade this idea because I have gained deeper insights over the years.
The second theme I discussed last time was the great transformation of intelligent business. Online, networked, and intelligent development has constituted the theme of corporate development over the past decade. At that time, I drew a diagram based on the size of seven companies and their progress across various dimensions. Most of these companies are still among the world's leading companies today. These were the three most important development directions I discussed: online, networked, and intelligent.
The characteristics of intelligent business are, first, the ability to serve a massive number of users at low cost and in real-time; second, the ability to meet the personalized needs of each user; and third, the ability to iterate quickly. Therefore, intelligent business is essentially a technology-driven reconstruction of commerce based on networks and algorithms.
At that time, I mentioned two cores of intelligent business, which I called the double helix of DNA: one is network collaboration, which involves large-scale, multi-decision, real-time interaction; the higher the collaboration efficiency, the greater the value generated. The second is data intelligence, which essentially means machines replacing humans in decision-making. It is based on cloud computing, big data, and algorithms, forming data intelligence through rapid iteration. Thus, the two core components of intelligent business are network collaboration and data intelligence. At that time, I made two judgments: one is that the future business is in the preliminary stage of intelligent business patterns, and the second is that the future is an intelligent era, where human brains connect with machine intelligence. It is somewhat comforting that both of these judgments have been largely correct; otherwise, I wouldn’t feel comfortable standing here today. Most importantly, over the past six years, I have developed many new ideas and insights regarding this preliminary judgment, so the core of today’s sharing is a deeper exploration of these two themes.
We will unfold this in three topics. The first is the true arrival of the intelligent era, as we have AGI, the revolution of general artificial intelligence; blockchain and crypto have undergone nearly 15 years of brewing and development and are ready to take off; the third is XR and the metaverse. These are the three core technologies, and they are the three areas we will focus on in this morning's lecture. The second part will share a methodology with you, which is how to understand the actual process of technology-driven business transformation. Through this methodology, we can understand what is most likely to emerge in the next three years or three to five years. This is a very critical milestone in strategic decision-making. You need to know that aside from looking at the long-term vision of ten years, how to set the goals for the next three to five years? This requires a mid-term judgment. So, the second section will discuss how to make this mid-term judgment. The third section will talk about some new thoughts on intelligent business.
The impact of artificial intelligence on future business
We will begin the first phase of the discussion, which is the impact of artificial intelligence on future business. This diagram may be familiar to everyone; it shows the significant development of artificial intelligence over the past 20 years. In the early search phase, it was called big data, and at that time, the term AI had not yet emerged. As you know, after the popularity of ChatGPT at the end of last year and this year, there are over 100 entrepreneurial teams in China focused on large models, known as the "Hundred Model War." In fact, the second phase, during the time of facial recognition, was when deep learning was first applied on a large scale in the visual field, with over a hundred visual companies being established in 2014. Facial recognition, which everyone now feels is ubiquitous, was actually the first large-scale application of AI using deep learning methods. The recommendation engine behind Douyin that everyone uses daily is also based on AI technology. Large language models, known as Large Language Models, represent a revolution in general AI. It is essentially a very simple algorithm that predicts the most likely next word following a given word. This simple algorithm has achieved a level of predictive accuracy that is sufficiently high and useful. In this sense, it appears to have mastered language. As mentioned in the book "Sapiens," language is humanity's greatest invention. Language allows us to communicate, and behind language inherently lies human wisdom, as well as the vast knowledge accumulated over the last 10,000 years, which has been distilled into text, audio, and video over the past twenty years of IT development. Therefore, mastering text and language essentially unlocks all human knowledge to date. We still do not fully understand the operational mechanisms behind large language models. They may not think like humans, but in specific areas, they exhibit human-like logical reasoning abilities. This will have fundamental implications for our future.
The development over the past thirty years, from the internet to wireless internet, to sensors, digital transformation, big data computing, etc., has gradually enhanced the capability boundaries of the software world, but it is essentially additive, a stacking process. However, AGI, or general artificial intelligence, connects these together, enhancing the adaptability and autonomy of all software, leading to a qualitative leap from quantitative change. For example, AGI can automatically program, which dramatically enhances software capabilities, representing a qualitative change. In this sense, it is generally believed that large language models represent the first iPhone moment of the AI era, marking a time of great transformation.
From another perspective, the era of general intelligence can also be seen as the era of robots, as AI serves as the brain, and its combination with various hardware forms different types of robots. For instance, autonomous driving vehicles are robots, and particularly in the future, Robotaxis will essentially be technology outsourcing service companies. Understanding it from this perspective will provide a more fundamental insight into the impact of technology on business. When people mention robots, they often think of various impressive robots from Boston Dynamics, but Boston Dynamics has been developing for about thirty years and may not have progressed as quickly as Tesla's humanoid robot has in the past two years. This is also a breakthrough brought about by AI technology in hardware, and we can see that robots will develop rapidly in the entire environment.
In addition to ChatGPT, I want to emphasize that the other two main lines of AI and AGI development are also very important. One is autonomous driving, which has different requirements from ChatGPT; it must ensure safety and fundamentally addresses the interaction between humans and the physical world. ChatGPT focuses more on human brain behavior. However, autonomous driving must solve the interaction between humans and the physical world, which is why Tesla, as an autonomous driving company, has accumulated so much in robotics, as it fundamentally needs to perceive the external world. Another very important area is AI for Science, which is even more fundamental. So far, AGI can only apply existing human knowledge and cannot create new knowledge. However, AI for Science uses AI to advance scientific development. It is highly likely to create entirely new paradigms, as it may discover new chemical equations or new physical laws, propelling artificial intelligence forward significantly. Even today, projects like DeepMind's AlphaFold for protein analysis and synthetic biology, which is a very new field in the past few years, are also driven by AI. Many fields have already made significant progress, though they may not be widely known, but these accumulations will lead to breakthroughs in the next steps. The previous section provided some background knowledge that you may have heard in different contexts, and the next two slides are among the most important slides of today.
The difference between AI and the internet era
We have transitioned from the internet era to the intelligent era, to the AI era. So what is the essential difference between the internet and AI? The internet essentially deals with massive amounts of data, solving the efficiency of information flow and matching. Its core value lies in addressing information asymmetry, allowing information to flow and match as much as possible, avoiding various frictions caused by information asymmetry. However, in the AI era, the essence of AI is to handle vast amounts of knowledge; it is no longer just data or information, but knowledge generated through processing data and information, which combines with existing knowledge to solve practical problems. Therefore, it addresses the efficiency and cost of decision-making. In other words, can machines replace humans? Because so far, all decisions have been made by humans, and if machines can replace humans in decision-making, it represents a qualitative leap. Its core value is actually to create new supply. This is something I have felt deeply over the past year. Initially, we were all worried about whether AI would replace humans in the future. There are many areas to discuss, but today, in practice, we see that the earliest users of AGI services were those who previously could not afford human services because human services are very expensive. Therefore, AGI services actually provide new supply.
Let me give two simple examples: online education. The previous wave of online education development aimed to use the internet to improve the teaching efficiency of high-quality teachers, which is a very typical effort of the internet and has achieved significant progress. However, online education in the AI era is about providing unlimited high-quality teachers to meet personalized learning needs. In principle, each student should have their own teacher, which can only be fulfilled by AI teachers. Similarly, one of the biggest problems in the world today is the high cost of healthcare and insufficient doctor services. If AI doctors emerge, the overall health status of everyone will experience a qualitative leap. Thus, AI essentially addresses the issue of insufficient supply.
In the past five years, the reason why digital transformation and online industrial internet efforts have been so challenging is fundamentally that these industries are not facing information asymmetry issues but supply shortages. For example, all those working on internet hospitals and medical service transformations face limited value because they cannot solve the core problem; the bottleneck in medical treatment always lies in the limited number of good doctors. No matter how you match information, it is ineffective. Therefore, the AI era brings a completely new opportunity, as we can truly create new supply, and massive supply will create new demand.
The most critical capability in the AI era is to establish decision models based on decision-making scenarios. I will elaborate on this term later; it is very important. All our decisions are based on specific scenarios. Many times, human decision-making is subconscious or even unconscious. How do we make these spontaneous decisions explicit so that machines can implement them logically? This is a fundamental challenge, and all the difficulty lies here, especially for AI application companies and some cutting-edge companies working on large models, where algorithms may be a significant bottleneck. However, for AI applications, the core is modeling capability, understanding decision-making in real-world scenarios. This difficulty arises because the decision-making methods of AGI differ from those of humans, so you need a translator. For example, applied mathematics has become a popular elective among undergraduates over the past decade because its core is modeling. This is a very important core capability. The interesting thing about this model is that once you establish it and form a closed loop, it can continuously iterate and optimize itself; it becomes a living AI system. In this sense, all our past developments can be described as a machine era. Even the most complex mechanical systems are simple systems that can only perform deterministic executions. However, even the simplest cognitive systems are complex systems. Therefore, AGI is developing a system that can grow organically, similar to biological systems. This will also be a fundamental development. How do we embrace a system that has certain capabilities and tendencies, along with self-learning and self-growth abilities? This is the essence of AGI, which differs from the internet era. The internet era was still about solving relatively deterministic information matching problems, while the AI era is about building cognitive systems. This is the first point I want to share with you today.
To summarize, based on the public lecture in 2017, I want to elevate the discussion: the driving force of the era is intelligence. I elevate "intelligence" to a higher level, making it the dominant force of this era. The internet era was essentially characterized by online, software, and networking, with the combination of online and software being the most popular SaaS over the past 20 years. Networking refers to the series of developments from PC internet to mobile internet to IoT. Its essence is connection, completing the infrastructure for network collaboration. Every new era builds upon the foundation of the previous one, so we can see the new driving forces of the intelligent era on the increasingly improved infrastructure of the internet era. On one hand, there is intelligence, which we have been discussing throughout this lecture, especially the growing power of general artificial intelligence. We do not know how powerful it will ultimately become; we only know it will become increasingly powerful. On the other hand, two foundational technologies support the development of the intelligent era: 1) the continuous enhancement of human-computer interaction capabilities, which is the XR topic we will discuss shortly, and 2) blockchain and crypto, which enhance our ability for network collaboration.
XR: Human-Computer Interaction
Next, we will delve into the discussion of XR technology, which encompasses AR, VR, and XR as a whole, representing a development process in human-computer interaction. First, let me briefly discuss the background. Ignoring the console era, which many people may not have seen, starting from the PC era, we can think of some of today's most impressive companies, such as Microsoft and Apple. The core invention was the GUI (Graphical User Interface), which led to all the internet revolutions we see today. From personal computers to mice to keyboards, the essence was keyboard input, leading to Microsoft's entire software system. Then came the mobile internet era, primarily characterized by touch screen input, along with some voice input. The third path began to develop in the past decade, with Oculus established in 2012 and acquired by Meta in 2014, leading to the emergence of VR headsets. Google Glass also appeared in 2014, and a batch of products was launched in 2015. In 2016, there was great excitement, dubbing it the "Year of Virtual Reality," as the first generation of Oculus Rift was released, Sony launched its own VR headset, and the popular game Pokémon Go made headlines. I remember taking my child to Yokohama to catch monsters for the game. That was the first hot game based on virtual reality, but it quickly fell silent for a while. As is well-known in the trajectory of high-tech development, there is often a phase of cliff-like stagnation. In 2019 and 2022, everyone was working hard. Magic Leap seemed to be a promising startup at that time, receiving support from companies like Google and Alibaba. In 2018, I visited Magic Leap to see their next-generation product nearing production. I was deeply impressed; it was not about the distinction between real and fake, but about the future where one could not distinguish between the two. The effect was such that it could completely confuse your eyes, as it provided real light sources to your eyes, making it impossible to judge whether what you saw was real or fake; it presented formed images and sent signals to the brain. That was my first impression. The second was that the founder of Magic Leap told us on the first slide of the presentation that they were not making glasses; they were creating the future of human-computer interaction. Just think about it: if you could execute commands just by moving your eyes and looking at the computer, wouldn't that be much faster and easier? Unfortunately, they encountered some final technical challenges and did not succeed in becoming a consumer product. However, this year saw two significant releases: one was Apple's VisionPro, marking Apple's first official product launch in this field, defining many new standards and raising expectations. The second was the release of Meta's Quest3 last week, targeting the mid-to-low end, while Apple focused on the high end. The technical routes chosen by both companies are largely consistent, indicating that industry standards are beginning to emerge, with both high-end and low-end options available. Additionally, Meta launched AiGlass, also aimed at the development of human-computer interaction. Although it is not a virtual headset product, it shows that visual interaction has once again become a focal point in the industry.
Returning to the hardware discussion, what is the core purpose of hardware? The core purpose of hardware is to create new opportunities for human-computer interaction. We mentioned that early PC computing was through keyboard input. Mobile computing on smartphones was through touch screens, and in the era of spatial computing, the core emphasizes vision and perception. Everyone's definition may differ; we do not need to get into the details.
I want to summarize why the XR field is crucial for everyone present. What is the essence behind its technology? This represents a qualitative leap in human-computer interaction. Previously, our interaction with machines, including AI, required us to actively operate the machines; we had to input commands. However, in the future, machines will actively respond to humans. We may not need to do anything; they will naturally perceive us. If we evolve into a brain-machine interface, they may even subconsciously know what we want and execute it. Thus, the future is about machines perceiving humans and taking proactive actions in the interaction interface, marking a completely different era. We will see more and more machines connecting human senses directly to the digital world. Currently, we have AR and VR glasses, wearable devices, and even clothing that resembles skin. The distance will shift from far to near, close to the skin, and eventually into the skin. The implantation of chips will inevitably happen sooner or later. This represents a significant development trend over the next ten to twenty years. What is the commercial significance of this trend? Starting with XR and VR glasses, we are beginning the digitalization of human perception and attention; humans are no longer independent of the digital world. We may be becoming Digital Natives, and in the future, we may all be digital natives, leading to different definitions of what it means to be human. We will become part of the digital world, which is very important.
The reason why the metaverse was once so yearned for is that it is a purely digital world, unbound by the laws of physics, where extreme personalization can be achieved, along with rich biological characteristics and diverse scenarios, offering endless services. The excitement surrounding the metaverse was also a hopeful vision for the future. However, XR and similar devices, in addition to hardware, also require improvements in software and computing power, which correspond to edge computing and algorithm miniaturization. In the future, each edge device will experience a qualitative leap in perception, computation, thinking, and decision-making capabilities. Therefore, this technology and AI technology complement each other, providing an infinitely broad scenario for AI to be applied more widely. Conversely, it will also promote advancements in AI technology because without progress in AI technology, it cannot support deeper, more complex, and real-time technical requirements. Thus, these two technologies are entirely complementary.
Blockchain and Crypto
Next, we will discuss the development of the third technological field: blockchain and crypto. Some friends may not be very familiar with this area, and it can be complex to elaborate on, so I will keep it concise. First, you can note some conclusive points to digest later. Why is blockchain technology so closely tied to crypto? It actually started with the first cryptocurrency, Bitcoin (BTC). In 2008, Satoshi Nakamoto published the white paper, leading to the mining industry and the emergence of Bitcoin, which gained consensus from a significant number of people. Bitcoin now has the consensus of hundreds of millions of people, many of whom have traded or purchased Bitcoin. This is an interesting new form of trust and consensus based on technology and algorithms.
Blockchain technology evolved from Bitcoin, and on this basis, Ethereum was developed, becoming a universal technology platform for smart contracts. Ethereum has undergone three rounds of development. The first round was ICO (Initial Coin Offering), which was particularly popular in 2017. Issuing tokens was the first smart contract, the simplest form of online automatic token issuance, forming a set of rules and systems. Token issuance was Ethereum's first killer application. Building on this, in the summer of 2020, decentralized finance (DeFi) emerged, which essentially restructured all simple financial services using blockchain technology, based on the concept of transitional collateral, replicating all simple financial services under controllable risk. This was also a remarkable achievement. Based on the accumulation from DeFi, GameFi began to appear in 2021, and many friends have played some GameFi games, including StepN's running shoes, which belong to GameFi games. Then there were NFTs (non-fungible tokens). Each product is backed by a type of smart contract, so these applications have promoted Ethereum's development in rounds. Of course, Ethereum itself is also undergoing scaling and layer 1 and layer 2 developments.
Fundamental challenges of blockchain technology
In 2022, the industry faced numerous negative events, with many instances of collapse, and the entire year saw no new developments. As a result, many people were confused about whether this field still had a future. Even some staunch believers began to waver. To answer this question, we must first address what the essence of blockchain is. The essence of blockchain is a value network; it is not an information network. The internet is an information network, but blockchain is a value network. Its core purpose is to enable the more effective circulation of digital assets. Another byproduct is that issuing tokens online has become very simple and reliable, allowing for a series of innovations in new incentive mechanisms. These are the two core breakthroughs of blockchain, fundamentally a breakthrough in production relations. This is a technological innovation aimed at transforming production relations.
This brings significant challenges because it is not a tool for productivity. It is difficult to enhance user experience, so the blockchain field has been waiting for a good application that can reach hundreds of millions of users to unfold the entire system. Therefore, from this perspective, the fundamental challenge facing blockchain is the lack of technological innovations that can directly enhance consumer experience. This field itself lacks such innovations. Secondly, they originally hoped to digitize traditional assets, such as various financial assets, but this has not progressed smoothly because the efficiency they provide and the value they create are not significant enough. At the same time, there are enough traditional interests and existing systems to maintain, so this transformation has also failed. Moreover, without new applications, there are no new digital assets. Without digital assets, having a value network to reduce the circulation of digital assets becomes meaningless, like a tree without roots.
What developments can we expect in this field next? One is that, following the current logic, it will continue to develop. Bitcoin will continue to serve as an alternative asset; in a certain sense, BTC will continue to move towards greater consensus, or Bitcoin may play a larger role in payments, promoting the development of inclusive finance based on payment networks. This is one path of innovation along the financial main road. The second is to rely on the development of new applications. In the past two years, many innovations in GameFi and SocialFi have accumulated, and perhaps in the next six months to a year, we will see some preliminary results.
AIGC: A significant breakthrough in productivity
I believe the most valuable breakthrough will be the creation of massive new digital assets through AGI. The first breakthrough area of AGI is AIGC (AI-Generated Content), which refers to deep-level AI that creates vast amounts of content. At some point next year, there will definitely be very useful tools for converting text and speech into video. Essentially, the barriers to creation will drop sharply from text to speech to images to videos, and the space for creating new digital assets will rise dramatically. Moreover, as we discuss the virtual world, these future digital assets will increasingly become mainstream, and their importance will grow. These assets will have value, and people will pay attention to them, leading to active circulation and trading. Therefore, on this basis, new digital assets will naturally utilize new value network technology platforms.
At the same time, as I mentioned, the core of Ethereum is smart contracts, but in the future, the cooperation between machines will differ completely from that between humans. They will require more, more automated, more efficient, and smarter contracts to be completed directly. Therefore, in this field, blockchain and crypto have vast development space, and in this sense, I also consider it an important component of the entire AGI intelligent era.
Whether from the earlier discussion of the crypto field's call for a creator economy or the value brought by AGI, I believe we are about to enter an era of creator economy. On one hand, this trend is very clear: AGI will gradually replace structured human knowledge and become increasingly intelligent. On the other hand, humans, with the help of machine intelligence, will have the opportunity to become more creative. Just like in the early days of the industrial revolution, there was great fear that humans could no longer derive value from physical labor and could not survive on physical strength. However, over the past 100 years, the white-collar class, knowledge workers, and software engineers have emerged, relying on their intellectual activities to create the prosperity of the past 100 to 200 years. I can envision a relatively positive scenario where machines or artificial intelligence liberate humans from tedious, repetitive, and boring mental labor, allowing people to spend most of their time developing their creativity and engaging in things they are truly passionate about and can excel in. These may be the two fundamental driving forces. On this basis, collaboration among humans, between humans and machines, and among machines will raise higher demands. In the internet era, collaboration between machines relied on APIs, where applications had to have agreed-upon standards for mutual assistance. However, due to the development of AGI, in the future, all services will interact using natural language. In other words, machines will learn to communicate like humans, completing collaboration between machines. Natural language will become the communication language among humans, between humans and machines, and among machines, raising higher demands for smart contracts.
If we look at these contents from a more macro perspective, Peter Drucker may be the greatest business thinker of the 20th century. He divided the industrial revolution into three historical stages. The first stage is the revolution of productivity, where factories replaced handicraft workshops. Traditionally, knowledge in handicraft workshops could only be passed down from master to apprentice. However, with the advent of factories, scientific management began to emerge. The second stage is the management revolution that began about 100 years ago, where the concept of enterprises emerged. Previously, there were only individual factories, focusing on production and sales. However, with management, matrix management and functional management emerged, leading to the establishment of human resources departments, strategic planning departments, etc. Business schools were established to supply a large number of high-quality managers who could be produced in bulk for the management revolution, which is also very important. The MBA program emerged as a standardized language, representing a set of commercial training for management. With the development of IT, starting in the 1960s and 70s, we entered the software revolution, where software engineers created the most value. Following the earlier discussion of AGI's replacement of structured human knowledge, humans must move towards the development of creativity. Therefore, I define the future fourth development stage as the stage of the revolution of creativity, where human value will be reflected in creativity.
We are about to welcome a new economic paradigm. The core of the intelligent economy can be understood as the creator economy from another perspective. The three core supports are the general artificial intelligence, crypto, and AR&VR we discussed earlier. Of course, these three development stages are different; currently, AGI is developing rapidly, crypto is in a relatively low valley and is brewing, while AR&VR may take another three to five years to produce large-scale sales of application-level products. However, the trends of these developments are very clear.
The evolution of human civilization
If we step back from the intelligent economy and look at the evolution of human civilization from a broader perspective, the core of human development relies on two aspects: one is the development of human networks, which includes language, writing, culture, systems, etc., all of which are considered soft institutional elements. The other very important aspect is the continuous creation of tool networks by humans, from the earliest fire to the use of tools, to agriculture, to physical networks, and today’s logistics, communication, and computing networks. The development of tool networks has promoted social progress and human development, and humans have invented more networks and tools, facilitating the development of new technologies. Therefore, technological progress and social progress have produced qualitative leaps through these two networks, leading to successive rounds of development.
If you look closely from a biological perspective, the capacity of a single human brain has only slightly improved. The progress is twofold. First, the development of the brain has gradually unfolded, and we still have a low ratio of brain development. This is why the creativity revolution is possible; we may develop many abilities we cannot currently imagine. Second, the emergence of collective intelligence is currently more important, achieved through social collaboration. In fact, the society we create generates greater value and accelerates development.
The progress of the tool network driven by technological change is the main line of human civilization development. Based on this, we can make a judgment about what state we are in today. From the discovery and application of fire to the use and invention of tools, to the agricultural economy, which is only about 10,000 years old, and then to the industrial revolution. The first industrial revolution was powered by mechanical energy, and the second industrial revolution was powered by electricity. Although some people refer to the information revolution as the third or fourth industrial revolution, I personally believe that separating the information revolution as a distinct concept may be clearer. Thus, we have the first information revolution, which was the invention of computers. The second information revolution occurred roughly from the late 1970s to the early 1980s, with the advent of personal computers and the internet, culminating in the integration of communication networks and computing networks, leading to the explosion of the internet over the past 20 years.
The past five years and the next five years represent a transitional period from the internet era to the intelligent era. I personally prefer to call it Internet 3.0, transitioning from Internet 1.0 (PC) to 2.0 (mobile) to 3.0 (the future). To clarify the concept, we can define the next decade, or even the next two or three decades, as the beginning of the intelligent era. Intelligent Era 1.0 is the opportunity of the era we are currently in, as well as its challenges. Regardless of our position today, everyone faces a common challenge: to become a native species of the intelligent era, as this is the only way to develop and even survive. This is the macro picture I want to convey.
The basic laws of technological change driving business transformation
Next, I will discuss some prospects and technological developments for the next three to five years, which is the second part of today’s content. Sometimes, looking ten years ahead may not seem too difficult; everyone can talk endlessly about the future. However, how do you project this ten-year vision into the next three to five years? Because your strategic core is formulated around these three to five years. A ten-year vision, or even further, is about vision. How to view the next three to five years, especially during significant technological changes, amidst such uncertainty? I have repeatedly pondered this question, which was triggered again by the rapid rise of ChatGPT, the fastest application to reach over 100 million users. Is ChatGPT the star of tomorrow? Is it the next Google we have been waiting for? This is a question I want to answer. After much reflection, I believe there is a concept I can share with you: the emergence of native applications, or services, referred to as native apps. What does this concept mean?
Let’s first look at the basic laws of technological change driving business transformation. First, a significant technological change often brings several waves of business transformation. Of course, during this process, it nourishes the technology itself, which also progresses and matures. We can see that the internet has gone through the first wave of PC internet, which we can consider commercialized from the listing of NetScape in 1993 to the emergence of Apple's App Store in 2008, which opened the door to mobile internet, and then to the later Internet of Things. Similarly, AI has a historical trajectory, having gone through the era of big data, then AI 1.0, and now AI 2.0, which may be AGI 1.0. Thus, it is often a wave after wave of continuous development until the technology matures and is replaced by new technologies.
From another perspective, I categorize the business transformations driven by technological changes into four stages. The first stage is very early development, during which there will definitely be bubbles. This is because it shows people too many possibilities, but the progress in realizing these possibilities is far below expectations. Thus, it is an exciting time, but this bubble will eventually burst. The internet bubble is the most memorable instance. Why is the internet bubble so memorable? It is because before the internet bubble, especially during the stock market crash in March 2000, people had experienced a century of steady development during the industrial era, becoming accustomed to linear growth. Suddenly, the internet emerged, representing exponential growth, and the potential seemed limitless, leading to a sudden and disruptive transformation. However, during the mobile internet era, people had the experience of the first PC internet, so the excitement was not as high. Therefore, we can see that significant technological advancements often come with various bubbles, then enter a penetration phase, and finally reach native applications, becoming a general technology that almost every industry will use, turning into infrastructure. This is how the internet has become the infrastructure of society.
Another important point is that infrastructure and applications evolve together. When we look at native applications, we will see this very clearly. Native applications typically emerge in the third stage of a technological revolution; they need time to mature. The technology must reach a certain level to create new value, but at this point, it can bring real mass-market users, becoming a national-level killer application, just like WeChat in the mobile internet era. Ultimately, it naturally becomes the leader of the new business paradigm, making it difficult for followers to catch up.
For example, Google is the first native application of the PC internet. This is my own judgment, and there may be different opinions on this. Some might argue that Yahoo or eBay could also qualify, but in terms of completeness, Google is undoubtedly the king of the PC era. It achieved this because it completed disruptive innovations on several levels. One was the search box. When Google introduced its minimalist search box, it was absolutely shocking. Moreover, it could return all the information on the internet in seconds based on a keyword input. This was previously impossible and represented a significant breakthrough in user experience. Such a breakthrough requires substantial innovation in underlying technology, which is what we refer to as cloud computing, or from a technical perspective, distributed computing. Today’s AI computing is essentially built upon the development of distributed computing. Thus, it initiated a new era of computing, but equally importantly, it established a new business model called Pay for Performance, which is the familiar concept of precision marketing today. It transformed advertising costs from an immeasurable metric into a precisely measurable one, allowing us to see how much we spent to acquire how many users, and payment occurs only after the customer clicks. Moreover, the price is determined by the market; if someone competes with you, the price goes up, and if no one competes, the price goes down. Through such market pricing, it can fully utilize massive clicks, leading to a significant siphoning effect, where advertising moves online, and online advertising flows to Google. This resulted in Google’s glory for over a decade, during which all the talent in Silicon Valley flocked to Google, and all innovations originated from Google, resulting in extremely high profit margins and rapid growth, culminating in a monopoly in search. This is a very typical example of a native service that initiated a new era.
From this perspective, we can observe the emergence of native services brought about by technology. As mentioned earlier, Google is the native service of the PC internet, founded in 1998 and listed in 2004. The second native service is Facebook, founded in 2004 and listed in 2012. Facebook is a very typical native application of the PC era. However, the year it went public coincided with the explosive growth of mobile internet, causing its stock price to drop by 40% immediately after its IPO. This forced Facebook to rapidly transition to mobile internet. In 2006, Twitter emerged, and in 2007, the iPhone was released, followed by the App Store in 2008. By 2009, the first batch of super applications began to appear, including WhatsApp, Weibo, and Uber. In 2010, Meituan and Instagram emerged, followed by WeChat in 2011, Toutiao in 2012, Kuaishou in 2013, and Douyin in 2015. Today, our lives are largely defined by Douyin and Pinduoduo. These are the true kings of the mobile internet, and the most native applications emerged densely during this period.
AGI era's native services
The next two slides are also very important for those working on AGI, as they may provide significant help in assessing whether you are truly at the forefront. The first is whether you are using the latest AI technology to engage in natural language dialogue, as large language models have solved the language problem. Therefore, you can engage in natural language dialogue and interact deeply, continuously, and communicatively with users through future XR glasses and wearable devices that provide visual spatial perception. Essentially, being always online will become the default in the future. The second very important point is that by cracking the language, you have unlocked the totality of human knowledge, specifically the totality of text-based knowledge, meaning you can access the entirety of human knowledge at any time. This is what all trained models accomplish. The third point is that it involves a certain level of reasoning ability, meaning it helps you make decisions. The purpose of using this technology is to create a qualitative leap in user experience. How does it redefine products based on scenario decisions, and can you effectively utilize large language models to leverage general knowledge? In this scenario, what specialized knowledge and skills do you need, and can you access the relevant knowledge and skills in real-time? Finally, innovative interaction is also very important because mature hardware can better support the underlying technology.
If we look at ChatGPT from this perspective, it may still be a semi-finished product. It has indeed created a new way of human-computer dialogue, whereas Siri only addressed the level of voice recognition accuracy. However, it cannot engage in semantic dialogue. Large language models have solved the problem of natural language dialogue, making dialogue a very mainstream interaction method in many contexts. In many cases, it is the most efficient. However, the product form of ChatGPT is very outdated, reverting to the most primitive PC webpage format, indicating significant room for innovation. The new user entry point may not necessarily be ChatGPT, as its user growth has clearly slowed down. Simple dialogue, akin to an encyclopedia, and some basic writing assistance may not constitute a killer application, and no new business model has emerged. Therefore, from these perspectives, I believe ChatGPT has merely sounded the horn; it is not a truly native application.
The future of Web 3.0
In the next three years, I believe the most noteworthy aspect is which entrepreneurial teams, including a few giants, have the opportunity to launch truly native applications, which will lead to an explosion of this era. Based on the concept of native services discussed earlier, I view the development over the next three to five years as follows: the next two to three years will be a nurturing period, and perhaps in some corner, there is already an entrepreneurial team working on such initiatives. We will see mainstream services emerging from AI and crypto, which may bring entirely new user experiences. The emergence of such mainstream services will drive a surge of native innovative services, potentially becoming an ecological entry point or platform infrastructure, similar to Apple's App Store, catalyzing the emergence of a series of killer native applications. Just as we saw a continuous emergence of significant applications between 2009 and 2016, it is likely that in five to ten years, the original leaders will begin to take the lead, followed by the emergence of a few second-generation, more native applications. Approximately ten to fifteen years from now, the first batch of leaders in the intelligent era will have established their leadership positions.
In the next three years, it will depend on who these native services are and who is most closely tied to that ecosystem. Who has the greatest development potential? Currently, it appears that, based on history, all three technologies may use gaming as a primary breakthrough point. The promotion of AIGC in gaming is evident, as is the application of GameFi in the crypto field and VR gaming. Therefore, gaming will undoubtedly be a major application. However, the metaverse may emerge as a true, native super application in about ten years, integrating innovations and making digital life a genuine part of our existence. However, the metaverse requires the maturity of these three technologies before undergoing another round of integration, so it is certainly not something that will happen within five years. This is my judgment about the future, for your reference.
The tipping point of intelligence: machines replacing humans
Next, we will discuss the third part of the content. With this macro perspective, how do we view specific changes in business? First, let’s look at the paradigm revolution of intelligent business. This has not changed much from the definition in 2007; it has just become clearer: intelligence, machine algorithms, and AI replacing humans, continuously evolving to make increasingly intelligent decisions, thus significantly enhancing user experience and business efficiency. The more steps replaced, the more complete the roles, the greater the value created. However, the ultimate goal remains to provide real-time, precise, low-cost services to massive users. The successful cases of 1.0 are well-known, from the early days of Taobao shopping to the browsing of short videos on Douyin, to the automatic scheduling of Didi rides and Meituan deliveries. Why has this round of development accelerated? It is because breakthroughs in AI have led to a qualitative leap in machine capabilities. Moreover, an increasing number of decisions will be replaced by machines, and they will become increasingly intelligent.
This may also be a very important slide that you will repeatedly apply in practice. How do you achieve intelligence? Over the past 15 years, everyone has struggled with digital transformation. I later realized a fundamental issue: the value created by this technology is insufficient. We are laying the groundwork for intelligence, so how do we achieve a breakthrough in intelligence? The core is the scenario because decisions are certainly made based on specific scenarios. So, who are your users? What challenges do they face in which scenarios? Then, based on that scenario, you need to provide a complete solution and service. That is the success of intelligent transformation. Therefore, on the left, you may need to call upon certain standardized products or service modules. In the future, there will be no product companies; there will only be service companies. Products are merely tools and carriers to meet the needs in specific scenarios.
We have always imagined scenario-based e-commerce, and I only now understand that scenario-based e-commerce is possible today because you must make decisions based on that scenario. You need to coordinate all knowledge and expertise to provide the best service solution for a specific person in a specific scenario at a specific time, all in one go. This is the personalization of the intelligent era, and in this sense, the C2B business model can truly be established. Or more completely expressed, it is a user-driven business model, which may be C2S2B2B, where the "S" represents the intelligent platform, as consumers need this intelligent platform to directly integrate all possible resources and provide them with the most personalized, real-time solution. In this sense, we are moving towards a broad, on-demand phase. The "S" in C2S represents the AI Agent, which is the living, growing AI system that continuously learns and grows, making better and smarter decisions.
2023-2033: The decade of nurturing and exploding intelligent business 2.0
From today’s perspective, looking at the next decade, we return to the theme of today: the decade of nurturing and exploding intelligent business 2.0. Moreover, in the next three years, we may see the emergence of that native application driving the development of the entire ecosystem. Artificial intelligence technology will become a general technology, empowering more and more industries to complete their intelligent transformations. The key lies in whether machines can replace humans in decision-making, and the core capability behind this is the ability to establish decision models based on scenarios. Establishing a living, learning, and growing AI system: AI Agent. Intelligent business will become the mainstream business paradigm.
What capabilities will future enterprises require in this intelligent era? Conceptually, we are clear that we need to pursue intelligence and get this intelligence flywheel turning, with one side being user experience and the other being knowledge and data. Nowadays, the startup PPTs you see basically start to illustrate this flywheel; we aim to be the drivers of intelligence. However, the real difficulty lies in the fact that this is merely a concept. We are in the very early stages of this ecosystem, and we do not know how the future will unfold. Just like autonomous driving, I have followed this field for ten years, and the more I follow, the less I know how it will end. How the competition in this second stage will unfold is actually a highly complex system with too many uncertainties. Therefore, how do you embrace the future in the early stages of the intelligent ecosystem?
Intelligent strategy: look ten years ahead, think three years ahead, act in one year
This brings us back to our old expertise, discussing intelligent strategy. Compared to the discussion in 2017, there is a significant change: I specifically emphasize "thinking three years ahead." Thus, "look ten years ahead, think three years ahead, act in one year." Looking ten years ahead is about visioning, which, through today’s discussion, you should understand the value of such foresight. It is the premise of all your strategic decisions; you must strive to understand the various possible evolutions of the future. The second point, "thinking three years ahead," is about strategy, which is to start with the end in mind. Based on the vision, establish your positioning and development path, etc. Acting in one year is about planning, ensuring the execution of this plan. Two points need to be emphasized: one is that I repeatedly mention that strategy should be based on the rapid iteration and feedback of vision and action, constantly adjusting. You need to actively engage in various attempts to understand and test whether your imagination of the future is correct, and then adjust your vision of the future based on feedback.
The second very important point is that looking ten years ahead, thinking three years ahead, and acting in one year are not three separate tasks; they are three perspectives on the same issue. Whenever I encounter new inputs, I will ask what the short-term, medium-term, and long-term impacts are. Is it a one-year matter, a three-year matter, or a ten-year matter? Therefore, it is not about thinking ten years ahead when you think ten years ahead and acting in one year when you act in one year. You must always consider what this matter looks like in the short, medium, and long term; it is a matter of trade-offs. This is the essence of strategy; you need to train yourself to think from all three perspectives, which is very critical. A small suggestion for you is to actively use this framework: go back and seriously consider what your three-year goals are. Can you specify them to a measurable indicator? Not your traditional KPIs, but a truly reflective indicator of the essence of your innovative business.
Let me reiterate: what are your three-year goals? Can this goal be distilled into a very essential indicator? Most students' three-year figures are habitual, linearly derived, and rarely do they bring the tension of looking ten years ahead into the formulation of three-year goals. Then, based on these three-year goals, you can work backward to determine what you should do next year and this year. As we approach the annual strategic planning period, you can seriously reflect on whether your three-year goals are clear. If they are unclear, it represents a tremendous opportunity; uncertainty also represents a tremendous opportunity, indicating that we have significant room for growth, and you can create the future. If there is only a current linear derivation, it merely indicates that your growth potential is very limited.
Intelligent strategy: emergence and growth
The second deep insight is that intelligent strategy is about emergence and growth. It is no longer the result of a powerful CEO's decision; it is a dynamic balance of short-term, medium-term, and long-term interests. Why is it called intelligent? This strategy is intelligent because you actively embrace uncertainty and adapt to changes in the environment. Intelligent strategy is about emergence and growth, maintaining possibilities, and even creating possibilities, rather than merely pursuing efficiency. When there is no map, you must create a compass! Today, we do not have a detailed map telling us what the future will look like; we can only create our compass. This is the most significant difference between intelligent strategy and traditional strategy. The core of traditional strategy is to reduce uncertainty, with relatively certain planning and efficient execution. However, because we are in a highly complex and rapidly changing era, uncertainty represents possibilities and opportunities for creation. Therefore, the essence of strategy today is creation and innovation. In this sense, strategy is no longer just a high-level executive's responsibility; it is about innovation and creation, closely intertwined with products, technology, and user experience. All these elements must reflect your strategic principles, and they must function like intelligent agents, providing feedback on whether what you are doing is correct. This is the potential future of strategy we envision.
However, such a strategy requires a completely different organization to implement. Let me add that early strategies were the result of the wise decisions of CEOs. Over the past five or six years, everyone has realized that no company can succeed without three to five co-founders. You will quickly discover that even if you have a dozen outstanding executives, they may not be able to manage everything. You increasingly need the entire organization to be vision-driven, strategy-driven, and future-driven.
The future of intelligent organizations
Why do we need a different organization? What will intelligent organizations look like in the future? It is because the environment requires organizations to continuously emerge with good strategic decisions and innovations. The term "continuously" is crucial; in today's environment, making one correct decision is not enough because you are entering a continuous elimination race. Our current goal may be the Asian Games, but most students are likely still in the selection phase for the city games, with several steps ahead. Therefore, getting it right this round is of no use; it merely buys you a ticket to qualify for the next round of competition. Thus, you need to establish an organization capable of continuously producing high-quality innovations and decisions. In this sense, it aligns perfectly with the context of the AI era, where the importance of simple, replicable work is declining sharply. Efficient execution will still be crucial for the future of organizations, but increasingly, efficient execution will be completed by AI systems. The focus of organizations will increasingly evolve into creating unique value. Therefore, at the individual level, there will be a tremendous demand for creative talent in the future. The talent of the future will need multidimensional perspectives and unique expertise. Particularly with the emergence of AGI, the positions of narrowly defined professionals will likely be eliminated once again.
The industrial revolution marked a contraction from general capabilities to specialized narrow capabilities, as cooperation strengthened specialization. However, if we return to a new era, this starting point will lead us into a new era of building general capabilities, where the most creative individuals will be those with synesthetic abilities. Therefore, we need individuals with general capabilities who can also understand and leverage specialized knowledge. The sharpness, impact, and vitality of an organization lie in its ability to rapidly break through open-ended problems, which will be the core capability of future organizations. You must continuously break through and create.
The collaboration of creative individuals and machines will be the mainstream working state in the future, and I currently see some early forms of this. The best organizational form for creating new jobs in the future may be very tightly coupled special forces teams, consisting of eight to a dozen people. Many entrepreneurs have already noticed that a team of around ten people is sufficient. In the previous wave of mobile internet entrepreneurship, around thirty to fifty people were needed to sustain a business. However, in this wave of entrepreneurship, around ten people are indeed enough, including the well-known case of Midjourney, which achieved significant success with just a small team in a very short time. However, tightly coupled teams must have corresponding foundational capabilities and support, which is very important. Therefore, this relatively loosely coupled structure can quickly mobilize internal organizational capabilities, combined with a broad, open external network for collaboration.
The evolution of organizational guiding principles
Returning to our discussion of intelligent business, network collaboration remains the core concept, and the same applies internally within organizations. From the inside out, we must transform into a networked organization. The core of management in the industrial era was hierarchical systems, and we must break away from hierarchical structures and move towards a networked organizational form. This is what intelligent organizations require: new guiding principles. When the management revolution began 100 years ago, we emphasized management. We spent a century learning management, and many entrepreneurs today are still learning management. Basic management is certainly necessary, but it is just the foundation. In the era of knowledge revolution that Drucker spoke of, we entered the era of software engineers, where a good software engineer could be worth a thousand average engineers. Because when an engineer sits there for a day, you cannot tell whether they are working hard or slacking off.
Thus, the guiding principles of organizations have shifted from management to motivation. You cannot measure what kind of rewards to give based on output, so motivation has been prioritized, which is equity. Since the 1970s, the entire internet revolution has been accompanied by equity systems. Over the past decade, especially in the last five or six years, many unreasonable aspects of equity systems have been fully exposed, as more often than not, you are not in a motivational model.
The empowerment model, which I discussed in 2017, is increasingly necessary. It provides motivation and drive. More often than not, we have reached the second-highest level of Maslow's hierarchy, which is self-actualization and self-motivation. At this point, what they need is not motivation; such excellent talent can find work anywhere and is not short on money. They need empowerment; they need assistance. They need you to provide a platform that allows them to have greater space to excel. Therefore, empowerment will be a very important foundational capability for organizations for a long time.
However, in recent years, we have seen that co-creation has become a very important mechanism, even though it has not yet permeated all aspects of enterprises. At least in the areas of strategy formulation and execution that I have observed, the core employees of the enterprise must participate in strategic discussions together. Then, dynamically adjusting the strategy based on feedback and changes in the external environment allows the strategy to emerge. We need to create a consensus within the organization, forming a system that can continuously adjust. Co-creation, of course, requires certain prerequisites: you need the right people, and you need to share; without shared outcomes, no one will want to co-create with you. Therefore, every principle is built upon a solid foundation laid by the previous principle. This also reflects the evolution of organizational principles. We are transitioning from the IT internet era to the intelligent era, moving from motivation to empowerment and co-creation.
Changes in market capitalization over the past 20 years
At this point, everyone will certainly ask another question. Let me show you a chart to provide some stimulating, real insights. Market capitalization is a decent litmus test. You may have mixed feelings when looking at this chart today. There is so much information here, as we can observe the evolution of the most outstanding companies over the past 15 to 16 years, which can provide us with much motivation.
You can see that from 2007 to 2017, it was truly the internet era. Companies like Google, Facebook, Alibaba, and Tencent experienced growth of about 20 to 30 times, achieving significant increases from a high base. They transformed from billion-dollar companies into multi-hundred-billion-dollar companies. However, from 2017 to 2023, the previous companies only grew by about three times, while Nvidia and Tesla experienced growth of 20 to 30 times. This is because Nvidia and Tesla are the pioneers of the intelligent era, having already begun their 30-fold growth. Who will be on this list in 2033? It is particularly interesting to see continuity and the phase changes over the decade.
One conclusion is that, as with the last three slides, we return to the PPT from 2017. The conclusion is essentially the same: it takes ten years to achieve greatness. If you want to accomplish something significant, you must align with a great era. If you truly follow this trend, you cannot achieve anything meaningful without ten years. My experience is that for companies that achieve greatness over a decade, the most challenging part is from 0 to 0.1, not from 0.1 to 1. Because 0.1 indicates that your prototype has some semblance, but since what you are creating is so new, you do not even know what it should look like. Therefore, it often takes about three years to have a rough idea, and five years to explain to others what you are trying to create. I have not seen any exceptions; over the past decade, including my journey with the Lakeside community since 2014, truly great companies spend at least three years in the strategic exploration phase.
So what is needed? It requires original intention and persistence. People often ask me how to look at ten years. One is that you must strive to look and persist in looking. The second is why you would look ten years ahead. What lies behind it? Why are you willing to sacrifice short-term interests for long-term benefits? It is because you have a greater pursuit, a mission, a vision, and values. You want to transform the world and make it a better place, and you have something different to offer others. If you do not have such intentions, your vision will naturally be limited, and you will not be able to see far ahead or gain support from others. The right timing, location, and people come from following the right path and the great trends of the future. You can use the best and most advanced technology to solve the problems of this era. This is the true entrepreneurial spirit and the foundation for achieving greatness. All the successful companies we discussed earlier are "companies of the era." This phrase is indeed accurate. However, two fundamental points define a company of the era: you must genuinely keep pace with the great trends of this era, and you must possess the intentions and capabilities that align with the demands of this era.
This is the underlying driving force behind everything and the reason why some can go far. However, such companies have all faced particularly challenging developmental processes. Google once considered selling, and Tencent also contemplated selling; they all faced moments when they thought they might not make it. At such times, what can you rely on? You can only rely on the power of belief, believing that tomorrow will be better. Moreover, the most critical aspect during this process is that there will be a leap of faith based on belief.
Ultimately, whether it is vision or mission, do you have faith and belief in this matter? Believe in yourself; at that moment, that is your only reliance. Of course, I wish everyone good luck in the end!