As a representative of China's emerging enterprises, Li Auto's corporate management ideology development trend has attracted the attention of the society. Recently, the media disclosed founder Li Xiang's comments on the recent changes in his management thinking:

The "founding company management method" mentioned by Li Xiang is exactly the "founder model" defined by Paul Graham, and the core logic of the two is highly compatible.

》（Founder Mode）

（https://paulgraham.com/foundermode.html）

Paul Graham (founder of Y Combinator , often referred to as the "Godfather of Silicon Valley") published a blog post titled "Founder Mode" on his 90s-style personal website in September 2024 (https://paulgraham.com/foundermode.html), which has caused a huge response in the scientific and technological community and the management community.

The "founder model" proposed by Paul Graham is a management method opposed to the "manager model" : it opposes the corporate governance logic of "hiring talents and letting go of management" , advocating that founders are deeply involved in the core business, breaking the hierarchical structure (such as "cross-level dialogue" instead of step-by-step reporting), focusing on core values rather than process delivery, and driving decision-making in an agile and direct way.

Li Xiang's statement fully echoes Graham's logic: he emphasizes "more dialogue, in-depth dialogue rather than reporting" (corresponding to Graham's "breaking down hierarchical communication"), "focusing on user value rather than delivery" (corresponding to "focusing on core value"), "rejecting the rigidity of the professional manager system" (corresponding to "opposing the manager model"), and at the same time using Nvidia and Tesla as a reference, advocating a return to the entrepreneurial management that the founder is good at.

Graham's article was originally intended to reflect on the growth model of Silicon Valley startups from the perspective of venture capitalists, not to demonstrate the general management phenomenon of large enterprises.

This article was directly inspired by a talk by Airbnb co-founder Brian Chesky at a Y Combinator event. Chesky shared that as he scaled his company, he followed Silicon Valley's standard advice – "hire good professional managers, give them autonomy, and then run the company through those managers himself." As a result, this approach almost ruined Airbnb.

（Steve Jobs）

Chesky later decided not to follow these recommendations and instead learned from Steve Jobs' management style – deep in detail, leapfrog management.

（Micro-management）

Graham realized that countless entrepreneurs actually felt this way: when the company grew, everyone told them that they had to transform into a "professional manager" model, and if they continued to get involved in the details, they would be considered "micro-management" or not know how to delegate power.

（Gaslit）

Graham argues that these founders were actually manipulated by the "gaslit effect." This so-called standard management advice (i.e., the "manager model") is not suitable for founders. Therefore, he proposed the concept of the "founder model" to justify founders who want to expand their companies in a unique way.

The gaslighting effect is a proper term in psychology, referring to a systematic form of psychological manipulation, which is basically equivalent to what we call PUA. By constantly denying the facts and distorting events, the manipulator makes the victim begin to doubt his memory, perception, judgment and reason, and eventually, the victim will believe that something is wrong with him ("Am I really crazy?"). "Am I mistaken?" Thus, completely dependent on and trusted the perpetrator's description of reality. The term comes from a 1938 stage play "Gaslighting" (also called "Angel Street"). The 1944 film of the same name starred Ingrid Bergman, for which she won the Academy Award for Best Actress.

In the movie, in order to obtain his wife's property, the husband systematically manipulates his wife psychologically, making her believe that he is slowly going crazy. The husband will secretly dim the brightness of the gas wall lamp in the house without his wife's knowledge. When the wife notices that the light has dimmed and asks her husband questions, the husband will firmly deny it, insisting that the brightness of the gaslamp is completely normal, and in turn accusing the wife: "You are thinking wildly again." ”

（Manager Mode）

（Founder Mode）

Graham divides management methods into Manager Mode and Founder Mode.

（Fakers）

Manager Mode is the traditional model taught in business schools. The principle is "hire good people, let them do things, and don't get involved unless things get messed up." The company's organizational structure is modular, and the CEO can only be managed by direct subordinates (executives) and cannot skip levels. Graham's criticism of this model is designed for professional managers, who often don't understand the technical details or product intuition, so they can only manage the company by managing "people". But if the founders do the same, the company becomes mediocre, bureaucratic, and easily taken down by fakers.

（Skip-level meetings）

The core logic of Founder Mode is that founders use their deep understanding and intuition of the company to break down hierarchies and directly intervene in important details. The company does not have a fixed organizational chart, but is more like a neural network with the founder as the node. This model behaves in the form of skip-level meetings, direct focus on product details, and even the freedom to switch between so-called "micromanagement" and "macro strategy".

Graham is not saying that professional managers are useless, but that forcing founders to take on the role of professional managers when founders are still in office is wrong, which wastes the core talent of founders.

I recommended Bain & Company's "Founder's Spirit" book in last year's article "China Management Model |

In fact, Graham's last sentence is:

I have another less optimistic prediction: once the concept of "founder model" is established, people will start abusing it. Founders who are reluctant to delegate power even in the matter of authorization will use the "founder model" as an excuse. Managers who are not founders may think they should try to imitate the founders. This may work to some extent, but if it fails, the result will be chaotic; And modular management at least limits the damage caused by a bad CEO.

（Top 100 retreat）

Graham clearly pointed out that Jobs was an example of the "founder model". It was by studying Jobs' method that Brian Chesky came out of the quagmire of the "manager model". Jobs was known for digging into every pixel detail and holding the annual Top 100 retreat.

Although Graham's article did not directly name Musk and Huang Renxun, public opinion generally believes that these two are also representatives of the "founder model":

Musk was once known for sleeping on the factory floor and personally reviewing lines of code, which fits perfectly with the definition of "breaking down the hierarchy and going into the details". Huang directly manages 40~50 direct subordinates (far more than 7~10 people in traditional management), and publicly expresses that he does not believe in traditional 1:1 meetings and likes to discuss problems in public, which is a typical practice of the "founder model" to break down hierarchical barriers.

The development history of Li Auto and Li Xiang's own management style are typical embodiments of the founder's model:

Extreme product obsession: Li Xiang is known as a "super product manager" who is known for his extreme sensitivity to product definition and "micromanagement" of details. He has personally formulated the product definition of Ideal ONE, L7/8/9 , and even paid attention to details such as the stitching of the seats and the pixel arrangement of the screen. This practice of directly grasping products across the hierarchy is a typical manifestation of the "founder model".

Flat and centralized organizational structure: The professional manager model tends to have a hierarchical reporting line, and Li Xiang has been adjusting the organizational structure to maintain efficiency. For example, he has implemented a matrix organizational structure and recently emphasized a return to entrepreneurial state, with the aim of breaking down the "big company disease" and departmental walls, so that founders can directly access the core of the business.

Breaking the superstition of management professions: Graham mentioned that professional managers are easily deceived by executives who are "not tech-savvy but good at upward management". Li Xiang emphasizes seeking truth from facts and cognitive iteration, and he prefers to look directly at the data and listen directly to the sound of front-line artillery fire, rather than just listening to carefully modified PPT reports.

Li Auto tried to introduce a standardized management process represented by Huawei's best practices and leaned towards the manager model, but after encountering market challenges (such as the turmoil in the early days of MEGA models), Li Xiang's response was to strengthen the "founder model":

Admit mistakes and correct them quickly: Professional managers tend to cover up mistakes to keep their positions, while founders (like Li Xiang) dare to openly admit strategic misjudgments (such as admitting MEGA rhythm issues) and adjust quickly because of their ownership and long-term vision for the company.

Acquisition and focus: Li Xiang often shrinks the front at critical moments, personally grasping the core business (such as sales, product definition) and reducing dependence on the middle layer. This kind of efficient decision-making in a "wartime state" is difficult for professional manager teams to have.

For China's fiercely competitive new energy vehicle market, this founder model with extremely fast response speed and extremely short decision-making chain may be the only solution to survive and win, and it is also a corporate governance model suitable for Li Auto, but it does not mean that there is no risk. For a company like Li Auto with an annual revenue of 100 billion yuan, the founder model faces the following challenges:

Founder's energy bottleneck: Musk and Huang can do it because they are superhuman energy monsters. Li Xiang If he had to do things for a long time, how to avoid energy overdraft was a problem.

Suppression and risk buffering of successors: The founder model relies heavily on the intuition and judgment of the founder's individual. This may lead to the decision-making ability of middle managers not being exercised, and once the founder makes a mistake (such as a mistake in judging the market), the company lacks a buffer layer for error correction.

In my two articles, "China's Management Model (1): Institutional Design" and "Ideals Give Up Learning from Huawei | The Pot of Management, Management Does Not Carry It", I talked about how professional managers have created modern Western capitalism from the macro perspectives of history, society, and economics.

In fact, the shift from "founder dictatorship" (such as Rockefeller, Carnegie, Ford) to "professional manager system" (such as General Motors' Alfred Sloan) is an inevitable result of the evolution of capitalism in the industrial age. Western societies (especially the United States) established the mainstream position of the professional manager system in the 20th century for the following four core reasons:

1. Visible hands solve organizational complexity

（Alfred Chandler）

Business historian Alfred Chandler discusses this in detail in his magnum opus The Visible Hand.

The limits of the founder's model: At the end of the 19th century, when the enterprise was just a single factory, Rockefeller had the final say alone. But with the rollout of the railway network, the company has become a behemoth that spans continents, has hundreds of thousands of employees, and involves raw material procurement, production, logistics, sales, and finance.

Bottleneck in human brainpower: A person's brain (even a genius) cannot process such a large amount of information at the same time.

Solution to complexity: The professional manager system is actually an algorithm for processing information in layers. Through the hierarchical system, complex decisions are dismantled, and middle managers are allowed to handle details, and high-level management is dealt with strategies. In the days without computers, it was the only "operating system" that could run very large organizations.

Henry Ford (the extreme of the founder's model) had insisted on refusing to update the Model T and firing all dissenting executives, nearly bankruptcy for Ford. And Alfred Sloan (the founder of the professional manager model) allowed General Motors to completely defeat Ford in terms of management efficiency by establishing a division system. This battle established the dominance of the manager system in the 20th century.

2. Capital will: the separation of ownership and management rights

As the business expanded, the founder's family did not have enough money and had to bring in external investors (stock market).

Wall Street demand: What external shareholders (shareholders, institutions) hate most is "unpredictability". Founders are often paranoid, risk-takers, emotionally unstable, and rely on intuition (risky) to make decisions. Professional managers are trained, rational, and "machine parts" (risk controllable) that value quarterly financial reports.

Fiduciary Responsibility: The professional manager system establishes a mechanism for supervising the CEO through the board of directors. For capital, hiring a replaceable CEO is far safer than being subject to an uncontrollable founder.

3. Biological Curse: The Heir Conundrum

This is the most insurmountable weakness of the "founder model": genes do not guarantee talent.

No more than three generations rich: It is impossible for the Rockefeller or Morgan family to guarantee that every generation of descendants will have the business genius of the founder. If family management (a continuation of the founder model) is adhered to, the business often declines due to the incompetence of the heirs.

Talent supply chain: The professional manager system (in conjunction with the MBA education system) actually creates a standardized "talent supply chain". Institutions such as Harvard Business School mass-produce "qualified managers", and companies can replace CEOs like replacement parts, so that the life of the business exceeds that of human life.

4. Management science prevails

In the early 20th century, Taylorism (scientific management) prevailed. The prevailing belief at the time was that management was a science, not an art. It is believed that the optimal solution can be calculated through processes, KPIs, financial statements, and organizational charts. The "intuition", "inspiration" and "personal charm" on which the founder model relies are considered primitive, unscientific, and even feudal.

Paul Graham's "founder model" is because the environment of the times has changed compared to a hundred years ago:

Digital transformation changes organizations: In the industrial age, managing 10,000 people requires complex hierarchies; But in the software age, technology leverage is extremely high. For example, when Instagram was acquired, there were only 13 people, and OpenAI changed the world with a very small core team. Technology makes it possible for founders to control the details directly, eliminating the need for so many middle layers.

Innovation> Efficiency: The main task of large enterprises in the 20th century (such as Coca-Cola and GE) is to improve efficiency and reduce costs, which is suitable for manager management. However, the 21st century technology war (such as AI and electric vehicles) is a battle that requires disruptive innovation. Professional managers are good at risk aversion, while founders are good at taking risks.  In the era of radical innovation, the conservatism of professional managers has become the biggest risk, such as the decline of Nokia.

The establishment of the professional manager system in the West is to solve the stability problem of super-large-scale organizations under low information technology means, which is the "optimal solution" in the era of industrialization. In the era of digitalization and artificial intelligence, the return of the founder model of Musk, Huang Renxun and others is because information technology has given individuals a stronger radius of control, and the market environment has changed from the pursuit of "stability" to the pursuit of "subversion".

The conclusion is that I hope that Li Auto can make good use of Feishu, SAP ERP, and O9 supply chain management information systems, and only sufficient digital transformation can support Li Xiang's return to comprehensive management.