When Management Detaches from Execution: The End of the Professional Manager Myth

By Geoffrey Chen

For a long time, modern enterprises have been described as objects that can be governed through “rational management.” Managers are portrayed as chess players positioned high above the board, coordinating organizations through models, reports, and abstract judgment, while execution is left to others. The legitimacy of the MBA system is rooted precisely in this imagination: as long as one masters sufficiently mature methodologies, management can exist independently of concrete action.

This narrative was not entirely wrong in the industrial era. When production processes were stable, technological change slow, and market demand relatively predictable, abstract management did indeed improve efficiency. Today, however, the problem is no longer whether this model is outdated, but whether it misunderstood the nature of the enterprise from the very beginning.

A company is not first and foremost an economic entity; it is a collective form of action. It is born from the willingness to confront uncertainty: someone decides to pursue something not yet proven feasible; someone bears the cost of failure; someone continuously adjusts direction under real-world constraints. Profit is not the reason a company exists, but the outcome of these actions over time. When outcomes are mistaken for causes—when a company is understood as a machine for producing economic value—management inevitably drifts toward abstraction and away from action itself.

The mythologizing of the professional manager is the institutionalized result of this misunderstanding. Over the past several decades, a repeatedly reinforced belief has taken hold: founders are suited for exploration and risk-taking, but once an enterprise enters maturity, it must be handed over to “professional managers.” In this narrative, professional managers are depicted as neutral, rational, and replicable figures. They need not deeply understand products or technology; mastery of processes, metrics, and organizational tools is presumed sufficient to “optimize” any enterprise.

The generation of professional managers represented by 王石 is a typical product of this historical logic. Their capability structure is not without value—on the contrary, they are highly skilled at handling the steering wheel, clutch, and brake. They excel at maintaining stability, controlling risk, and avoiding mistakes on well-defined roads. The problem is that when the road itself begins to change rapidly—when terrain becomes unclear and rules are constantly rewritten—this capability structure proves inadequate. What determines survival is no longer whether operations are standardized, but whether one possesses the instinctive capacity to respond directly to complex and shifting conditions. And such instinct cannot be added externally through training or frameworks; it can only grow naturally through long-term, real execution.

The core competence of professional managers has never been the creation of new value, but the allocation and preservation of existing value. Their effectiveness presupposes that problems are clearly defined, boundaries stable, and risks compressed. Yet the stages that truly determine an enterprise’s fate—periods of technological transition, product reconstruction, or unresolved business models—are precisely those in which problems remain undefined. In such contexts, abstract rationality is not merely unhelpful; it systematically excludes the risk-taking and judgment that are actually necessary.

This has already manifested in dramatic fashion in recent years. Two years ago, OpenAI experienced a sudden and intense leadership upheaval. The board attempted, within a very short period of time, to replace core executors on procedural and governance grounds, only to trigger strong collective resistance from research and engineering teams. The outcome itself revealed the underlying issue: in an organization operating at the technological frontier, where paths are still being formed, abstract governance structures lose legitimacy the moment they attempt to override execution reality. The crisis was resolved not because governance logic proved more “correct,” but because the organization was forced to acknowledge a fact—within domains of extreme uncertainty, the authority to judge direction can only belong to those deeply embedded in execution and technical reality.

This is not a new problem. Decades earlier, Apple paid a high price to learn the same lesson. When Steve Jobs was forced out of Apple in its early years, the company gained a more “professional” management structure, steadier processes, and a more respectable organizational form. What it gradually lost, however, was not efficiency but direction. Apple did not collapse immediately, but it drifted over time. Jobs’s return was not merely a return of managerial style, but a return of execution logic: managerial authority was once again bound to product judgment, technical detail, and responsibility for action. Apple’s later revival did not come from better abstraction, but from management returning to its proper place—making things actually happen.

When management detaches from execution, enterprises often enter a state that is outwardly rational yet internally hollow. Financials remain healthy, processes highly standardized, and decisions seemingly logical. Yet within the organization, few genuinely believe that what they are doing is still worth doing. Corporate spirit is not a slogan or a vision statement, but something more fragile and more real: it is the internal reason people are willing to go the extra mile to do things well, and the respect afforded to judgments and exploratory efforts that have not yet been quantified. Once this layer is entirely replaced by KPIs, models, and procedures, a company may not fail immediately—but it has already lost its internal momentum.

The emergence of artificial intelligence has further illuminated this problem. In discussions about AI, people often worry that symbolic reasoning will replace understanding and meaning. But at a deeper level, AI has not betrayed rationality; it has pushed rationality to a limit humans are reluctant to confront. When a system can reliably produce effective outcomes purely through symbolic relationships and statistical structure, what it reveals is not the disappearance of understanding, but the extent to which human “understanding” itself is largely an internal effect of symbolic structures.

The true dividing line lies in constraint. AI’s symbolic reasoning is constantly corrected by real-world feedback; errors translate directly into performance degradation, leaving no room for explanation. By contrast, the managerial language cultivated within MBA systems often circulates without such constraints. Reports are treated as reality, narratives replace correction, and formal rationality gradually floats free from practice. In an era where symbolic capability is systematically surpassed by machines, this kind of abstraction appears increasingly hollow and fragile.

AI can optimize paths, but it cannot answer a fundamental question: why a path is worth taking in the first place. The true value of an enterprise is never designed in advance; it is validated repeatedly through interaction with reality. It arises from an internal logic: why this work solves real problems, why long-term risk is worth bearing, and why continued effort remains justified even after failure. These are existential questions, not optimization problems, and they cannot be outsourced to MBA models, processes, or algorithms.

This is not a rejection of management itself. What is truly coming to an end is the illusion that management can exist independently of action, risk, and real-world feedback. In the age of AI, the only irreplaceable asset of an enterprise is not its models, processes, or managerial rhetoric, but the internal logic that has been proven—again and again—worth acting upon in reality. Only those who stand within execution are qualified to judge whether that logic still holds.

When management returns to this foundation, it ceases to be a ladder for escaping the front lines and becomes a higher-density form of responsibility. The enterprise, in turn, ceases to be merely a container of economic value and re-emerges as a human structure grounded in action and calibrated by reality.

《当管理脱离执行：职业经理人神话的终结》

By Geoffrey Chen

在很长一段时间里，现代企业被描述为一种可以被“理性管理”的对象。管理者被塑造成高处的棋手，通过模型、报表与抽象判断，协调一个由他人执行的系统。MBA 体系正是在这种想象中获得了它的合法性：只要掌握足够成熟的方法论，管理就可以脱离具体行动而独立存在。

这种叙事在工业时代并非完全错误。当生产流程稳定、技术变化缓慢、市场需求相对可预测时，抽象管理确实能够提升效率。但今天，问题已经不在于这种模式是否过时，而在于它是否从一开始就误解了企业存在的本质。

企业首先并不是一个经济体，而是一种集体行动的形式。它诞生于对不确定性的承担：有人决定去做一件尚未被证明可行的事，有人承担失败的代价，有人在现实约束中不断修正方向。利润不是企业存在的原因，而是这些行动在时间维度上的结果。当结果被误认为原因，当企业被理解为一台“制造经济价值的机器”，管理便开始不可避免地走向抽象化，远离行动本身。

职业经理人被神化，正是这一误解的制度化结果。在过去几十年里，一个被反复强化的观念是：创始人适合探索和冒险，而一旦企业进入成熟阶段，就必须交由“专业经理人”接管。在这一叙事中，职业经理人被描绘为一种中立、理性、可复制的角色，他们不需要深度理解产品或技术，只需掌握流程、指标和组织方法，便可以“优化”任何企业。

以 王石 为代表的这一代职业经理人，正是这种时代逻辑下的典型产物。他们的能力结构并非毫无价值。恰恰相反，他们极其熟练地掌握方向盘、离合器和刹车踏板，擅长在既定道路上维持稳定、控制风险、避免失误。但问题在于，当道路本身开始快速变化，当地形不再清晰、规则不断被重写时，这种能力结构便显得力不从心。真正决定生死的，已不再是操作是否规范，而是是否具备在复杂路况中做出即时判断的本能反应。而这种反应能力，并不是通过培训或模型外加上去的，它只能在长期、真实的执行过程中自然生长出来。

职业经理人的核心能力，从来不是创造新价值，而是分配和维持既有价值。他们依赖的前提是问题已经被定义清楚，边界已经稳定，风险已经被压缩。但在真正决定企业命运的阶段——技术跃迁期、产品重构期、商业模式尚未成型之时——问题恰恰是未定义的。在这种情境中，抽象理性不仅无助，反而会系统性地排斥必要的冒险与判断。

这一点，在近年的现实中已经以极为戏剧化的方式显现出来。两年前，OpenAI 爆发了突如其来的管理层动荡，董事会在极短时间内试图以程序性、治理性理由替换核心执行者，却迅速引发内部研究团队与工程团队的集体反弹。事件的走向本身已经说明问题：在一个尚处于技术前沿、路径仍在生成中的组织里，抽象治理结构一旦试图凌驾于执行现实之上，就会立刻失去合法性。危机最终得以平息，并非因为治理逻辑更“正确”，而是因为组织不得不承认一个事实——在这种不确定性密度极高的领域，真正能够承担方向判断的，必然是深度嵌入执行与技术现实之中的人。

这并非新问题。早在数十年前，苹果公司就经历过一次代价高昂的验证。当 Steve Jobs 早期被迫离开苹果时，公司获得了更“职业化”的管理结构、更稳健的流程和更体面的组织形态，但它逐渐失去的，并不是效率，而是方向感。苹果没有立刻衰败，却在长期中陷入漂移。乔布斯的回归，并不只是管理风格的回归，而是执行逻辑的回归：管理权威重新与产品判断、技术细节和行动责任绑定在一起。苹果后来的复兴，并不是因为抽象管理做得更好，而是因为管理重新回到了“让事情发生”的位置上。

当管理脱离执行时，企业往往会进入一种表面理性却内在空心化的状态。报表仍然健康，流程高度规范，决策看起来合乎逻辑，但组织内部已经无人真正相信这件事“值得继续做”。企业精神并不是文化口号或愿景陈述，而是一种更脆弱、也更真实的东西：它是组织成员愿意为把事情做好而额外付出的内在理由，是对尚未被量化的判断和探索空间的尊重。一旦这一层被 KPI、模型和流程完全替代，企业可能不会立刻失败，但它已经失去了继续向前的内在动力。

人工智能的出现，使这一问题进一步被照亮。围绕 AI 的讨论中，人们常常担忧符号推演是否会取代理解与意义。但从更深的层面看，AI 并没有背叛理性，而是把理性推进到了人类不愿面对的极限。当一个系统仅凭符号关系与统计结构，就能稳定地产生有效结果时，它揭示的并不是理解的消失，而是人类“理解感”本身在很大程度上也是符号结构的内在效应。

真正的分水岭在于约束。AI 的符号推演始终被现实反馈严格校正，错误会直接转化为性能下降，没有解释空间。而 MBA 体系下的管理语言，却常常在脱离现实约束的情况下自我循环。报表被当成现实，叙事替代了修正，形式理性逐渐漂浮于实践之上。在一个符号能力被机器系统性超越的时代，这种抽象理性反而显得空洞而脆弱。

AI 可以优化路径，却无法替企业回答一个根本问题：为什么这条路径值得走。企业的真正价值，从来不是被设计出来的，而是在现实中被反复验证出来的。它源于一套内在逻辑：做这件事为何能解决真实问题，为何愿意承担长期风险，为何失败之后仍然值得继续尝试。这些都是存在论问题，而不是优化问题，无法通过 MBA 模型外包给流程或算法。

这并不是对管理本身的否定。真正被终结的，并不是管理，而是那种试图脱离行动、风险与现实反馈而独立存在的管理幻觉。在 AI 时代，企业唯一不可被替代的东西，不是模型、流程或管理话术，而是那套在真实世界中一次次被证明“值得继续行动”的内在逻辑。只有真正站在执行之中的人，才有资格判断，这种逻辑是否仍然成立。

当管理回到这一点上，它不再是逃离一线的晋升阶梯，而是一种更高密度的责任形式。企业也不再只是经济价值的容器，而重新成为一种以行动为核心、以现实为校准的人类创造结构。

Email: [email protected]

GPG Public Key: https://keys.openpgp.org/[email protected]

(2025.12.31 Australia)