Jack Dorsey’s Bold New Vision: Replacing Corporate Hierarchies With AI-Powered Intelligence Systems / 杰克·多尔西的大胆构想：用AI智能系统取代企业层级架构

We’ve all felt the frustration of corporate bureaucracy: waiting three layers of approval for a simple decision, sitting through endless alignment meetings that could have been an email, watching good ideas get stuck in the structural limbo between teams. For 2,000 years, this hierarchical model was the only way we knew how to coordinate large groups of people — from the Roman legions to 20th century Fortune 500 companies, we built organizations around the simple limitation that a single leader can only effectively manage 3 to 8 people at a time. But what if that limitation no longer applies?

最近读了Jack Dorsey和红杉团队联合发布的这份组织创新思考，最让我震动的是他们直接戳破了所有组织创新实验的本质困境：过去几十年不管是Spotify的小队模式、Zappos的合弄制还是Valve的扁平化架构，最终都会在规模扩大后退回传统层级，核心原因是我们一直没有找到能替代人类管理者承担信息流转和协调功能的工具。而现在，AI可能就是那个打破2000年组织逻辑的变量。

Dorsey’s team at Block is not just using AI as a productivity copilot for existing teams — they’re rebuilding the entire company around two interconnected AI models, eliminating the need for most middle management entirely. First, there’s the company world model: a continuously updated, real-time map of every part of the business, from ongoing projects and resource allocations to bottlenecks and progress updates, pulling all its data from the recorded artifacts of their remote-first work culture. This replaces the entire job of managers relaying status updates up and down the chain of command. Second, there’s the customer world model, built from the honest, unfiltered transaction data from both Cash App users and Square merchants, giving the system an incredibly clear, per-person view of actual financial needs rather than survey responses or hypothetical feedback.

最有意思的是他们重新设计的三个核心角色，完全颠覆了我们对传统公司岗位的认知：不再有层层上报的经理，只有深耕技术和产品能力的个人贡献者、对特定客户问题全权负责的直接责任人，以及专注于技能指导和人才发展的“球员教练”。AI系统承担了信息同步、优先级对齐、跨团队协调这些过去消耗管理者80%精力的工作，让所有人都能把时间花在真正靠近用户、靠近核心创造的工作上。甚至连产品 roadmap 都不需要产品经理凭空规划了：当AI尝试为用户组合解决方案却发现缺少某项能力时，这个需求会自动成为团队的下一步工作方向。

Of course, this is still an experiment in progress, and there are obvious questions to ask: How do you avoid bias baked into the AI models? Who takes responsibility when the system makes a bad call? What happens to employees who thrived in the old hierarchical system? But the core insight here feels unavoidable: if your company’s only use for AI is to cut headcount and optimize existing processes, you’re missing the biggest opportunity of the technology. The real value of AI isn’t to make your old hierarchy work slightly better — it’s to build a completely new kind of organization that’s faster, more responsive to customers, and lets people do the work they’re actually good at.

其实最值得我们思考的问题是：你的公司真正拥有的、别人难以复制的深层认知是什么？如果没有这个核心认知，AI最多只是帮你降本增效的工具，迟早会被更聪明的对手淘汰。而如果你的认知足够深，AI不会只是你的辅助，它会彻底把你的核心优势放大到过去不敢想象的程度。这种“围绕智能而非层级构建公司”的模式，或许真的会成为未来10年所有组织的标配。

来源：https://sequoiacap.com/article/from-hierarchy-to-intelligence/