From Hierarchy to Intelligence — Sequoia Capital
URL: https://sequoiacap.com/article/from-hierarchy-to-intelligence/
Author: Jack Dorsey and Roelof Botha (Sequoia)
Published: March 31, 2026
Summary
Jack Dorsey and Roelof Botha argue that AI makes it possible to replace organizational hierarchy itself, not just augment it. Block (formerly Square) is redesigning coordination around a continuously-updated real-time model of the company, using AI to perform what middle management has always done: route information, pre-compute decisions, and maintain alignment.
Key Claims
- Hierarchy is an information routing protocol. The Romans invented nested span-of-control (8→80→480→5000) because humans can effectively manage 3-8 people. This constraint never changed. Every org chart since is a variation on this solution.
- Middle management = pre-AI information routing. Scharnhorst’s General Staff, Frederick Taylor’s Scientific Management, McCallum’s org chart — all are attempts to route information efficiently under human cognitive limits.
- Previous alternatives failed to scale: Spotify squads reverted to conventional management. Zappos Holacracy saw attrition. Valve’s flat structure breaks above a few hundred people. The underlying problem (information routing) was never solved.
- AI changes the underlying constraint: For the first time, a system can maintain a continuously updated model of the entire company and coordinate across it. The bottleneck was always human cognition; AI removes that bottleneck.
- Block’s approach: Not “AI copilot for existing hierarchy” but AI as the coordination mechanism itself. Building the company as an intelligence.
- The frame: “Most companies are giving everyone a copilot. We’re building a company as an intelligence.”
- Historical arc: Military → railroads → corporations, all solving the same problem with the same hierarchical logic. AI is the first thing that could actually break this.
Implications
This is a macro-level framing for what AI agents mean organizationally. The “AI as coordination infrastructure” frame is relevant to how companies like Kaon AI might position themselves — building intelligence infrastructure that replaces not just individual tasks, but coordination layers.