The traditional corporate hierarchy is an artifact of information scarcity. For a century, the pyramid structure served as a biological routing network, moving data up to decision-makers and pushing directives down to laborers. Jack Dorsey’s recent restructuring of Block—eliminating forty percent of the workforce—is not a conventional cost-cutting measure disguised as innovation. It is an architectural teardown. By replacing the traditional organizational chart with a circular model, Dorsey is attempting to build what he terms an "intelligence layer." In this configuration, artificial intelligence sits at the center, holding institutional context and operational memory, while human employees operate at the edges. This represents a fundamental inversion of the firm, shifting the burden of coordination from middle management to machine intelligence.
The Intelligence Layer
The concept of a centralized intelligence layer challenges the foundational logic of the modern corporation. Since the early twentieth century, companies have relied on the Taylorist model, which fragmented work into specialized tasks coordinated by an expanding class of middle managers. Information was the currency of the hierarchy. At Block, Dorsey is betting that large language models and autonomous agents can absorb this coordination tax. By simplifying the company into just three roles, the structure mirrors the architecture of a neural network rather than a traditional factory floor.
Roelof Botha of Sequoia Capital notes that this structural evolution is already standard practice among early-stage, AI-native startups. These new entities are born without the legacy debt of deep organizational charts. They scale output without scaling headcount. However, retrofitting an established publicly traded company like Block to operate like an agile, AI-native startup requires acting from a position of strength rather than reacting to market distress.
The math behind a forty percent workforce reduction forces a binary outcome: either the company collapses under the weight of missing personnel, or the remaining system becomes hyper-efficient. By placing AI at the core, Dorsey is gambling that the lost human coordination can be entirely automated. The center holds the context; the edges execute the novel, creative, and highly specific tasks that machines cannot yet manage.
Erasing the Management Pyramid
Transitioning from a pyramid to a circle fundamentally alters the physics of corporate power. In a pyramid, proximity to the top dictates influence, and middle managers act as gatekeepers of information. In a circular organization with an AI core, information is omnidirectional and universally accessible to the human nodes at the perimeter. This flattens the hierarchy to an extreme degree, effectively eliminating the need for human routers.
Botha points out that while the tools of execution are changing, the timeless qualities of a chief executive remain. The CEO must still define reality and allocate resources. Yet, the mechanism of that allocation is now radically different. When a company operates as a "mini-AGI," the friction of internal communication drops to near zero. Decisions that previously required weeks of committee alignment can be processed instantaneously by the central intelligence layer, which synthesizes cross-departmental data without political bias.
This structural shift mirrors the transition from mainframe computing to decentralized networks, but applied to human capital. The risk is profound. Eliminating the organizational chart removes the traditional scaffolding that catches errors and mentors junior talent. If the AI core hallucinates or lacks sufficient context, the human edge may lack the structural support to course-correct, turning a highly efficient circle into a highly efficient failure.
Dorsey’s experiment at Block is a leading indicator for the future of the enterprise. If the circular, AI-centric model succeeds, it will render the traditional corporate pyramid obsolete, triggering a mass extinction event for middle management across the knowledge economy. The lingering question is not whether AI can route information better than humans, but whether a company stripped of its human connective tissue can maintain a cohesive culture. The firm of the future may be infinitely more efficient, but it will be remarkably empty.
Source · The Frontier | AI
