Jack Dorsey suggests that artificial intelligence ought to take the place of middle managers following Block’s reduction of 4,000 positions.

Dorsey’s strategy eliminates middle management, with AI managing coordination, product choices, and internal synchronization.

What to know:

• Jack Dorsey maintains that his company’s choice to eliminate around 4,000 positions from its workforce of over 10,000 was not merely a cost-cutting measure but a fundamental restructuring aimed at substituting middle managers with AI.
• Dorsey previously indicated that the restructuring was prompted by a shift in capabilities he observed in December in tools such as Anthropic’s Opus 4.6 and OpenAI’s Codex 5.3, which he noted had become proficient in managing extensive codebases.
• The corporate hierarchy has historically existed to address one issue: the flow of information through organizations too large for any single individual to manage, a challenge that AI is now tackling, according to Dorsey.

In Jack Dorsey’s perspective, the role most vulnerable to the AI transformation is that of the middle manager.

Dorsey argues in a recent essay, "From Hierarchy to Intelligence," co-authored with Roelof Botha, managing partner at Sequoia Capital, an investor in Block, that the company’s decision to reduce its workforce by approximately 4,000 employees was not a cost-saving initiative but a lasting restructuring to replace middle management with AI.

The essay contends that corporate hierarchy has always been established to solve one fundamental problem: managing the flow of information in organizations too extensive for any single person to oversee.

Managers consolidate context from lower levels, act as intermediaries from upper levels, and ensure alignment across teams. The authors argue that AI is now capable of executing these tasks continuously and at scale, rendering the messenger role unnecessary.

In lieu of management tiers, Dorsey and Botha propose two AI-driven "world models."

One model consolidates internal data from code, decisions, workflows, and performance metrics to generate a constantly updated overview of company operations, thereby replacing the context traditionally held by managers.

The other model analyzes customer and merchant behavior using transaction data from Cash App and Square.

These models contribute to what Block refers to as an “intelligence layer” that dynamically constructs financial products to meet market needs.

If executed correctly, these models will take on the coordination responsibilities that previously justified the presence of middle management.

Rather than relying on fixed roadmaps, the essay suggests fragmenting Block’s business into modular capabilities, including payments, lending, card issuance, and payroll.

When the system identifies a requirement—such as a merchant experiencing a seasonal cash flow shortage—it assembles a solution from existing capabilities. If a capability is lacking, it defines what needs to be developed next, thus replacing the product roadmap with a system-generated backlog.

The organizational structure is adjusted accordingly. Block aims to function with three roles: individual contributors who construct the system, directly accountable individuals who manage specific outcomes on 90-day cycles, and player-coaches who remain engaged while mentoring others.

Dorsey informed Wired in early March that the restructuring was initiated by a capability shift he witnessed in December involving tools like Anthropic’s Opus 4.6 and OpenAI’s Codex 5.3, which he claimed had gained the ability to operate effectively within extensive codebases.

However, current and former Block employees informed the Guardian that around 95% of AI-generated code alterations still necessitate human adjustments, and that AI tools are not yet capable of leading in regulated sectors such as banking and money transfers.