40% Workforce Reduction: The New 'Organizational OS' for the AI Agent Era as Demonstrated by Jack Dorsey's Block
Owen Jennings, Business Lead at Jack Dorsey's Block (formerly Square), spoke calmly yet revealed a shocking fact in an interview with a16z. Block recently carried out a massive reduction in force (RIF), cutting approximately 40% of its total workforce, but this was not merely for cost-cutting purposes.
Now that they have the 'power drill' that is AI, this was a bold, founder-led update to address the fact that the previous 'organizational structure' itself had become inefficient. This article explains the full scope of how Block was reorganized (Reorg) into an AI-centric organization.
1. The 'Binary Change' that occurred in December 2025
There have been several stages in the evolution of AI, but according to Jennings, a definitive paradigm shift occurred at the end of 2025.
1-1. The birth of intelligence that understands existing complex code
Until then, AI was suitable for writing code for new projects (green space), but it was insufficient for handling existing, massive, and complex codebases. However, the arrival of next-generation models like Claude Opus 4.6 and Codex 5.3 changed that situation entirely.
'A handful of engineers using these tools began to demonstrate 10x, 20x, and sometimes 100x the productivity of the past. It was clear that AI systems were far superior to hiring 1,000 humans to do the work.'
1-2. The collapse of the correlation between headcount and output
For the past few decades, the growth of tech companies has correlated with 'headcount.' However, with the rise of AI agents, this correlation has completely broken down. Block faced this reality head-on and spent the entire first quarter debating 'what an organization in the AI era should look like.' The result was the decision to cut more than 40% of the staff.
2. The true nature of the 'AI Agent OS' built by Block
Block did not just reduce its headcount; it had been preparing the infrastructure to incorporate a new workforce of 'AI agents' into the organization for several years.
2-1. 'Goose': A model-agnostic foundation
'Goose,' developed in-house by Block, is a model-agnostic agent harness.It orchestrates the entire development process by utilizing approximately 120 different models, such as Anthropic, OpenAI, or open-source models, depending on the task. Currently, much of the automation within Block is executed through Goose.
2-2. 'Builderbot': Automating development
Even more groundbreaking is an internal tool called 'Builderbot.'This tool autonomously merges GitHub pull requests (PRs) and completes 85-90%, and sometimes 100%, of feature development using only AI. 'The era of hand-writing code is over. Now, humans finish the final 10% of PRs created by AI agents after understanding the context. That is the current development style at Block.'
3. 'Generative UI': The changing form of apps
The transformation of the organization is also changing the form of the products users hold in their hands.
3-1. From static UI to dynamic UI
Apps until now have featured fixed UIs (static UI) defined in advance by engineers. However, 'Generative UI' is being adopted in 'Moneybot' (within Cash App) and 'Manager Bot' (within Square), which are being deployed by Block.
For example, if a user asks to 'visualize my recent spending,' the AI generates charts and graphs on the spot. This is not a UI written into the source code in advance, but one that the AI 'invents' each time. 'Your Cash App and my Cash App will look and function completely differently. The era where apps adapt to humans has arrived.'
3-2. Proactive Intelligence
What Block prioritizes is not having users type prompts. It is a 'proactive' experience where AI anticipates the user's context and suggests appropriate actions. This allows AI to act like a personal CFO in the complex domain of finance.
4. What is the 'True Moat' for Companies Surviving in the AI Era?
Now that every company can implement AI, what distinguishes competitive advantage? Mr. Jennings offers a very sharp insight into the future 'moat'.
4-1. Do You Understand 'Something Others Cannot'?
Simply being able to generate code is no longer a differentiator. 'The biggest moat boils down to how deeply a company understands something that is extremely difficult for other companies to understand.'
In Block's case, that is the vast data and deep insight into 'how buyers and sellers participate in the economy.' Feeding this unique 'signal' back into the AI and keeping the learning loop running—the company that can run this improvement cycle hundreds or thousands of times a day is the one that will ultimately survive.
4-2. Returning to a Lean Organization
Block's organization has been reorganized from its former hierarchical functional structure into small 'squads' of 1 to 6 people. Management layers have been reduced by more than 50%, and information flows freely. 'Small, lean, few layers, and a wide span of control. And just getting back to building. That is Block today.'
Summary: Will Your Organization Disappear Due to 'Vibe Coding'?
Owen Jennings' story is not a prophecy of the future, but a reality that is already happening.
The image of Block, which utilizes AI agents like '1,000 subordinates' and maximizes the speed of decision-making, serves as a mirror for all business professionals. 'If you cannot answer what your company deeply understands that others cannot imitate, your company might be wiped out by AI (Vibe Coding).'
We are now witnesses to, and participants in, the most intense 'organizational evolution' in history.
Recommended Articles