Learning from Block's 40% Layoff—You're Probably Only Learning How to Cut Staff

Deep thoughts on AI and aspiration —— ByteThink Circle

Block cut 40% of its workforce in late 2025, from concept to execution in under three weeks. Outsiders see the percentage; insiders see the algorithm: the team ran a three-step minimization exercise—how many people to keep services running at 100%, how many to meet strict regulatory compliance, how many to deliver on growth commitments plus agent reconstruction—add a buffer for error tolerance, and you have the final number.

Many CEOs will be tempted when they see this news. But companies that directly copy this move will likely only learn how to lay people off, not the conviction behind it. Where that conviction comes from is the truly valuable part of this story.

A Two-Thousand-Year-Old Hierarchy Problem

To understand what Jack Dorsey is dismantling, you first have to acknowledge why hierarchy exists. The modern corporate organizational model has its prototype in the Roman legion: 8-person squads nested into centuries, cohorts, and legions, with commanders at each level responsible for aggregation and communication. It was designed around a human limitation—a leader can effectively manage about 3 to 8 people, what we now call span of control. In the 1850s, railroad companies drew the first organizational charts; Taylor then refined them into functional pyramids. For the past two centuries, management theory has basically been making improvements within this framework.

Hierarchy is essentially an information routing system that coordinates large-scale human effort under conditions of limited information transmission. It has an unavoidable cost: narrowing span of control requires adding layers, and more layers slow information flow. Tech companies have made various escape attempts—Spotify promoted cross-functional squads, Zappos tried holacracy, Valve operated flat—and at scale, all returned to hierarchical coordination. It’s not that they didn’t try hard enough; it’s that alternative routing mechanisms didn’t exist at the time.

Dorsey’s judgment: AI is the first real alternative. A system that can maintain a continuously updated model of the entire business and use it to coordinate work—this used to require human managers relaying messages layer by layer.

Replacing Routing with Models Requires Two Foundations

Block’s approach isn’t complicated: instead of giving each person an AI assistant, let AI directly take on the coordination function that hierarchy performs. The company needs two world models—one about itself, one about customers.

The one about itself depends on a condition most companies lack. Block is remote-first; all decisions, discussions, code, design, and plans exist as digital records. AI can continuously maintain a panoramic view: what’s being built, what’s blocked, where resources are allocated, what works and what doesn’t. In traditional companies, this is a manager’s job—passing contextual information up and down the command chain. In an all-text company, the machine reads it itself.

The one about customers puts Block in a position few can replicate. Money is the most honest signal. People lie in surveys, ignore ads, abandon shopping carts, but every transaction, transfer, and payment actually happens. Through Cash App and Square, Block sees both ends of millions of daily transactions; every customer’s and merchant’s financial reality continuously accumulates in the system.

With two models, you can then talk about an intelligence layer. Payment, lending, card issuance—these capabilities are no longer products but building blocks. The intelligence layer combines blocks into solutions at the right moments. When a restaurant’s cash flow tightens before a seasonal decline, the system automatically assembles a short-term loan with an adjusted repayment plan and delivers it before the merchant thinks to look for money. Where does the roadmap come from? When the intelligence layer tries to combine a solution but fails because a capability doesn’t exist, that failure signal becomes what to build next. Customer reality directly generates the to-do list; no more product managers guessing from scratch.

People retreat to the edges, but the edges are where action happens. The entire organization converges into three roles:

From Dorsey to any employee, it’s now at most 5 layers deep. His goal is to compress it to 2 or 3.

Why You Can’t Copy It Usually Lies in the Foundation

My judgment: the replicability of Block’s approach is severely overestimated. It has two foundations, each filtering out most companies.

First, the work itself must be machine-readable. If your company’s core knowledge lives in veterans’ heads, at dinner tables, in twenty-minute voice messages, AI has nothing to read. The world model is an empty frame. Learning from Block and cutting middle management then means removing the router before the new network is installed. Second, customer signals must be honest. Most companies understand customers through surveys and satisfaction scores—low-fidelity channels. The intelligence layer, working with low-fidelity signals, can’t combine the right solutions, at best combining pleasing ones.

The narrative layer also needs calibration. The most widely circulated version says AI makes most decisions; even Block board member Roelof Botha doesn’t see it that way. His version is much more moderate: AI helps transmit alignment, the management team sets the framework, people at the edges handle error correction, information input, and directional adjustment. Radical claims suit propagation; hybrid reality suits operation. When learning from others’ organizational experiments, step one is distinguishing between these two—don’t mistake fundraising narrative for an operating manual.

One change is particularly practical though. Block’s meetings shifted in two months from flipping through slides to each person bringing AI-constructed prototypes in real time, using simulated or real data to slice a facet of their work, editable on the spot. What changed isn’t really meeting format but evidence quality: discussion centers on what’s actually being built, not what’s planned, suddenly expanding explorable breadth. Any size company can learn this tomorrow without buying a single layoff tool.

For a 100-Person Company, Two Things You Can Do Tomorrow

Dorsey gave small companies a starting point. First, ask yourself: is this management layer necessary, where does it obstruct actually solving customer problems? A 100-person company has at most two or three layers; now is the time to change while traveling light. Wait until thousands of people and the cost becomes an entirely different magnitude. Second, put the information generated daily—group chats, documents, meeting notes—into a conversational intelligent system. His exact words: this will increase your understanding of the company by two or three times, because previously you could only rely on people telling you, and for various reasons people don’t always speak up.

Finally, there’s the litmus test he left behind, which I think is more worth copying than the entire four-layer architecture: what genuinely difficult-to-understand things does your company comprehend, and is this understanding deepening every day? If the answer is no, AI is just a cost optimization story for you—cutting headcount, improving a few quarters of margins, then getting absorbed by something smarter. If the answer is profound, AI is revealing what your company actually is. This question has no uniform answer—a gas station and a law firm respond completely differently—but every company must answer it, and the sooner the better.