Flat, Fast, and Intelligent: The Future of How Creative Teams Work
There's a debate that has quietly divided the business world for decades, and it's getting louder.
Do flat organizational structures actually work, or are they just a romantic idea that falls apart the moment a company hits a certain size? And if they do work, do they only work for small, scrappy startups doing creative work, or can they scale into something bigger and more durable?
It's a fair debate, considering that many mega companies tried and failed at it. Zappos went all-in on flat structure and saw significant attrition. Valve built an entirely manager-free studio that struggled to scale beyond a few hundred people. Spotify tried cross-functional squads and eventually drifted back toward conventional management as it grew.
Every time someone champions a flat structure, the critics point to these examples and say, "See, it doesn't work at scale."
But Jack Dorsey's recent essay, From Hierarchy to Intelligence, published with Sequoia's Roelof Botha, reframes the debate entirely, and we think he's onto something important.
The Real Problem With Hierarchy
To understand Dorsey's argument, it helps to understand where hierarchies actually came from in the first place because they weren't invented to consolidate power or slow things down. They were invented to solve a very specific problem.
Two thousand years ago, the Roman Army needed to coordinate thousands of soldiers across vast distances with almost no communication infrastructure. Their solution was a nested structure: eight soldiers per tent reporting to a decanus, who reported to a centurion, who reported upward from there.
It was, at its core, an information routing protocol. A human-powered system for making sure the right context reached the right person at the right time.
That logic carried forward through the Prussian military, into the American railroad industry, and eventually into every major corporation in the modern world.
Middle management wasn't originally about bureaucracy. It existed to route information, pre-compute decisions, and keep large organizations aligned across complexity.
The problem is that this system was designed for a world where humans were the only available mechanism for doing that routing. It was the best available solution to a coordination problem, not a fundamental truth about how organizations should work.
And if the coordination problem has a better solution now? The entire justification for traditional hierarchy starts to unravel.
Why Flat Works for Creative Work
Before getting to Dorsey's AI argument, it's worth making the case for flat structure on its own terms because the case is strong, particularly for teams doing creative, fast-moving work.
In a layered organization, failure is political. A bad decision doesn't just affect the project. It affects how you're perceived by the person who approved it and the person above them. So people stop optimizing for boldness and start optimizing for defensibility.
They choose the safe option, the proven approach, the thing that already worked somewhere else. The org chart becomes a risk-aversion machine.
Creativity, on the other hand, is a condition more than it is a trait. You create the right environment, in which trust, proximity to the problem, and psychological safety to be wrong prevail, and creative work follows naturally. Flat structures create those conditions architecturally, not just culturally.
This is exactly how things work at Blink22. We follow a flat structure, so when someone has an idea, they don't need to schedule a meeting to get permission to explore it. They explore it. They build something rough, bring it to the team, and the conversation happens around something real rather than a slide deck.
When a decision needs to be made, the person doing the work almost always makes it because they have the most context. There are no bottlenecks disguised as process. No six-week approval cycle for a two-day experiment. No ideas that die quietly in a chain of command before anyone with authority even hears them.
The result is a team that moves fast, takes real creative risks, and builds things that genuinely surprise people, including us, sometimes! Not because we hired unusually brilliant individuals, but because we built a structure that lets the intelligence already in the room actually surface and move.
The Scaling Problem and Dorsey's Answer
Here's where the critics have a point, though. Flat structure is relatively easy when a team is small. Everyone has context. Communication is natural. You don't need formal coordination because informal coordination fills the gap.
As organizations grow, that informal coordination breaks down. You get misalignment, duplicated work, decisions made without the right information, and people pulling in different directions without realizing it. The traditional answer to this problem has always been to add a management layer. Create reporting structures. Build the hierarchy.
Dorsey's answer is different: let AI handle the coordination that hierarchy was always doing.
His argument is that most companies using AI today are simply giving everyone a copilot, making the existing structure slightly faster without changing it. That's not the opportunity. The real opportunity is to replace what the hierarchy does at a structural level.
A system can now maintain a continuously updated picture of an entire business—what's being built, what's blocked, where resources are, and what's working—and use it to coordinate work in ways that previously required humans relaying context through layers of management.
In Dorsey's model, this compresses the entire organization down to three roles:
Individual contributors who build and operate close to the work
Directly responsible individuals who own specific problems and outcomes with full authority to act
Payer-coaches who combine doing real work with developing the people around them
The model has no permanent middle management layer. No role whose primary function is passing information from one level to another.
The AI carries the context, and people make the decisions.
Where We Agree and Where We'd Push Back
We find this genuinely compelling, and it aligns with the direction we're thinking at Blink22. AI absolutely has a role to play in making flat structures more scalable and more effective. Using it to surface the right information to the right person at the right time, without requiring a manager to relay it, is a real and valuable thing.
But we'd push back on one part of the vision: the idea that human judgment in coordination and management can be fully replaced.
Information routing is one thing. But what managers and team leads often do isn't just relay information. It's interpreting it.
They notice that a team member is struggling not because the data shows a productivity dip, but because of something in how they spoke in a meeting. They recognize when a technical problem is actually a motivation problem or when a deadline conflict is really a clarity problem. They give a perspective on why a product isn’t working the way it’s supposed to.
An AI world model can tell you what is happening. However, it's much harder for it to understand why and nearly impossible for it to navigate the deeply human dimensions of creative work.
Our view is that the right model isn't AI replacing human judgment in coordination; it's AI handling the mechanical parts of information flow so that the humans doing the coordination can focus entirely on the parts that actually require human judgment. Less time in status updates. More time in real conversations about real problems.
That's a version of Dorsey's vision we're very much on board with.
Our Verdict
For us, the debate about whether flat structures work isn't really about structure. It's about information. Who has it, how fast it moves, and what decisions it enables.
For two thousand years, hierarchy was the only answer to that problem. It doesn't have to be anymore. But the solution isn't to remove humans from the equation. It's to finally free them to do the parts of the work that only humans can do.
At Blink22, that's the company we're building. And we think it's the most interesting place to be right now.
