The Edge Assumes a Self
The Block essay that Jack Dorsey and Roelof Botha published last week is worth engaging with. The historical arc and the argument that AI can now perform the information-routing function that hierarchy has always performed are real.
There’s a word that appears nowhere in the essay: development. As in human development. As in, the question of whether the people being asked to operate at the edge of a de-layered, AI-coordinated organization are actually capable of doing that.
Hierarchy did two things: the Block piece collapses into one. It routed information — who knows what, who decides what, how context travels up and down the chain. And it routed meaning. It told people what mattered when nothing was clear. It held people inside a legible world.
AI can plausibly replace the first function. That’s the Block thesis. But the second — the meaning-making scaffolding that hierarchy provided — is not a technical problem. It’s a developmental one.
Robert Kegan spent decades mapping how adults make meaning. In The Evolving Self and later In Over Our Heads, he argued that adult development doesn’t stop at adulthood — that the complexity of mind we bring to our experience continues to evolve, or can, under the right conditions. His research found something uncomfortable: most knowledge workers operate from what he called a socialized mind — a self largely constituted by the expectations of the systems in which it’s embedded. Identity, values, judgment: externally derived. The hierarchy told you what to do. The loop closed. A socialized mind could navigate that world well enough.
What Kegan called the self-authoring mind is different. It holds its own compass. It can act without waiting for direction — not because it’s reckless but because it has something stable inside to act from.
Jennifer Garvey Berger, building on Kegan’s framework in Changing on the Job, made a point that’s directly relevant here: the demands of leadership routinely exceed the developmental stage of the people being asked to lead. Organizations promote people into complexity they aren’t yet equipped to navigate — not because those people lack intelligence or skill, but because the form of mind required hasn’t yet developed. The answer, she argued, isn’t better training. It’s creating conditions for vertical growth.
The Block essay describes DRIs owning problems with full authority, player-coaches developing people without hiding in status meetings, and ICs making decisions without waiting to be told. Each of those descriptions assumes a self-authoring person. But self-authorship isn’t a skill you can assume people are interested in. It’s a practice that requires work and attention.
This matters because the instinctive response will be productization. Give everyone an AI copilot. Build tools that make existing people more efficient. Increase output per head. That’s a reasonable response to a productivity problem. It isn’t a response to a developmental one. You can automate the tasks a person does. You cannot automate the development of the person doing them.
Here’s where meaning becomes an economic variable, not just a philosophical one.
In a hierarchy, meaning was upstream of performance but downstream of structure. The structure told you what mattered. You performed accordingly. Leadership was largely about managing that loop — keeping people oriented, motivated, and aligned. In a de-layered organization, that loop has no structural anchor. The person has to provide it themselves. And a person who can provide it — who can orient themselves, hold direction under pressure, and make calls from their own values rather than waiting for permission — is categorically more valuable than one who can’t.
This is what self-actualization looks like when it enters an economic frame. It’s the difference between a person who expands the capacity of every system they touch and one who executes competently within it. The former is curious about development — specifically, vertical development.
Most organizational learning is horizontal: adding capabilities, skills, frameworks. More tools in the same toolbox. Vertical development is the expansion of the capacity to make meaning, not just the content of it. Not increasing what you know, but how you know it.
Garvey Berger’s work describes what that growth actually requires: not more information, not better frameworks, but experiences that confuse the current form of mind enough to prompt a genuine reorganization of it.
Leadership, in this context, is no longer a role or a title. It’s a developmental movement. It’s what happens when a person has enough internal structure to hold direction for others — not because they were appointed, not because they have information others lack, but because they have a self stable enough to be a reference point.
There’s something else the essay passes over. When Block describes what humans do at the edge — intuition, ethical judgment, cultural context, the feeling in a room — they’re right about the list. But they don’t say much about where those capacities come from.
Creative and intuitive capacity exists before language. Before analysis and articulation, there’s a pre-linguistic sensing — a felt orientation toward a problem, an aesthetic response that arrives before the argument for it. The best judgment you’ve exercised in a difficult moment probably preceded your ability to explain it.
This is not something AI can simulate, because AI operates entirely within language. It recombines what has already been said. It cannot sense what hasn’t been named yet. The edge worker Block is counting on is precisely someone who can do that. But that person is cultivated through a specific kind of development — sustained contact with a challenge, a reflection, and the experience of sitting with uncertainty long enough for something to clarify rather than resolving it prematurely.
Block is right that the structural constraint is changing. The question isn’t only whether AI can replace what managers do. It’s whether people can become what the new structure assumes they already are.
That’s a slower problem than an org redesign. It won’t be solved by tooling. And the companies that figure out how to build for it — not just around it — will have something that doesn’t appear on any capability roadmap.
A thoughtful person, operating from an evolving self and curious about their practice, is not a resource. They’re a compounding advantage. And right now, almost nobody is building the infrastructure to produce them.