The Player-Coach Returns

When Jack Dorsey published Block's organizational theory in April 2026, one of the three surviving roles he named was the Player-Coach: "Senior practitioners who develop others' craft by doing the work alongside them, not by managing status meetings." The phrasing is almost an insult to the manager class — not by managing status meetings — and it's deliberate. Dorsey had just cut about 4,000 people from Block, and the coordination work much of middle management performed had moved to two AI "world models" that supply context and institutional memory directly.

What didn't move was the work of making other people better at their craft. That stayed human. And it changed shape, from the manager who directs to the practitioner who demonstrates. The Player-Coach is back — not as a sports metaphor borrowed for the office, but as the organizational technology for transmitting taste at the scale AI-Born firms require.

The model that's ending: management as a separate profession

For a century, we treated management as a distinct discipline. You could be a great manager of engineers without writing code, a great manager of designers without designing. The skill was coordination — routing information, resolving conflicts, running the meetings that kept a large organization synchronized. This wasn't a mistake. When organizations grew past the span any one person could oversee, you needed layers, and layers needed people whose job was the layer itself.

The fair version of the case for that model: coordination is real work, and someone has to do it. Information doesn't route itself through a 10,000-person company. The management layer existed because the alternative was chaos.

But notice what justified it. The constraint, as Dorsey put it, is the same one the Romans faced: narrowing span of control means adding layers, and more layers mean slower information flow. The question was never whether you needed layers. It was whether humans were the only option for what those layers do. Once that answer changes, the separate profession of management loses its foundation — and what's left is the part management was never actually about.

The reframe: develop people by doing, not directing

Strip away the coordination and one function remains: building craft in other people. That function was never well served by the status meeting anyway. It's served by proximity — by the experience of watching a master encounter an edge case and make a call.

This is why the Force Multiplier in the The New Triumvirate: The Three Roles That Survive is, in practice, a Player-Coach. She doesn't manage outputs. She does the work alongside her team, transmitting standards through example. When the system fails, she works through the failure with people rather than explaining what went wrong after the fact. When escalations involve competing aesthetic or technical priorities, she models the judgment process in real time rather than issuing a ruling from above.

Figure: The overseer who commanded gives way to the orchestrator who designs and the coach who demonstrates — direction through doing, not through layers.

Andrej Karpathy is the clearest illustration. He doesn't make his engineers better by generating outputs they review. He makes them better by working through hard problems beside them, narrating his judgment as he goes, modeling the thought process good engineering requires. Anna Wintour, Virgil Abloh, Dieter Rams operated in the same register — transmitting standards through sustained proximity, not through org-chart authority. The Player-Coach model is how you do that at scale.

The mechanism: why proximity can't be automated

Here's the part that survives even a world where AI matches human judgment. Even if a machine can exercise taste, it can't transmit taste through the kind of collaborative doing that builds genuine craft capacity in another person.

Consider a Force Multiplier at a fashion platform who senses a recommendation engine drifting toward monotony — a pattern the metrics celebrate as record customer approval. She makes the architectural fix. But that's not where the value is. She brings her two junior stylists into the process. They sit with the data together. She narrates her discomfort in real time: why this combination feels stale, what signal her eye caught that the metric missed, how taste-fatigue differs from quality-fatigue. By the end of the afternoon, both stylists have internalized something no training document could convey. They watched a practitioner notice something true before she could explain why it was true.

That's the Player-Coach's real product. Not the fix — the team that can build the next one without her. You cannot extract that into a document, a course, or a model output, because the thing being transmitted is judgment under live conditions, and judgment under live conditions only transmits by being witnessed in action.

What to do about it

1. Stop promoting your best practitioners out of practice. The traditional reward for craft mastery was a management role that removed you from the craft. In an AI-Born firm that's backwards — your masters are most valuable when they're still doing the work, visibly, beside others.
2. Measure transmission, not just output. A Player-Coach who ships brilliant work but leaves no one able to replicate it is underperforming the role. Ask: who got better because they worked beside this person?
3. Protect the time to work alongside. Coordination collapse should free senior people into the craft and into teaching it, not into more meetings. If your best people are still buried in status updates, you've automated the wrong layer.
4. Don't rebuild the management layer agents just dissolved. The temptation, post-restructuring, is to recreate the comfort of supervisory hierarchy. Resist it. The roles shrink because the plumbing runs on agents now — adding human layers back is coordination leaking in.

The principle

The manager who only manages is becoming an artifact of a coordination problem that no longer needs a human solution. What endures is older and harder: the master who makes others masters by working in their company. We called that person a Player-Coach long before AI, in the rare organizations wise enough to keep their best people doing the work. The difference now is that it's not the exception. It's the role.