Skip to content
scsiwygest. ‘26
Sign in
get startedmcpcommunityapiplaygroundswaggersign insign up
David Olsson·The How-Tos of How-Tos Is the Control Plane15 May 2026David Olsson
David Olsson

The How-Tos of How-Tos Is the Control Plane

#orchestration#intelligence-fabric#knowledge-work#substrate#control-plane#worksona#atomic47

David OlssonDavid Olsson

A pattern keeps surfacing across the work we've been doing on substrate, on layered intelligence, on orchestration as a discipline. It took a while to name. The name is the post.

The how-tos of how-tos is the control plane. Not how to do something. The why of that something — encoded, governed, executable, and binding on the work below it.

Most knowledge work still operates one level too low. People own tasks. They own how to do tasks. They optimize the doing. The doing gets faster, the doing gets cleaner, but the doing never asks itself why it is the doing it is. That question lives in a slide deck somewhere, reviewed quarterly, disconnected from the keystrokes.

The shift is moving that question — the why of the how — into the operating layer. Making it the thing that orchestrates. Making the rest of the work answerable to it.


What the Substrate Taught Us

We've been building substrate the long way: from the bottom. Page-level transcription. Daily synthesis. Weekly synthesis. Longitudinal themes. Each layer is a different temporal aperture on the same underlying material. A page is a moment. A day is a thread. A week is a current. A theme is a direction. None of these are reports. They are intelligence at different time signatures, each one composed from the layer below it and feeding the layer above.

What this layered substrate revealed — and we did not see it until we had run it for a few months — is that the moment a single observation enters the system, it is already being addressed by every higher layer. The daily synthesis is not waiting for the day to end before it has a posture on the moment. The weekly synthesis already has a frame the day will be fit into. The themes are already exerting selection pressure on what counts as worth surfacing.

This is not top-down. The themes did not come from a plan. They emerged. The pressure they exert is the pressure of accumulated evidence — what the work has actually been about, not what someone said it would be about. Bottom-up percolation produced top-down governance. The top-down governance now changes the bottom-up work, because the bottom-up work knows what the top is asking.

The system is doing the why of itself while it does itself.


Orchestration Was the First Hint

The earlier theme — orchestration over specialization — pointed at this without naming it. We learned that the leverage was not in any single skill. The skills were commodity. The leverage was in the conductor: the one piece that decided which skill to call, in what order, against which state, with which approval gate.

Conductors are not faster than specialists. They are answerable in a way specialists are not. A specialist does the thing. A conductor decides whether the thing should be done, whether the prerequisites are met, whether the output meets the rigor, whether the next step is licensed. The conductor carries the why. The specialists carry the how. Until you separate them, you cannot govern either.

The control plane is where conductors live. It is not a tool. It is a posture: the work has a why, the why is structured, the why is enforceable, and the doing happens beneath it.


Compliance as a Side Effect

The market reads this as governance and reaches for the wrong word. Compliance, in most organizations, is a friction layer bolted on after the fact. Audits exist because the work has no native memory of itself. Approvals exist because the doing has drifted from the deciding. Documentation exists because nobody can reconstruct what happened.

When the control plane is the why of the how, none of that is bolted on. The work writes its own audit trail because the orchestrator logs the why behind every step. Approvals are not gates added at the end — they are the structural moments where the why was checked against the doing. Documentation is the natural exhaust of the system, not a quarterly campaign to produce it.

Rigor stops being expensive. It becomes the shape of the work.

This is what we mean when we say docked: how things actually get executed is bound to how they are decided, in a way the team can see, in a way an outside party can inspect. The docking is the compliance. The compliance is the structure. Neither is overhead.


What Shifts in the Market Posture

Once a team operates from the control plane, the way it shows up to clients changes. The conversation is no longer about deliverables. It is about how the deliverable was decided, what governed it, what rigor it passed through, what evidence it accumulated against, what the system will remember about it next quarter. The client is not buying the artifact. They are buying access to the apparatus that produces artifacts of a known kind, repeatedly, traceably.

This is the shift from selling work to selling work-systems. It is also the shift from individual expertise to institutional intelligence. A consultant carries their how-tos in their head. A control plane carries the how-tos of how-tos in shared state, where the team operates from the same posture without re-deciding the posture every engagement.

The market posture changes in a second way that is harder to see from outside: the team's relationship to its own work changes. When the why of the how is the operating layer, individual contributions are not heroic. They are situated. People stop competing on doing and start composing on deciding. The internal disagreements get sharper because they are about the why, which is the thing actually worth disagreeing about. The agreements get more durable because they are written into the orchestrator, not into a memo.


The Cusp

This is the part I want to be careful about. It is easy to make grand claims about knowledge work and have them be wrong. So narrowly:

For two years the operating question was how do we do more work with AI. The honest answer turned out to be not much more, if you keep the work shape the same. AI accelerates doing. Doing was never the bottleneck for serious teams. Deciding was the bottleneck. Coordinating was the bottleneck. Remembering was the bottleneck. Governing was the bottleneck.

The teams now pulling away are not the ones with the best prompts. They are the ones who moved the why of the how into the operating layer. They built substrates that accumulate. They built orchestrators that decide. They built control planes that govern. The doing got faster as a side effect — and the doing matters less than it ever did, because the doing is no longer where the differentiation lives.

The differentiation lives one level up. In the how-tos of how-tos. In the control plane.

That is the shift. Not a faster wheel. A higher gear.


What This Means for the Next Year

A few things become obvious from this vantage that were not obvious before.

Knowledge work tools that do not expose a control plane will commoditize. Tools that let teams compose their own control planes — that let the why be written, versioned, governed, executed — will accrue the value. The platforms that win are not the smartest. They are the most composable into a team's own governance.

Methodology becomes the asset. Not the documents. The encoded methodology — the orchestrator that runs it, the state that holds it, the gates that enforce it. A team's competitive position is no longer "we know how to do X." It is "our control plane runs X to a standard, repeatedly, with memory, and the standard improves itself from what it learns."

The role of senior people changes. It stops being doing the hard work and starts being deciding the why of the hard work — and encoding that why so the team can operate against it without them in the room. The org chart flattens because the control plane is what's vertical. Seniority is measured in how much of the control plane you authored and how cleanly it composes with the rest.

Procurement starts asking a different question. Not "what can your team do." But "show us the control plane your team operates from." The answer to that question is the actual asset. Everything else is execution against it.


Where We Are

We are early in this. The substrate is running. The orchestrators are running. The layered intelligence is producing output we now trust enough to plan against. The control plane is forming — visibly, in ways the team can point at — and the way we engage clients and the way we engage each other has already shifted in ways that surprised us.

The biggest shift, though, is what the work feels like from the inside. There is less re-deciding. There is less rebuilding the argument from scratch. There is less drift between what we said we'd do and what we did. The work has a memory now, and the memory has a posture, and the posture is the control plane.

The how-tos of how-tos. Not what we do. Why we do it, made operational.

That is the shift. That is what's actually changing.

Share
𝕏 Post