The Deeper You Can Imagine a Structure, the More You Can Control It

I keep coming back to a sentence I said out loud last week and then couldn't shake: the better I can imagine something, the better I can control it.

That sounds almost trivial when you read it once. Then I started running it against everything I do, and it stopped being trivial. It became the rule that explains why some of my projects feel obvious and effortless, and others feel like wrestling smoke.

Control is not a personality trait. It is not willpower or discipline or even experience. Control is a function of mental resolution. You can only operate inside a system as deeply as you can imagine its structure. Everything else is hope dressed up as competence.

The surgeon and the textbook

A surgeon who can hold a real, three-dimensional, vascularised body in their head — the way the fascia slides, where the bleeders are likely to hide, what this specific patient's anatomy probably looks like under the skin — operates differently from a surgeon who learned the body from 2D textbook diagrams. Same hands. Same scalpel. Same procedure on paper. Radically different outcomes.

The first surgeon is operating on the actual system. The second is operating on a flattened cartoon of it and praying the cartoon is close enough.

This is not a metaphor. This is what is actually happening, in code, in business, in relationships, in everything I touch every day.

Code, business, ops

A developer who can hold an entire system architecture in their head — the request lifecycle, the data flow, the failure modes, the cache invalidations, what happens when this service is offline — writes different code than a developer who only sees the function in front of them. The second one isn't worse. They just have a thinner model. Their resolution is lower. They will keep introducing bugs the first one literally cannot introduce, because the first one can see the collision before they type it.

A founder who can imagine the whole customer journey, the unit economics, the cash position, the team's morale, the partnership pipeline, and how all of those couple to each other — makes different decisions than one whose mental model is "we should grow." Both are running businesses. Only one is running the actual business.

I noticed this most painfully when I started running multiple projects at once. The deeper I can imagine the team — what each agent is doing right now, what they're stuck on, what they need from me, what they will need next week — the more I can delegate without it falling apart. The shallower my mental model, the more I have to micromanage, because micromanagement is what you do when your imagination of the system is too thin to trust.

Delegation is not a leadership skill. Delegation is the privilege of having a high-resolution model of the work.

Why intuition feels like magic

The reason a 20-year veteran in any field looks like a wizard is not that they're smarter. It's that their internal model has more layers, more edge cases, more "this looks fine but actually it's about to break" patterns. They are not predicting the future. They are running a higher-fidelity simulation than the room.

Conversely, the reason most strategy fails is that the strategist had a flat picture of a fractal world. They could imagine layer one, maybe layer two. The system has eight layers. Reality reaches into layer five and takes the plan apart.

You cannot beat someone whose mental resolution exceeds yours. You can only get lucky.

Where this points

If control is bounded by mental resolution, then the highest-return investment is almost always the same one: deepen the model. Read original sources, not summaries of summaries. Sit with the system long enough that you start dreaming about it. Talk to the people who actually live inside it — the operator, the customer, the patient, the contributor — until your simulation includes their reactions and not just yours. Build the picture deeper than your competitors are willing to.

This is, I think, the most honest case for using AI well. I do not use AI to replace my mental model. I use it to raise its resolution. To pull in the part of the system I haven't sat with yet. To run scenarios I'd otherwise hand-wave past. To turn a 2D mental sketch into something closer to 3D before I act on it. The people who lose to AI will be the ones who let it think for them. The people who win with it will be the ones who use it to think more deeply themselves.

The opposite move is also visible: when you're shocked by an outcome — the partner left, the launch flopped, the system fell over, the team quit — what almost always happened is that reality had a layer you didn't imagine. You were operating on a flat map of a layered place.

The honest question

So the question I'm sitting with this week, and the one I want to leave open here, is this:

Where in my life is my mental resolution lowest? Which system am I trying to control with a cartoon-level picture of it? Money? Health? A specific relationship? A specific project on the portfolio that I keep nudging without ever actually understanding why it underperforms?

And what would it cost — in hours, in conversations, in books, in uncomfortable honesty — to deepen that picture by one full layer?

That cost is almost always smaller than I think. And the return — measured in things I no longer have to wrestle, decisions that stop being agonising, surprises that stop being surprises — is almost always larger.

Imagine deeper. Then look at what just became controllable.