Ascend Before You Descend
“Draw the map, then descend” hides an assumption: that you can already draw the map.
Most people cannot, not at first. The map is not lying there waiting to be traced — it only becomes visible from altitude, and you do not begin with altitude.
Before the first line of the map, there is a climb.
The descent gets all the attention, because it is where the work looks like work. The ascent is the part nobody sees, and it is the part that takes years.
I did not have the altitude when I started. Early on I solved the problem in front of me, the way everyone does. What I did not notice was that each problem was a rep. Enough exposure to enough domains, enough processes, enough constraints, and something quiet happens: you stop seeing isolated problems and start seeing shapes you have seen before.
The terrain stops being unfamiliar. That recognition is altitude, and it is earned, not issued.
The mechanism is less mystical than it sounds. It works like a practiced reframe — again and again, stepping back to the whole before touching the part, until the reframe stops being a practice and becomes your default. Not motivation. Repetition hardening into perception.
After enough reps you no longer visualize the system by effort; the system is simply how the thing now appears to you. The map stops being something you make. It becomes how you see.
And from up there, one thing changes that changes everything: ambiguity stops reading as a threat.
At ground level, a black box is a risk — an unknown that could blow up the plan. From altitude, the same black box is just an unshaped space, and you already have a feel for how spaces like it tend to resolve. The fear becomes an opening.
This is the real starting point of turning ambiguity into governed execution — not the execution, but the moment ambiguity flips from problem to opportunity.
Here is the part that took me longer to understand.
In a culture wired for tactical outcomes, the climb is invisible and the artifact is everything. You draw a diagram in ten minutes and people conclude you are good at diagrams. The diagram is the thing they can see, so the diagram gets the credit.
They are crediting the exhaust and missing the engine.
And now AI can produce the diagram too — faster than I can, often well enough. If the diagram were the skill, the skill would be over.
It is not.
And the reason is the whole point. When the artifact gets cheap, everything that was never the real skill falls away, and what is left standing is the only thing that ever mattered: the perception that knew what the diagram should contain.
AI did not erode the moat. It dissolved everything around it and left the moat exposed.
The climb was always the asset. The artifact was only ever proof the climb happened.
So if you are early on this path, here is what I would tell you. Do not optimize for the artifact. The artifact will keep getting easier — let it. Optimize for the reps that build altitude: take the problems one size too big for you, sit in a domain long enough to feel its constraints, force yourself back to the whole before you reach for the part.
It will not feel like progress for a long time, because the climb never announces itself. Then one day the map is just there when you look — and the ambiguity that used to frighten you looks, instead, like an opening.
That is the ascent.
The descent is the visible half.