An observation about awareness, attention and adaptability vs prediction, perfection and plans.
I've been thinking about the geometry of stability.
Stability in life, work, software and design.
We tend to think of it as something we build—a broad, flat surface made stronger through more information, more certainty, more planning and more control.
Nature suggests something different.
Imagine a sphere balancing on the apex of a pyramid.
The pyramid may be skewed. Its base may be tilted. Its geometry may be uneven.
And yet there can still be a configuration in which the sphere balances on its apex.
At a point.
Not along a line. Not across a plane.
Flattening the pyramid is an attempt to create the simplest possible environment: one that remains predictably stable all the time.
Sometimes that's exactly right.
Checklists. Protocols. Backups. Standards. These took decades to build, and they hold when it matters most. Where the stakes are high and the ground rarely shifts, the best move is to design the judgment out.
But a flat surface is built for the world as it was.
When the ground keeps shifting, the flattening never finishes.
So what if we focused instead on helping the sphere sense the dynamics of the pyramid—and continuously seek a point of equilibrium?
This distinction has been sitting with me.
Maybe resilience isn't built by predicting every possible future.
Maybe it's built by becoming exceptionally aware of the present.
Prediction still matters.
Models matter.
Second-order effects matter.
But I've started thinking about prediction differently.
Prediction isn't the objective.
It's one input into awareness.
Awareness informs judgment.
Judgment shapes action.
In an age when computation is becoming abundant, attention may become our scarcest resource.
Perhaps AI's greatest contribution won't be making decisions for us, but expanding what we're able to perceive—surfacing weak signals, revealing hidden relationships and widening our field of view.
The machine helps us see more.
But not more at the expense of attention.
More in order to focus it.
We still decide what matters.
We've spent decades optimizing for predictive intelligence—the ability to forecast what comes next.
I'm increasingly interested in something different:
Catalytic intelligence.
Not intelligence that predicts the future with perfect accuracy, but intelligence that continuously improves the conditions from which better decisions can emerge.
It values awareness over certainty.
Calibration over optimization.
Response over prediction.
Predictive intelligence seeks certainty.
Catalytic intelligence seeks leverage.
In a world of abundant intelligence, perhaps the central design question becomes less about how precisely we can predict what comes next, and more about how continuously we can discover where equilibrium exists now.
If this is real, it should show up somewhere.
Try this.
Think of the last five things that caught you off guard.
For each one, ask when it first became visible—and when it finally reached someone who could act.
The distance between those two moments is the thing.
If it's wide, awareness is what's holding you back.
If you saw it all early and still nothing moved, awareness was never the problem—and this essay isn't the one you need.
The goal isn't to make the pyramid still.
It's to help the sphere keep finding its balance.