PUBLICATION / 001—∞
Creative direction, engineered experiences, artificial intelligence, and the strange signals connecting them.
ENTER ARCHIVETHREE NOTES
ONE PRACTICE
I WRITE TO FIND THE SHAPE OF A THING.
THEN I BUILD IT TO SEE IF THE SHAPE HOLDS.
The work begins before there is a screen, a frame, or a line of code. It begins as pressure: a sense that something should exist and a refusal to let it arrive without a point of view.
Creative direction is often mistaken for choosing the surface. I think it is closer to designing gravity. It decides what belongs, what moves, what remains still, and what every later decision will orbit.
I SEE THE SYSTEM BEFORE IT EXISTS.
Engineering enters at the same moment. A visual world that cannot survive contact with a real interface is only an image. A system without feeling is only infrastructure. The useful territory is the overlap: where atmosphere becomes behavior and intention becomes an executable rule.
Start with one feeling precise enough to reject the wrong ideas. Build a small system around it. Let type, motion, sound, silence, and code repeat the same underlying logic in different forms.
The result should feel discovered rather than decorated—as if the final object was already waiting inside the original instinct.
An image can be a destination. It can also be an instruction.
I like moving back and forth between visual material and software until the boundary becomes unhelpful. A crop suggests a layout rule. Compression becomes texture. A motion test becomes an interaction model. Code returns the image changed, alive, and sensitive to the person looking at it.
THE INTERFACE IS NOT A FRAME AROUND THE WORLD. IT IS PART OF THE WORLD.
This loop is why prototypes matter. They expose qualities that static explanation hides: weight, delay, friction, rhythm, and surprise. The prototype is not merely proof that something works. It is a way of discovering what the thing wants to be.
Look closely. Extract a rule. Build the rule. Watch what it produces. Feed the result back into the image. Repeat until the system has a pulse.
The least interesting use of a new material is to paste it onto an old shape.
Artificial intelligence becomes useful when it changes the structure of an experience: what can be made, how quickly an idea can be tested, how a system responds, and where authorship is exercised. It should not appear merely because the current moment expects it.
INTELLIGENCE NEEDS DIRECTION. POSSIBILITY NEEDS TASTE.
The model is not the point of view. It is a field of possibility with particular textures, failures, and tendencies. Direction is the act of shaping that field—building constraints that are strong enough to produce character without reducing everything to repetition.
Remove the label “AI” from the project. Is the experience still meaningful? Does the intelligence create a capability, relationship, or feeling that could not exist otherwise? If not, the material is decoration.
The goal is not to make software pretend to be human. The goal is to make a new kind of instrument, then learn how to play it with intent.