Experiment 001
Three-Body Problem
Three celestial bodies locked in gravitational dialogue—their paths deterministic yet unpredictable. A real-time simulation of one of physics' most beautiful unsolved problems, rendered in light and motion.
On Chaos and Craft
In 1687, Newton solved the two-body problem—two objects in gravitational orbit follow paths you can write as a single equation. Add a third body, and the equation dissolves. No closed-form solution exists. The system becomes exquisitely sensitive to its opening conditions: shift a decimal point and the entire future rewrites itself.
Deterministic does not mean predictable. The rules are simple. The outcomes are not.
The three-body problem lives at the intersection of order and chaos—a space we navigate in every design project. Brand systems are gravitational: identity, audience, and market pull on each other in ways that are technically deterministic but practically impossible to predict from first principles alone.
What you see above are periodic orbits—rare, stable choreographies discovered by mathematicians over the last three decades. The figure-eight, found by Cris Moore in 1993 and proven by Chenciner and Montgomery in 2000, shows three equal masses tracing a single infinity loop. The butterfly and moth orbits, catalogued by Milovan Šuvakov and Veljko Dmitrašinović in 2013, reveal that stability hides in the strangest geometries.
Craft is finding the periodic orbit inside the chaos.Read the full essay →
Experiment 002
Rössler Attractor
Three coupled equations. An infinite, non-repeating trajectory folding through space. Otto Rössler's 1976 system is one of the simplest recipes for chaos—a spiral that periodically spikes into a third dimension, never quite retracing its path.
On Simplicity and Surprise
In 1976, Otto Rössler was not looking for beauty. He was looking for the simplest possible system of equations that could produce chaos—deterministic behavior so sensitive to its starting conditions that prediction becomes impossible. He found it in three lines of calculus.
The simplest rules can produce the most complex behavior. That is the lesson of the Rössler attractor—and of every good brand.
The system traces a flat spiral in two dimensions, unremarkable on its own. But periodically, the third variable spikes—the trajectory folds upward, rotates, and re-enters the spiral at a slightly different point. This fold is where the chaos lives. The path never exactly repeats. It is deterministic yet infinite in its variation.
We see a parallel in the work. A brand identity is a set of simple rules—a color palette, a typeface, a mark, a voice. The rules are finite. But the encounters are infinite: a sign glimpsed from a highway, a menu held in lamplight, an email opened at a desk. Each encounter folds through its own context, its own moment. The same system, never the same experience.
The Rössler attractor reminds us that constraint is not the enemy of surprise. It is the source.
Read the full essay →Experiment 003
Chaos Typography
The Lorenz attractor—a mathematical model of atmospheric turbulence discovered in 1963—becomes the engine behind a new typographic language. Twenty-six letters of Helvetica Extra Bold, each driven by its own chaotic system, rendered as 3D wireframes in real time.
On Chaos and Language
In 1963, MIT meteorologist Edward Lorenz re-entered a value from a printout—0.506 instead of the full 0.506127—and watched a completely different weather pattern emerge. This accident became the founding observation of chaos theory: deterministic systems can produce wildly divergent outcomes from nearly identical starting conditions.
The more neutral the letterform, the more visible the chaos that disrupts it.
Each letter of the alphabet is extracted from Helvetica Extra Bold through pixel-level contour tracing, simplified with Douglas-Peucker reduction, and refined through a corner-aware Catmull-Rom spline. Each letter is then assigned its own independent Lorenz attractor, initialized with slightly different starting conditions. Because of sensitive dependence, every letter evolves along a unique chaotic trajectory, displacing its 3D wireframe in real time.
The tension between the typeface’s rigid rationalism and the attractor’s deterministic unpredictability is the entire point. What emerges is a layered palimpsest of typographic forms—each one obeying its own attractor, each one pulling away from the others in a way that is deterministic but visually unpredictable.
Read the full essay →Experiment 004
Emotion Gradient Backgrounds
Fifteen animated background gradients, each tuned to a specific emotional state. A WebGL experiment testing whether color and motion can carry feeling as clearly as typography carries voice.
On Color and Feeling
Pull up ten homepages from the last six months. Any sector, any budget. Nine of them have the same background: a cyan bleeding into purple, bleeding into pink. A soft blur. A subtle animation. The effect is pleasant and instantly forgettable. We call it the default AI gradient, not because AI made it (most of these were built by human designers) but because it is what an AI would make. A color choice that offends nothing and speaks of nothing.
A well-tuned background is something you feel, not something you see.
Every designed surface has a mood, whether the designer chose it or not. When the mood is chosen, the surface does emotional work. When it is not, the surface falls to the default, and the default is always generic. These fifteen gradients are an attempt at choosing mood on purpose, calibrated in OKLCH color space for perceptually even transitions, with motion signatures tuned to the frequency of the emotion itself. Slow drift reads as peace. High-frequency jitter reads as anxiety. Each mood has ancestors: Sugimoto seascapes, Fragonard pastels, Byzantine mosaic, wet asphalt under cellular-tower light.
Read the full essay →