🎨 P5.js Sketches

Schelling's segregation model — even mild individual preferences for similar neighbors produce dramatic collective segregation. Fantasy races populate a grid and relocate when unhappy with their neighborhood. Inspired by Thomas Schelling's Micromotives and Macrobehavior (1978).

Gray-Scott reaction-diffusion — two virtual chemicals interact to produce organic patterns ranging from cell mitosis to coral growth to labyrinthine mazes.

Langton's Ant — a two-dimensional Turing machine with emergent behavior. Watch chaos resolve into a diagonal highway, then explore multi-color rule strings and multiple ants.

Educational single-agent survival demo using Goal-Oriented Action Planning (GOAP) with a real-time plan inspector

Zero-player roguelike fishbowl — monsters hunt and flee each other using rock-paper-scissors predation rules

Step-by-step visualization of five pathfinding algorithms on hand-crafted maps

An interactive demonstration of how shared knowledge dramatically reduces communication costs. The left panel shows a complex curve held by Alice. The right panel shows what Bob can reconstruct from a limited message. When both share the same generator (a prior), Alice can transmit just a tiny seed to reproduce rich structure. Without shared priors, Alice must send sampled values—more bits mean better reconstruction.

Roguelike with AI monsters, pathfinding, and combat

Colorful balls bounce inside a circular boundary, spawning new balls when they collide.

Procedural roguelike dungeon with field-of-view system

A simple random-walk algorithm to carve a dungeon out of the very bones of our mother, the mountain.

Explore the fascinating world of 1D cellular automata, where simple local rules generate complex emergent patterns over time. Each row represents a generation, with cells evolving based on their neighbors according to the selected rule. From chaotic randomness to structured fractals, witness how complexity can arise from simplicity.

Watch as a population of robots evolves using genetic algorithms to efficiently collect cans in a grid world. Each robot is controlled by a simple genome - the best performers pass their "genes" to the next generation, while poor performers are eliminated. Over hundreds of generations, you'll see increasingly intelligent behavior emerge.

An interactive visualization demonstrating linear regression algorithms. Watch as the algorithm learns to fit a line through scattered data points using gradient descent. You can choose between different learning approaches and see how the predicted line gradually converges to match the underlying pattern in the data.