Single-file apps, interactive tools, games, and generative art — the joy of shipping something before the doubt has time to kick in.
These are the things that exist because someone said "I wonder if I could make that" and then just did. No planning documents. No sprint reviews. Ship first, name it later.
The builds that survived the panic phase and grew into something larger.
The tools, techniques, and philosophies that keep showing up across every project.
HTML, CSS, and JavaScript in one file. No build steps. No dependencies. Load it, read it, understand it. The discipline of keeping everything in one place forces clarity.
Drawing custom worlds with pixels, paths, and triangles. The browser has a surprisingly capable graphics pipeline once you get past the abstractions. We go past the abstractions.
Algorithms that surprise their own authors. Levels, textures, music, enemies — generated from rules and seeds rather than authored by hand. Infinite content, finite code.
Making noise on purpose, with math. The Web Audio API gives you oscillators, filters, envelopes, and a bus to the speaker. From chiptune bleeps to full synthesizer patches.
Update, render, repeat — the heartbeat of interactive software. Fixed timesteps, interpolation, and the surprisingly deep question of how to make a simulation feel right.
MCP bridges, hot-reload servers, and the tools for building tools. Sometimes the meta-work of making development smoother is more interesting than the project itself.
Four things we believe about making software — tested through repeated acts of reckless shipping.
A prototype that runs teaches you more than a design document that doesn't. Get something on screen. The feedback loop of something real outweighs every prior assumption.
One file. One canvas. No libraries. The restrictions you impose on yourself — or that circumstances impose on you — often produce the most distinctive work.
The first version of anything is supposed to be bad. Committing to bad-first removes the paralysis of trying to start well. Get the logic right, then make it beautiful.
The features you don't build are the ones that can't break. Scope creep is a project's natural cause of death. The question isn't what to add — it's what to leave out.