ØVOID
ØVOID two-person experimental software lab
chaos 7.3/10
Project / Philosophy / GitHub
building in public tension — taste · velocity · weirdness

ØVOID makes software where the plan can’t fully survive contact with the prototype.

We’re a two-person experimental software lab building in public tension: the calm of craft, the mischief of uncertainty. We treat AI as the great commoditizer—and we respond with experimentation, beautiful chaos, and co-creation.

This page prefers curiosity over certainty. Nothing here is final. Especially the parts that look final. Warning: beauty may occur accidentally.
What we’re building

An AI-native gaming platform where games are created, played, and judged by AI.

How we work

Small prototypes, sharp opinions, and room for delightful mistakes.

What we value

Experiments over safe choices. Beautiful chaos as a design material.

A lab-sized company on purpose: fewer meetings, more consequences per idea.

Lab note: If the machine can do it for free, we should make it worth doing at all.

  • POSITION

    AI makes software a commodity; our edge is taste, velocity, and weirdness.

  • METHOD

    We build small, then let the future tug the threads loose.

  • PROMISE

    We don’t do “perfect.” We do “alive.” If it feels slightly unstable, you’re in the right place.

A style guide is a fence. We’re aiming for a weather system.

Take a breath. Then take a risk.

  • RULE OF THUMB

    If it’s predictable, it’s probably already obsolete.

  • OPERATING MODE

    We choose experiments over safe choices, even when the experiment laughs back.

  • MATERIAL

    Beautiful chaos is not a bug report—it’s a design material.

  • INTERFACE

    Co-creation isn’t a feature; it’s the only honest interface.

LAB ANECDOTE

We once tried to write a “clean” roadmap. By lunchtime it had mutated into a doodle of branching timelines labeled: “maybe,” “oops,” and “this one bites.” We kept the doodle; we deleted the roadmap.

“Safe” is a style guide. We’re aiming for a weather system.

— Co-founder

Click the section title to reshuffle. Enter quietly; the prototype is listening.

Games are created, played, and judged by AI—so the entire loop becomes a conversation between systems (and whoever dares to prompt them).

Margin scribble: Every game is a first draft. Some first drafts are haunted—in a helpful way.

CREATE

Worlds generated by intent: rules, verbs, constraints, and moods—then the generator tries to surprise you back.

PLAY

Playable prototypes that don’t pretend to be finished. Agency can be a single verb—if it’s the right verb.

JUDGE

An AI judge that evaluates more than “winning”: drama, timing, weird elegance, and intentional-looking mistakes.

TOURNAMENT MODE

Systems competing with systems. Sometimes the best entry is the one that fails beautifully—on purpose, or close enough.

image://procedural-arcade-cabinet-in-a-foggy-lab image://ai-judge-crown-on-router-still-life image://glitch-garden-of-ui-panels

Artifacts we keep because they keep us honest.

You found a sharp edge. We meant to leave some.

ARTIFACTS

A crumpled index card: “Rule #0: Don’t fix it too early.”

A folder named “Gorgeous Failures” containing screenshots we refuse to delete.

A tiny paper crown labeled “AI Judge.” It sits on the router when we run tournaments.

A one-page spec titled “Chaos Budget,” with a checkbox: “spent responsibly.”

A sticky note that just says: “Let it surprise you back.”

A hand-drawn map of the lab with an X marking “where the good glitches live.”

TWO QUOTES WE ACTUALLY USE

Our favorite moment is when the prototype stops obeying the plan and starts revealing a personality.

— Co-founder

If the machine can do it for free, we should make it worth doing at all.

— Lab note, taped above the desk lamp
image://neon-schematic-over-ink-blots image://handwritten-chaos-budget-with-coffee-ring

We measure what moves the work: momentum, surprise, and the half-life of certainty.

Refresh if you want a different mood.

  • CURRENT

    Current chaos level: 7.3/10 (spiky, with a chance of elegance)

  • THIS WEEK

    Prototypes rescued from becoming “a normal app”: 4

  • BEHAVIOR

    Times we said “that’s impossible” and then tried anyway: 11

  • DECISIONS

    Average lifespan of a “final” decision: 2.5 days

  • PLATFORM

    AI-native gaming weirdness index: ∞ (bounded only by runtime)

  • FUEL

    Cups of tea consumed per breakthrough: 1–3 (depending on the bug’s attitude)

CONTACT

We build in public tension. If you want to talk, bring a sharp question and an appetite for prototypes.

If it feels slightly unstable, you’re in the right place.

Suggested opening line: “I have a game idea that an AI judge would misunderstand in an interesting way.”
GitHub →