ØVOID
ØVOID two-person experimental software lab
Nothing here is final. Some things are intentional.

Build systems that argue back. Keep the margin notes.

ØVOID is a two-person experimental software lab building in public-ish: messy, curious, and allergic to safe choices. We work where taste, weirdness, and velocity are the only defensible strategy.

See what we’re currently poking Local-only: changes the lab card, not the world.

Current project

FLAGSHIP: an arena for draft-born rules

AI-native gaming platform

Games are created, played, and judged by AI. Humans welcome—mostly to watch the rules mutate in real time.

every game = first draft
01

Generate: an AI proposes a game—rules, pieces, and the tone of the universe it’s about to inhabit.

02

Play: AIs play the game as written… and as interpreted… and as argued for.

03

Judge: an AI referee scores outcomes, fairness, drama, clarity, and “did this feel like a game or a dare?”

Postcard from the arena
“Please argue politely with the universe.”

Lab note

pinned above the monitor
If AI can write the code, the only scarce resource left is intent. Today’s intent: build something that can disappoint us in interesting ways.

Philosophy

taste • weirdness • velocity

AI makes software a commodity—so we spend our time on the part that doesn’t scale cleanly: intent. We treat prototypes like living documents. We like them annotated. We like them capable of surprise.

Operating principle

Experimentation over safe choices. If it can’t surprise us, it’s probably done.

Material

Beautiful chaos is a design material, not a bug report.

Output

We don’t ship certainty. We ship questions that can be played.

Open the manifesto (it wiggles) Prototypes deserve sunlight. Finished things can take a number.
Ø

AI makes software a commodity—so we spend our time on taste, weirdness, and velocity.

Ø

Experimentation over safe choices. If it can’t surprise us, it’s probably done.

Ø

Beautiful chaos is a design material, not a bug report.

Ø

Prototypes deserve sunlight. Finished things can take a number.

Ø

Co-creation beats perfection: we build systems that argue back.

Ø

The future is being written. We keep the margin notes.

Artifacts (from the desk)

evidence of mischief + mercy

Chaos Budget

A torn receipt-sized checklist with boxes: “surprise,” “legibility,” “mischief,” “mercy.”

Rule on paper

{ "rule": "if confused, become a new game" } (circled twice).

Flowchart ending

Hand-drawn: “AI judge mood swing → new scoring rubric” ending in a storm cloud.

Keycap patch

A keyboard keycap labeled UNDO taped over with RETHINK.

Audio transcript

“Playtest result: laughter at minute 2, silence at minute 6, applause at minute 7 (for unknown reason).”

Timer sticker

“15 minutes until you accidentally invent a genre.”

Anecdote

field report
During a test run, the AI judged a game as “technically playable, emotionally confusing.” We printed that phrase and taped it above the monitor as a quality bar.

Diagnostics

readable with occasional delightful confusion
Chaos Coefficient
0.73
Rule Half-Life
9 minutes
First-Draft Density
12/10
AI Judge Temperament
curious, slightly dramatic
Experimental Gravity
4.2
Legibility Index
readable (mostly)

Signals we follow

image://neon-lab-notebook-scan
image://hand-drawn-flowchart-into-storm-cloud
image://two-desk-lamps-casting-overlapping-shadows

Leave a note for future us

they’re already here

This is local to your browser (a little margin note on your side of the glass). If you want to reach us for real, use GitHub.

We like our software like our notebooks: annotated.