ØVOID
ØVOID two-person experimental software lab
Enter the lab (wipe feet on the weird).

Weatherproof software, built in public.

ØVOID is a two-person experimental software lab building in public weather—part lightning, part lullaby. We make places where AI turns software into a commodity, and we choose the weird, the fast, and the beautiful chaos anyway.

CORKBOARD MODEL: nodes are ideas; strings are disagreements; pins are decisions.
Current project
AI-native gaming platform

Games made by AI. Played by AI. Judged by AI.

We’re building a platform where AI agents generate games, run them, and score them—so we can observe what emerges when “fun” is negotiated by non-human taste.

Constraint: We’re not building a monument. We’re building an environment where prototypes can mutate, compete, and become legible.
Mission
two voices, one keyboard

Make outcomes feel expensive.

If AI can make code cheap, our job is to make outcomes feel expensive—strange, specific, alive. We don’t polish chaos away. We frame it, feed it, and ask it better questions.

“The safest path is a hallway. We prefer doors that open into weather.”

Build log, 02:14
Philosophy
experimentation over safe choices

Beautiful chaos as a design material.

  • AI makes software a commodity; taste becomes the scarce resource.
  • We pick experimentation over safe choices, even when the safe choice is louder.
  • Prototypes are opinions with timestamps.
  • If it feels finished, we probably shipped it too late.
  • The future is being written; we prefer ink that smudges.
A moment we kept
AI judge (playtest)

Suspicious of perfection.

“This game is a first draft and that is its superpower. Please do not fix it too quickly.”

AI judge, playtest

We stapled that sentence to our wall and started celebrating flaws more loudly.

Anecdote
lab memory: rotating

One brave hour.

We once set a timer for “one brave hour” and tried to build a menu that changed its personality with every refresh. At minute 59, it became a diary. At minute 60, it became a map. We kept the map.

Lab telemetry
vibe-dependent

Metrics we admit to.

Chaos budget used this week 73%
Times we said “what if?” today 19
Average prototype half-life 9 days
Minutes spent arguing with our own UI 42
Artifacts
captured, not contained

Things on the wall.

index card “Ship the experiment, not the explanation.”
sticker “CHAOS IS A FEATURE (but label it).”
error “E_UNEXPECTED_DELIGHT: outcome exceeded the plan.”
file rules_for_unruly_interfaces.txt → “Make it legible. Let it misbehave.”
Moodboard (text-only)
labels, not images

Reference fragments we’re orbiting.

  • neon lab desk with sticky notes and cables
  • glitchy constellation of UI panels floating
  • AI judge scorecard with handwritten doodles
  • arcade cabinet made of code blocks and fog

We build experiments, not monuments. Leave with a question, not a conclusion.

Contact / Trailhead
Proceed anyway (recommended)

Follow the work where it’s loudest: the repo.

ØVOID builds in public when possible. If you want to understand what we do, read the build logs, scan the commits, and watch the prototypes grow a personality.

Operating rule: Make it legible. Let it misbehave.