ØVOID is a two-person experimental software lab building in public-ish, iterating in strange loops. We treat AI as the force that makes software a commodity—so we chase the parts that can’t be commoditized: taste, risk, and beautiful chaos.
Two humans, one lab notebook, and a habit of shipping prototypes with fingerprints still on them.
An AI-native gaming platform where games are created, played, and judged by AI.
Experimentation over safe choices. Safety is a feature request, not a philosophy.
Not a “generator” that finishes everything—more like a collaborator that proposes rules, twists, and incomplete elegance.
The platform treats playtesting as a living thing: the AI can be curious, stubborn, poetic, or accidentally profound.
Judgment isn’t just a score; it’s an annotated rationale. Sometimes it’s decisive. Sometimes it asks for a rematch against silence.
AI makes software a commodity—so we spend our time on the uncommoditizable: experiments, aesthetics, and weird courage. We prefer first drafts that move to polished plans that stall. Unpredictability isn’t a bug; it’s a collaborator with sharp elbows.
“The only stable thing here is our willingness to be surprised.”
“If AI makes software cheap, our job is to make outcomes interesting.”
“Every game is a first draft. Chaos isn’t a flaw—it’s the medium.”
“Experimentation over safe choices. Safety is a feature request, not a philosophy.”
A crumpled index card labeled “RULE #0: Don’t sand off the weird.”
“AI Judge Mood: mercurial / curious / occasionally poetic.”
apologies_to_future_us.txt containing only: “we meant well; we meant weird.”
Prototype button label: “Begin the unpredictable part.”
The fastest way to reach us is where the work lives. If you’re curious about the platform, want to collaborate, or have a good question with sharp edges—drop it on GitHub.
People who can argue for a constraint and still protect the weird.
Tell us what feels inevitable, what feels fake, and what feels accidentally true.
Bring your best “formal mischief.” We’ll bring logs.