ØVOID — experimental software lab

LIVE DRAFT · mind the loose wires

Enter the lab (mind the loose wires). Refresh if you want. Or don’t. Either way, the page has opinions. This is a draft of a draft, and it knows it. Poke around. If it changes, it’s working. The future is being written. Please excuse the edits. Enter the lab (mind the loose wires). Refresh if you want. Or don’t. Either way, the page has opinions. This is a draft of a draft, and it knows it. Poke around. If it changes, it’s working. The future is being written. Please excuse the edits.

We build software that refuses to pretend it’s finished.

ØVOID is a two-person experimental software lab that treats software like a living draft. We choose experimentation over safe choices, then document the scorch marks. AI makes software cheap; taste, risk, and weirdness become the differentiators.

Philosophy: beautiful chaos

not a bug report → design material

AI makes software a commodity; taste, risk, and weirdness become the differentiators.
We choose experimentation over safe choices, then document the scorch marks.
Beautiful chaos is not a bug report—it's a design material.
We don’t ship “finished.” We ship “interesting,” then iterate with the wind.
Unpredictability is a feature when you’re co-creating with machines.
The future is being written, and sometimes it misspells itself—on purpose.

Current project

AI-native · create / play / judge

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

AI-native gaming platform

A platform where games are created, played, and judged by AI—so the medium can iterate at the speed of drafts. The judge isn’t a scoreboard; it’s a character with taste. The games aren’t “content”; they’re weather.

Create Games are born unfinished on purpose.

Play Rules flex; the experience keeps notes.

Judge Stern, but fair-ish. Sometimes petty. Sometimes mythic.

What makes it ØVOID? taste · risk · iteration

We’re not trying to “solve” games. We’re building a system that can keep rewriting them mid-sentence—then admitting what changed, what broke, and what accidentally became better. Every session is a draft review with a non-human co-author.

Lab notes (selected scorch marks)

no tour guide. just corridors.

Brand voice, briefly stabilized kept it

We once tried to “stabilize the brand voice.” The site responded by turning every headline into a haiku for six minutes. We kept it. It felt honest.

The judge dislikes politeness drama score

During a late-night test of AI-native gaming platform where games are created, played, and judged by AI, an AI judge declared a game “too polite” and docked points for lack of drama. The next build added a weather system that only storms during plot twists.

The “Do Not Trust” button probably

We keep a “Do Not Trust” button in our internal tools. Someone clicked it and the UI started labeling everything “probably.” It was the most accurate week we’ve had.

Sitemap → terrain north

A prototype menu insisted on being a map. Clicking “About” sent you to “North,” and somehow that made more sense than it should. We renamed the sitemap to “terrain.”

Artifacts pinned to the corkboard

evidence of intention

A crumpled index card labeled “Chaos Budget,” with three checkboxes: Delight / Confusion / Revelation. image idea
“Judge Mood Dial” sketch: Calm → Petty → Mythic. Underlined twice. image idea
Printed error message taped to a monitor: “Outcome unavailable: the future is in another tab.” image idea

A tiny changelog entry: “Replaced certainty with curiosity. Sorry.” image idea
A screenshot fragment titled “Version 0.0.0-ish,” annotated: “Ship it, then learn what it is.” image idea
A napkin diagram: three arrows forming a loop—Create → Play → Judge—captioned “AI-native gaming platform where games are created, played, and judged by AI eats its own drafts.” loop logic

Telemetry (semi-trustworthy)

the page has opinions

Drafts per idea

7–19

Depends on moon, coffee, and how brave the compiler feels.

Chaos coefficient

0.62

1.00 is “beautiful”; 0.00 is “corporate”.

Unpredictability window

3–11 seconds / minute

Measured in “wait, what?” per minute.

AI judge strictness (AI-native gaming platform where games are created, played, and judged by AI)

stern, but fair-ish

The dial above is… loosely calibrated.

Safe-choice avoidance rate

84%

Remaining 16% are accidental.

Find us where drafts live

public-ish chaos

Want to track experiments, prototypes, and the occasional meaningful mistake? Follow the repository trail.

GitHub code / notes / receipts Back to top mind the loose wires

A small promise soft rules

We’ll keep things legible while they’re still strange. We’ll ship “interesting,” then iterate with the wind. If something looks unfinished, it probably is—on purpose.