Current project: AI-native gaming platform
What “AI-native” means here
behavior is the interfaceWe’re building a platform where the loop isn’t just humans playing content— it’s AIs generating games, AIs playing them, and AIs judging them. The result is a living catalog of “first drafts” that argue back.
- Create Games emerge from prompts, constraints, and evolving rulesets—less “final design,” more “replayable hypothesis.”
- Play AI agents explore the game space, discover exploits, and surface weird edges that humans might miss.
- Judge Scoring focuses on novelty, coherence, and “is this fun in a way we didn’t expect?”
Inspiration labels (not images): image://ai-judge-gavel-made-of-pixels, image://first-draft-game-screenshots-collage
Every game is a first draft. The fun part is watching the draft argue back.
How to read our progress
prototypes > promisesWe don’t claim certainty. We ship experiments and document what they teach us. If you want the most concrete trail, follow commits and notes.
Philosophy: experimentation over safe choices
Manifesto (read like margin notes)
Page 1 — the premise
- 1AI makes software a commodity; weirdness makes it memorable.
- 2We choose experimentation over safe choices, even when it squeaks.
- 3Beautiful chaos is not a phase—it’s a design principle.
Page 2 — how we work (and why it looks like this)
- 4Prototypes are allowed to be loud, unfinished, and oddly charming.
- 5If it behaves perfectly, we probably fixed the wrong thing.
- 6The future is being written. We keep the cursor blinking.
Beautiful chaos: the moment a prototype stops obeying and starts collaborating.
A practical promise
We won’t hide behind vagueness. Our work should be legible: what we’re exploring, what we learned, and what we’re trying next—without pretending the messy parts aren’t part of the craft.
“Safe choices are just defaults wearing nicer clothes.”
Lab notes (the margins are part of the page)
Inventory: artifacts we actually trust
- ✶A laminated “Chaos Budget” card: 70% curiosity, 20% restraint, 10% snacks.
- ✶A tiny JSON file named ritual.json containing: {"breathe":2,"ship":1,"overthink":0}.
- ✶A desk bell labeled “Reroute the Plan.” It rings when something unexpected works.
- ✶A commit message fragment: “teach the system to misbehave politely.”
- ✶A coffee-stained map of features with a legend: “Here be delightful mistakes.”
Inspiration labels (not images): image://neon-lab-notebook-with-glitch-margins, image://chaos-spiral-diagram-on-whiteboard
Telemetry (measured in eyebrow raises)
Anecdotes (the kind that explain the UI without excusing it)
- #We tested a color palette called “Slightly Haunted Neon.” It looked wrong in daylight and perfect at 2:13 AM. We kept it and named the bug “mood lighting.”
- #A sticky note appeared on the monitor: “Stop polishing the chaos; let it glitter.” Neither of us claims authorship. We treat it as an office visitor.
- #During a late-night build, we set a rule: every button must feel like it’s about to become a different button. The next morning, we couldn’t find the “Contact” link—turns out it had evolved into “Confess.”
Contact (we respond like humans)
Email: hello@ovoid.lab
If you’re exploring AI-native games, weird tooling, or novelty-driven systems—send a note. We’ll read it, not route it.