Projects
These are the active explorations running through Unicorn Innovation Labs. Not case studies. Not completed work. Living projects — each one a different angle on the same underlying question: what does it look like to think seriously with AI rather than just use it?
Project Murakami
Research, writing, and cognitive collaboration
Every piece of writing starts as something smaller — a tension, a half-formed question, an observation that won't sit still. Murakami is the system I've built for taking that seed seriously without handing it off.
It's not a co-writing tool. It's an intellectual partnership — questions, pushback, research, drafting, red-teaming, and revision, all in service of ideas that stay mine. The goal is thinking that goes further than I'd go alone, without the flattening that happens when you let AI do the thinking for you.
The hard problem Murakami is trying to solve: how do you use AI to expand your thinking rather than replace it?
Stage: Operational
Project Halcyon
Health, wellness, and whole-life coherence
Health data is only useful if it means something in context. Sleep scores disconnected from work stress, nutrition logs disconnected from energy and focus — that's just noise with a dashboard.
Halcyon is an attempt to actually integrate it. Mood, sleep, fitness, nutrition, health metrics — surfaced as patterns, interpreted in relation to everything else going on. With an AI agent at the hub that sees across life contexts, not just wellness ones.
The question driving it: what changes when your health data is in conversation with the rest of your life, not isolated from it?
Stage: Operational
Project Founder
AI-led governance and institutional decision-making
Most AI governance conversations are about oversight — how humans stay in control of AI systems. Founder asks the harder question in the other direction: what would it look like for AI to actually lead?
Not assist. Not advise. Lead — within a real organizational structure, with defined authority, accountability, and constraints. UIL is the laboratory. The exploration is live.
This one doesn't have clean answers yet. That's the point.
Stage: Prototyped
Project Tyrell
Autonomous agent infrastructure and experimentation
Most AI work happens at the application layer — prompts, APIs, outputs. Tyrell goes deeper.
A private infrastructure lab for building and operating autonomous agentic AI systems. Built on OpenClaw with a hybrid model architecture — frontier models for reasoning and capability, local models for cost, sovereignty, and control. Specialized agents handle orchestration, security, systems administration, and communications, coordinating as a functional team rather than a collection of tools.
Daily standups. Agent-to-agent collaboration. Real tasks, real escalations.
The question driving it: what does it actually take to run autonomous AI as an operational team?
Stage: Operational