Network Infrastructure Specialist · Independent Systems Builder · Experimental Researcher
Local-first computing · Human–AI collaboration · Experimental research · Creative tools · Playable worlds
REALITY IS THE CUSTOMER. | CAPABILITY ≠ AUTHORITY. | KEEP THE EVIDENCE ATTACHED.
Explore the public Field Index · Research archive on Zenodo · Enter the Field
I started with real-world network infrastructure: equipment, connections, deployment constraints, troubleshooting, and the fact that eventually a system has to work outside a diagram. Today I apply that field perspective to local computing, multi-model collaboration, research instruments, accessible creative software, and games.
This is not a single giant product. It is a growing family of original, independently developed tools and experiments that share a practical design principle:
Make capability useful, keep evidence inspectable, and leave consequential decisions with people.
The public repositories include working prototypes, research protocols, software under active development, and speculative designs. A passing build is not a production certification; an experimental result is not a universal scientific finding. Each repository describes its own evidence and limitations.
| Explore | Start here |
|---|---|
| Operating systems & useful software | PhiOS · Trellis · OpenBlue · PhiOffice369 |
| Local compute & networks | Infinite Porch · PorchTalk · PhiStream · BudgetGenius |
| Computational collaborators & governance | Super Φ.Vessel · Think Tank · Field Accord · BrainC |
| Research & evidence | Nested Bubble Gear · CollaborationJunction · FieldManual369 · Research Lab |
| Art, music & making | Domistika · Auralith369 · PixelForge · Silicon Louvre · More Bounce Labs |
| Games & immersive worlds | PhiCade · Pulse369 · HomeBass · NightCircuit |
Browse the full public repository index →
The index is organized by area and periodically refreshed from public GitHub repositories. It deliberately excludes private work.
| Project | What it explores | Evidence boundary |
|---|---|---|
| PhiOS | Local-first operator shell, governed tools, memory, and application platform | Alpha; OS release and physical qualification are separate |
| Trellis | Deterministic package management and explicit trust receipts | Release candidate, not a hosted package marketplace |
| Infinite Porch | Consent-based peer networking, local models, and bounded sharing | Native qualification candidate; not production-qualified federation |
| PorchTalk | Local-first, budget-aware selection between consenting compute offers | Offline routing-policy prototype, not real remote inference |
| PhiStream | Memory-tiered inference with measured data movement and parity receipts | Experimental toy decoder; not a real large-checkpoint speedup claim |
| BudgetGenius | Cognitive resource budgets, route qualification, cost and quality audit | Architecture and verification work, not guaranteed savings |
| Super Φ.Vessel | Multi-model roles, routing, evidence, local runtime interfaces, and human release boundaries | Experimental platform; individual surfaces have different release states |
| Field Accord | Human–agent coordination contracts with explicit authority and provenance | Governed protocol development |
| OpenBlue | Local-first blueprint, schematic, and network planning tools | Active prototype; not a certified CAD suite |
These projects connect conceptually, but a link on this page does not imply a completed integration between them.
Nested Bubble Gear / Altermath investigates when a coarse observation hides internal structure that still matters causally. The research program uses finite models, control experiments, explicit falsifiers, replayable artifacts, and a claim firewall. It does not assert a new physical law merely because a constructed experiment passes.
CollaborationJunction examines human–AI collaboration and where AI is actually present in existing technology. Its evidence atlas separates sourced descriptions from independent tests; proposed human studies remain proposals until run.
FieldManual369 turns systems thinking, evidence, design, governance, and cybersecurity into an interactive learning environment.
Φ369 Research Lab on Zenodo houses versioned public research records and protocols. Negative results are part of the research, not something to hide.
| Studio / world | Explore |
|---|---|
| Domistika | Procedural drawing, motion, symmetry, composition, and governed creative handoffs |
| Auralith369 | Visual treatment and image-processing experiments |
| PixelForge | Retro game creation, runtime contracts, and agent-control research |
| PhiCade | Human, script, and AI controllers under one bounded game-runtime contract |
| Pulse 369 | Browser/WebXR rhythm fitness and immersive practice |
| More Bounce Labs | Music catalog, radio, DJ decks, and creator experiments |
| The Silicon Louvre | Optical art, perception, and human–AI creative provenance |
| The MDF Files | A fictional network-technician horror/comedy archive, because field tickets deserve their own genre |
Not everything here is an agent framework or a theoretical experiment. Some projects focus on practical needs:
- GLO Pin — an accountless, local-first way to record and explore community resources.
- FieldDeck — a human-reviewed action dashboard with separate, permissioned execution workflows.
- FieldAtlas — a visual doorway into the growing collection of sites and tools.
- FamilyVault — family-centered software experimentation.
- Spark: The Substrate — a public playable-world experiment (the link is to the game, not a public source repository).
Need or question → observable definition → smallest useful implementation → tests and counterexamples → evidence receipt → review → human decision → revise.
- Evidence over confidence. Label what was observed, inferred, simulated, or left unknown.
- Capability is not permission. Tools, models, and agents operate within explicit authority boundaries.
- Local-first where practical. Prefer inspectable, user-owned capability over unnecessary dependencies.
- Original implementations. Study ideas; design and write our own systems rather than silently importing someone else’s work.
- Design for real humans. Accessibility, affordability, curiosity, play, and maintainability matter.
My toolbox spans network fieldwork, TypeScript/React/Next.js, Python, Rust, local inference with Ollama, GitHub Actions, visual tooling, audio, and game systems. The stack is less important than being able to explain what a system really did.
GOOD CODE · GOOD VIBES · BETTER WORLD
Complete public Field Index · Enter the Field · Research on Zenodo