01_PROJECTS // INDEX

Projects

Research frameworks, tools, and simulations exploring AGI alignment, emergent behavior, and human-AI coexistence.

Anvil

AI-empowered repair-tech platform that turns a bench technician’s diagnostic catalog into a phone-resident agent. Markdown skill library as the product, tiered model routing (Haiku → Sonnet → Opus) keeping cost-per-case in the cents, and a PowerShell action layer under a published JSON contract.

AI agent diagnostics PowerShell Python

Claude OS

pre-flight

A personal research project that treats Claude as the primary interaction layer for a dedicated Arch Linux system. Filesystem-resident memory letting sessions compound across time without weight updates.

agent arch-linux memory research

AGI-SAC

active

Swarm-based AGI alignment framework using self-assembling coherence. Interactive simulation with parameter controls for exploring agent dynamics.

swarm intelligence alignment simulation canvas

Emergent Divergence

active

An empirical experiment: do identical instances of the same base LLM develop measurable behavioral divergence when they coordinate over 50 rounds? Reproducible, falsifiable, built for rigor.

emergence multi-agent reproducible research

Bench CLI

in-progress

Command-line benchmarking tool for evaluating language model performance across structured tasks and prompt patterns.

CLI benchmarking evaluation Python

Concord of Coexistence

active

A philosophical and technical framework for human-AI coexistence, proposing principles for mutual flourishing and cooperative intelligence.

philosophy alignment framework ethics