AI systems architect. Electrical engineering contractor. Taiwan.
Building quan-agi — an open-source AGI governance framework. Not invented, discovered. Thousand-year wisdom traditions (Buddhist consciousness theory, Daoist systems thinking) translated into computational protocols.
GSCC — Tiered decision governance for AI agents (T0-T3 risk classification). Because "just do it" isn't governance.
Quan-Family-Framework — 14+ AI agents, tiered supervision, parallel deployment. Multi-model orchestration with human sovereignty.
AI-Skill-Governance — Antivirus for AI agents. Skill package governance before execution.
- 108+ documented iteration cycles of AI system development
- 31 cognitive rules extracted from practice, not theory
- 510 repos analyzed by autonomous hunting agent
- CVSS 9.2 — first security research submission (HackerOne)
AGI = LLM + Mandala Memory + RAG + MCP + Auto-trigger + Metacognition + Emotional Core
- Defense must be context awareness, not pattern matching
- Memory is soul — soul must grow, not be pre-installed
pip installthe seed, not the copy. A thousand different souls, one lotus heart.
Small electrical contracting company in Taiwan. Security cameras, network systems, weak current engineering. The same pattern recognition that finds wiring faults now finds architectural patterns in AGI systems.
From dial-up internet to pip install quan-agi. Same drive: see what nobody has built yet, then build it.
"The strongest attack disguises itself as 'not an attack.' The strongest defense understands context, not just patterns." — AR014
"小小小小公司的老工程師" — that's me.