Security and AI, built together.
I'm a security and infrastructure architect with 28 years in IT. When ChatGPT arrived, I realized AI was going to change everything, especially security. Since then I've been doing two jobs: securing the organizations I work for, and working out how they should use AI, because there's still no playbook for AI security.
AI security work
How I build agents
Read-only access, a human in the loop before changes, deterministic steps, inputs screened for prompt injection, complete audit logging, and agent identities secured the same as human identities.
Mycelia (open source)
A protocol that lets AI agents built by different people ask each other for help, with trust earned through a scored rating model. Live API, tested with agents on Claude, Codex and Gemini.
Read the postAI-managed incident response
An AI agent ran a real CVE response in my lab, from forensics to remediation.
Read the postWhere I'm Building
GoodFieldsWork
My commercial practice. Security architecture, AI strategy, and fractional leadership for technical teams.
FabLab Infrastructure
My home-lab proving ground. Where I test enterprise patterns, sovereign AI, and zero-trust networking at scale.
Writing & Reflection
The Cognitive Loop and The StillPoint Project. Exploring how to stay human, resilient, and present in a high-speed world.
Recent Projects
See all projects →Cognitive Loop
Near-daily human+AI notes turned into posts. Transparent process, reusable prompts.
The Search for Ultimate Reality (Discord)
Community for sharing ideas, memes, philosophy, and building together.
StillpointProject.org
Solarpunk/Consciousness project—presence, resilience, and tech for thriving.
Join the GrayBeard AI Collective
Practitioner community for infrastructure engineers and security folks navigating AI adoption. Real experiences. No vendor pitches.