Building
C2
A governed operating system for an AI-native company, with durable context and explicit human authority.
Now · August 2026
What has my attention right now.
A concise, dated snapshot of active builds, research questions, and the systems I am trying to understand more clearly.
This page changes when the work changes—not on a fixed content schedule.
Researching
Edge AI security
How efficient local detection and larger central reasoning can share the right evidence.
Testing
Agent reliability
Where evaluation, observability, permissions, and recovery belong in an agent architecture.
Publishing
A connected public record
Linking questions, writing, experiments, projects, and ventures as the same ideas evolve.