I Turned Arch Linux Into an AI Engineering Organization
One developer. A persistent herd of agents. An entirely different way to build software.
Technical field notes on AI-native engineering, agent systems, software architecture, security and human-machine communication.
One developer. A persistent herd of agents. An entirely different way to build software.
Running 100 AI agents is a distributed-systems problem, not a chatbot-management problem.
Talking to machines exposes how vague humans are—and may be training us to communicate better.
AI generates the code. The engineer reviews, questions, repairs—and becomes a student again.
Autonomous coding agents can look a lot like an intrusion from the outside. That changes the security question.