Harnesses & tools
Placeholder—the tools I'm currently comparing and where each one fits: herdr, buzz, Cloudflare's agent tooling, Claude managed agents, Azure's agentic containers and sandboxes. Which harness for which job, and why.
Agentic AI
Where I figure out my agentic AI setup in public: which harnesses and tools for which job, which models are worth their cost, and how to deploy and orchestrate agents—solo and in a team.
In my setup
What I’m figuring out as I build my agentic AI setup. Placeholder entries for now—this fills in as I learn.
Models: cost vs. value
Placeholder—which model I use for which task. Where a frontier model is worth its price and where a cheaper or local model is plenty.
Deployment
Placeholder—how I run agents: locally, hybrid, or on a server. Trade-offs in cost, privacy, reliability, and setup effort.
Orchestration & teamwork
Placeholder—how agents work together, and how they fit into a team or organization: handoffs, shared context, review, and ownership.