Start here

I build agentic AI systems for institutions where autonomous failure is a legal event, then publish the review. There is a lot here; pick the door that matches why you came.

You build agents

You are shipping agentic systems and want the parts the demos skip.

Evaluating Agentic Systems the four-axis method: cost per case, accuracy, tool-skip rates, temperature Tool Contracts why agents skip your tools, and the spec rules that fix it

You govern them

You own the risk, the audit, or the architecture decision.

The Control Plane inner loop vs outer loop, and the five obligations frameworks cannot own Where Agents Actually Work field notes from 50+ enterprise builds, by industry and workflow shape

You want the protocols

You are here for the wire: how agents talk to tools, each other, humans, and money.

A2A agent-to-agent: cards, task lifecycle, orchestration vs swarm AG-UI streaming the interface, not just the answer AP2 agent payments: signed mandates, not trust Agents, Orchestrated the series: each protocol built for real, reviewed in public

You run a business or non-profit

You are not here for protocol internals. You want a straight answer about AI for your organization.

Let me review your AI the questions worth asking any vendor, and an open door 🍁

Prefer to browse? Everything lives in the archive, and the person behind it is on the about page.