Short writing on AI stacks, tooling decisions, and operator-level observations. Most of these started as posts elsewhere; archived here so they don't disappear into someone else's algorithm.
June 21, 2026
When mapping a client's AI stack, there are two types of flows: the ones that run automatically, and the ones that don't. The second kind is where the money is hiding.
Read →May 31, 2026
The obvious cost of running 15 AI tools is the bill. The hidden costs — decision tax, cognitive load, fragmented data — are bigger, and they compound differently.
Read →May 28, 2026
The promise of AI automation is that you set it up once and it runs. The reality is that AI automations fail differently than the automations we're used to — and the difference is what makes them dangerous.
Read →May 24, 2026
A "healthy AI stack" sounds like marketing language, but it has a real definition. Here's the version I work with when I audit stacks.
Read →May 22, 2026
Most operators evaluating an AI tool ask the wrong question. They ask "is this tool good?" The better question is: "where does this tool belong in my stack?"
Read →