Journal · field notes & essays

Notes from the practice.

Short, practical writing for the people I work with, owner-operators and their teams. No hot takes. Nothing ghostwritten. Just what I'd tell you over coffee, if we had the time.

01 · Essay · 9 min read · Apr 2026
When to use AI, and when to use conventional code. A plain-English field guide for owner-operators who are being sold the same thing by everyone.
Read the essay
02 · Field note · 6 min read · Apr 2026
What "handoff-ready" actually means. A one-day test your vendor should pass, or fail loudly, in writing.
Read the field note
03 · Essay · coming soon
Stewardship of attention. Why your ad account should be run like a fiduciary, not a funnel.
Draft in progress
04 · Essay · 8 min read · Apr 2026
Why your Claude outputs are inconsistent. It's not the model. It's not your prompts. It's the layer underneath — the one most teams never build.
Read the essay
05 · Essay · 7 min read · Apr 2026
The vending machine trap. There are two ways companies waste their AI subscriptions. Most teams are stuck in one of them right now.
Read the essay
06 · Industry · 6 min read · Apr 2026
The associate's invisible tax. Every matter, associates re-explain the client, the case, and the firm's conventions to Claude. That's billable time paying for context setup.
Read the essay
07 · Industry · 5 min read · Apr 2026
Why real estate AI sounds like every other brokerage. Local expertise doesn't make it into the output because nobody told the tool it was there.
Read the essay
08 · Industry · 6 min read · Apr 2026
Grant writing is language work. Treat it that way. The science is done. Why is the PI spending 40 hours on narrative sections that describe work that's already finished?
Read the essay
09 · Industry · 5 min read · Apr 2026
AI makes independent retailers sound like chains. Generic AI output actively erodes the voice that differentiates you — unless the system knows who you are first.
Read the essay
10 · Industry · 6 min read · Apr 2026
Mission-driven language is precise language. In ministry work, the wrong word isn't just generic — it misrepresents the mission. Generic AI output is a liability here.
Read the essay
11 · Industry · 6 min read · Apr 2026
Your product documentation reads like an engineer wrote it. Software teams produce more language-heavy work than almost anyone — and deprioritize it more than anyone too.
Read the essay
12 · Essay · 12 min read · May 2026
How I run my entire business with AI. Nine clients, one person, and a system built on plain-text files, scheduled routines, and an AI superintendent named Gabe.
Read the essay
Modern tools. Craftsman's word.

Want to talk through your own situation?