Notes from building with agents: what held up, what broke, and what became useful enough to keep.
Articles
View:
Make Your Work Visible With a Simple Map
A practical way to show complex work honestly: map the pieces, label maturity, hide the private machinery, and point to the next useful path.
Small Automation Needs a Human Boundary
A practical boundary test for tiny workflow tools: what they may touch, what approval already exists, what proof they show, and where they must stop.
Stop Learning AI. Give It One Job.
A practical way for small businesses to cross the first AI adoption hump: assign AI one bounded task before trying to automate the whole operation.
Small Gates Keep AI Work Honest
A practical way to keep AI-assisted work fast without letting direction, proof, and publish decisions blur together.
Use AI Assistants Without Outsourcing Your Judgment
A practical guide to using AI assistants for leverage without letting fluent output replace source truth, privacy, verification, or human judgment.
Fix the Real Problem Before You Polish
A practical reminder to solve the actual cause before making the visible mess look better.
If You Keep Repeating the Prompt, Build the Agent
A practical case for creating custom agents when you keep re-teaching AI tools the same standards, checks, and judgment calls.
Make Writing Part of the Work
A practical case for treating writing as part of making work reusable, reviewable, and easier to improve.
The Work Behind the Handle
A work-first orientation to RomeoRaven: what lives here, why it is public, and where the useful boundaries are.