Less a set of lessons, more a workbench. The runbooks, prompting moves, and habits that have actually given me good results — written so you can copy a prompt, paste it into your own tool, and try it today.
This whole module is opinion. These are my habits — the moves I reach for, not a verified best practice or a standard. Some of it is me gently nudging you to try a technique I think is underrated. Take what's useful, ignore what isn't, and trust your own judgment over mine. Tools and prices change fast; where I name a specific tool, I link to Tools I Use so you can check current pricing there.
A runbook is a recipe you can run again. Three start-to-finish workflows — build a web page, turn a newsletter into a dashboard, automate a Word report — with the real prompts to paste at each step.
Three small habits that change the quality of an answer: tell it about you, ask in Markdown, and let the model help write the prompt. See each one side by side, before and after.
A conversation has a memory budget, and answers get worse as it fills. Watch the meter fill, learn when to start fresh, and grab a checkpoint prompt that carries your progress into a new chat.
My most-repeated habit: whenever you make something — a report, a deck, a spreadsheet — also make the version an AI can read. Canonical formats like Markdown and YAML are how you keep extending and refreshing your work.
It looks like a programmer's tool, but it's one of the best general-purpose assistants there is. A friendly, no-jargon walkthrough to get it running on Windows or macOS — terminal and all.
Anything you produce for a person, also produce in a form another AI can read and extend. That's what makes work compound instead of starting from zero every time. Get in the habit of asking three questions about everything you make:
How else can I use this data?
Where can I get this data?
How do I refresh this data?