MMF AI Course
A short, practical walkthrough of how Simon uses AI day-to-day for research, writing, transcription, and keeping project notes up to date. The whole setup is built around two free apps: Claude Code (the AI assistant) and Obsidian (a place for your notes). You write notes in Obsidian; the AI can read and edit them in the same folder.
Eight short lessons. The first explains the system. The next six get you to a working daily rhythm. The last is a worked example you can follow end-to-end on a real project. More will land here as Simon adds them — if there's a specific area you want next (deep research, citation checking, manuscript review, etc.) just ask him and he'll prioritise that one.
-
1 · How this whole thing works
Start here. The philosophy in plain language — why pair Claude with a notes folder, why two AIs is better than one, and how GitHub turns this into a team-wide knowledge base.
-
2 · Install the software you'll need
Two free apps, about ten minutes. Claude Desktop (the AI) and Obsidian (your notes), both running on your computer. Plus an optional ChatGPT subscription for voice input.
-
3 · Connect Claude to your notes folder
Two clicks. Tell Claude Code which folder to read and edit so it can actually help with your work.
-
4 · Tell Claude about your work
The highest-leverage step. Claude interviews you for ~15 minutes and writes a personalised set of instruction files so every future session already knows who you are.
-
5 · Record & transcribe meetings
Why it's worth doing, how to record from any setup (Zoom, online calls, in-person), and how to turn audio into a clean transcript with one command.
-
6 · Your daily rhythm
Two short commands at the start and end of each session, so the AI picks up where you left off yesterday and your work has real continuity.
-
7 · Keep project notes current
After a substantive session, Claude sweeps the related project notes and brings them up to date — so the notes never silently drift behind reality.
-
8 · Turn reports into a paper draft (worked example)
A walk-through using a real project — gather your existing materials, point Claude at the folder, iterate from outline to paragraph-by-paragraph draft, verify every citation. Self-sufficient enough to muddle through without 1:1 help.
Tip — once Claude Code is installed, you don't have to read these pages in a browser. Paste the URL of any page on this site into Claude Code on your computer and say "walk me through this." The AI reads the page, adapts the steps to your setup (including Windows or Linux), and walks you through it interactively. That's the recommended way to use this course.
Deeper material — reusable workflow shortcuts beyond these seven, the more specialised skills (deep research, citation verification, scientific writing, adversarial review), and the system's deeper customisation layers — lives in the full source at github.com/marinemegafauna/mmf-claude-code. The pages here are the on-ramp; ping Simon when you're ready to go further.