Chapter 11. Obsidian × AI — compounding what you know

Key points

• Accumulate safely on your machine, and cut out an anonymised extract when you go to an AI
• Build the loop: read → note → connect → structure with AI → return
• Aim at a state where knowledge compounds the more you learn

The idea of cutting out an extract

As knowledge accumulates in Obsidian, you will naturally reach the point of wanting a group of notes organised by an AI.

Here the Chapter 2 idea pays off.

Do not show the AI everything; cut out the part that may be shown, and hand over that.

• the knowledge base itself (containing personal information and unpublished ideas) → kept safely inside Obsidian, locally
• the part you want to consult about → copy an anonymised extract only, into ChatGPT or Claude

Make that cut a habit and you can use AI's full strength while keeping safety intact.

The loop that compounds knowledge

Combine Obsidian with cloud AI and a loop starts turning.

1. Read — papers, books, articles.
2. Note — write the key points into Obsidian, in Markdown.
3. Connect — link related notes with [[ ]].
4. Structure — once enough has accumulated, hand several anonymised notes to an AI together and ask "lay out the themes these share", "structure what can be said from these".
5. Return — put the AI's output back into Obsidian as a new note.

Once the loop turns, you reach a state where

the more you learn, the more knowledge compounds

Scattered knowledge connects, becomes structured, and generates new questions. That is a powerful base for a therapist's clinical reasoning and research alike.

Concrete uses for a therapist

• Literature review — a note per paper, connected by theme, then read across by AI
• Clinical reflection — accumulate anonymised observations and have the AI find the patterns
• A notebook of material to publish — daily notes become the source of blog drafts
• A store of teaching material — generate study-group material from accumulated notes each time

What used to be processed in your head goes outside and grows. AI tills that ground with you. This is one of the destinations of working with local files.

A prompt to try

You are an editor who knows intellectual production well. Below are several notes I have accumulated in Obsidian (anonymised). Organise the themes they share into three, and for each propose one question that could be developed from it.

[Notes] (paste the anonymised notes here)

Chapter 11 summary

• Keep the knowledge base local and safe; hand AI only anonymised extracts
• Build the loop: read → note → connect → structure with AI → return
• That loop compounds knowledge, and becomes the base for clinical work, research and publishing