📊 View Lecture Slides Full-screen presentation with navigation

Session 3: The Personal Exocortex

Session Duration: 2 Hours     Block: 1 — Foundations & The Exocortex

Learning Objectives

By the end of this session, students will be able to:

  • Define the concept of an exocortex and explain its role in AI-assisted systems
  • Set up and structure an Obsidian vault using Markdown conventions
  • Create Maps of Content (MOCs) to organise knowledge hierarchically
  • Explain why structured local knowledge is the foundation of a RAG system

Hour 1: Externalising Memory (Instructor-Led — 60 minutes)

1.1 What Is an Exocortex?

The term “exocortex” refers to an external cognitive system — a digital extension of your own memory and reasoning capacity. The idea is simple: your brain has limited working memory and is unreliable at long-term storage. An exocortex offloads this work to an external system you design and control.

For AI product engineers, the exocortex serves a second critical function: it becomes the knowledge layer your AI systems query. Instead of asking an AI model to answer from its static training data (which has a knowledge cutoff and may hallucinate), you ask it to answer from your curated, up-to-date knowledge base.

This is the foundation of Retrieval-Augmented Generation (RAG) — the subject of Sessions 8 and 9.

1.2 Why Obsidian?

Obsidian stores all notes as plain .md (Markdown) files in a local folder (called a “vault”). There is no proprietary database, no cloud lock-in, and no special format. This has three important consequences:

  1. AI-ready: Markdown is exactly the format AI models work best with. No conversion required.
  2. Portable: You can copy, search, process, and embed your vault with any tool.
  3. Durable: In 20 years, your notes are still readable plain text files.

No other mainstream note-taking application shares this combination.

1.3 The Anatomy of a Good Obsidian Vault

A well-structured vault has four components:

Atomic Notes: Each note contains one clear idea. Title is a complete sentence or clear concept. Body is a self-contained explanation.

# Why Context Window Size Matters for RAG

A context window defines how much text a language model can process at once.
In a RAG system, the retrieved documents must fit within the context window
alongside the user's query. Larger windows allow more retrieved content,
but increase cost and processing time.

Links: Use [[double bracket notation]] to link related notes. This creates a knowledge graph.

Tags: Use #tags to create categories that cut across the folder structure.

Maps of Content (MOCs): Index notes that serve as navigational hubs. A MOC for “AI Applications” might link to notes on RAG, agents, embeddings, and evaluation.

1.4 Folder Structure for This Course

vault/
├── 00-Inbox/           # Quick captures, not yet organised
├── 01-Concepts/        # Theory: what things are
│   ├── AI-Foundations/
│   ├── RAG/
│   └── Agents/
├── 02-How-To/          # Practical: how to do things
│   ├── Prompting/
│   ├── APIs/
│   └── Deployment/
├── 03-Project/         # Your AI product
│   ├── Product-Spec.md
│   ├── Prompts.md
│   └── Architecture.md
├── 04-References/      # External sources, papers, docs
└── MOC-Course.md       # Master index of this vault

Hour 2: Practical — Build Your Course Vault (60 minutes)

Lab 3.1 — Install and Configure Obsidian

  1. Download Obsidian from obsidian.md — free, available for all platforms
  2. Create a new vault in a dedicated folder (e.g., Documents/ai-product-vault/)
  3. Create the folder structure above
  4. Enable the following core plugins: Backlinks, Graph view, Templates

Lab 3.2 — Create Your First Atomic Notes

Write three atomic notes from this session:

Note 1: What is an exocortex? Note 2: Why Markdown is the ideal AI knowledge format Note 3: What is a Map of Content?

Each note should be 3–5 sentences. Link them to each other using [[double brackets]].

Lab 3.3 — Create Your Project MOC

Create a file called 03-Project/MOC-AI-Product.md. This becomes your running project index.

# MOC: AI Product Engineering Project

## Product Idea
- Product: [[Product-Spec]]
- Problem statement: [your problem]
- Target user: [your user]

## Knowledge Base
- [[Domain-Background]]
- [[Key-Concepts]]

## Session Notes
- [[Session-1-Notes]]
- [[Session-2-Notes]]
- [[Session-3-Notes]]

Add to this MOC after every session. By Session 15, it will be the map of your entire project.

Lab 3.4 — Import Your First External Content

Find one external resource relevant to your project (an article, a documentation page, or a paper). Create a note that:

  • Summarises the key point in 2–3 sentences (your own words)
  • Includes the source URL
  • Links to at least one other note in your vault

Key Takeaways

  • An exocortex is a structured external memory system — the knowledge layer your AI will query
  • Obsidian stores notes as plain Markdown files, making them directly AI-ready
  • Atomic notes, links, tags, and MOCs are the four structural elements of an effective vault
  • Your vault is the raw material for the RAG pipeline you will build in Sessions 8–9

Further Reading

  • Obsidian documentation: help.obsidian.md
  • “Building a Second Brain” (Forte, 2022) — foundational personal knowledge management theory
  • Zettelkasten method: zettelkasten.de