The Three Core Cards: Your Knowledge Hierarchy
Overview
The AI2U Index Cards system uses a powerful three-tier hierarchy to organize your knowledge efficiently. This structure mirrors how our minds naturally organize information: from big concepts to specific details to granular examples.
Why Three Levels?
- Cognitive Load Management: Three levels match our brain's chunking capacity
- Search Optimization: Main and Sub cards appear in searches, Nano cards don't clutter results
- Context Efficiency: Relevant details (Nano cards) automatically included with their parents
- Flexible Depth: Enough levels for complexity without overwhelming navigation
Main Cards (M-XXXXXX)
Purpose
Main cards are your primary knowledge containers - the big ideas, major topics, and overarching concepts that form the foundation of your knowledge base.
Characteristics
- ID Format: M-XXXXXX (e.g., M-7B9K2L)
- Size: 4�6 (largest card size)
- Color: Blue theme in UI
- Parent: None (top-level cards)
- Children: Can have multiple Sub cards
Features
- Always searchable: Appear in all card searches
- Standard scoring: Get full relevance scoring in chat context
- Tag inheritance: Tags flow down to child cards
- Graph view: Form the primary nodes in visualization
- No parent required: Independent, standalone units
When to Create a Main Card
- Starting a new project or topic
- Documenting a major concept or theory
- Creating a category or collection
- Building a learning curriculum
- Organizing a complex system
Examples
M-PROJECT1: "SaaS Product Launch"
M-LEARN01: "Python Programming"
M-CONCEPT: "Quantum Computing Basics"
M-SYSTEM1: "Company Knowledge Base"
Sub Cards (S-XXXXXX)
Purpose
Sub cards break down Main cards into actionable pieces - the specific tasks, detailed explanations, and focused topics that make big ideas manageable.
Characteristics
- ID Format: S-XXXXXX (e.g., S-9M3N7P)
- Size: 3�5 (medium card size)
- Color: Green theme in UI
- Parent: Must belong to a Main card
- Children: Can have multiple Nano cards
Features
- Parent-boosted scoring: Get 20% score boost when parent Main card is relevant
- Independent search: Can be found on their own merit
- Status tracking: Support Todo/In Progress/Done states
- Priority levels: Can be marked High/Medium/Low
- Context carriers: Bring their Nano children when selected
When to Create a Sub Card
- Breaking down a project into tasks
- Adding chapters to a learning topic
- Detailing components of a system
- Creating step-by-step procedures
- Documenting specific features or functions
Examples
S-TASK001: "Research competitor pricing" (parent: M-PROJECT1)
S-LESSON1: "Variables and data types" (parent: M-LEARN01)
S-THEORY1: "Wave-particle duality" (parent: M-CONCEPT)
S-PROCESS: "Employee onboarding steps" (parent: M-SYSTEM1)
Nano Cards (N-XXXXXX)
Purpose
Nano cards capture granular details - the examples, code snippets, quick facts, and specific references that support your Sub cards with concrete information.
Characteristics
- ID Format: N-XXXXXX (e.g., N-5K8L2Q)
- Size: 3�2 (smallest card size)
- Color: Yellow theme in UI
- Parent: Must belong to a Sub card
- Children: None (leaf nodes)
Special Behavior
� Important: Nano cards NEVER appear in search results directly. They are automatically included when their parent Sub card is selected.
Features
- Search invisible: Don't clutter search results
- Auto-inclusion: Come along with parent Sub card
- Score inheritance: Get 50% of parent's relevance score
- Lightweight: Perfect for quick notes and examples
- Bulk friendly: Can have many without affecting search
When to Create a Nano Card
- Adding code examples
- Storing specific commands or syntax
- Recording quick facts or statistics
- Saving URLs or references
- Capturing meeting notes or quotes
- Documenting edge cases
Examples
N-CODE001: "list_comprehension = [x**2 for x in range(10)]" (parent: S-LESSON1)
N-PRICE01: "Competitor A: $49/month for basic tier" (parent: S-TASK001)
N-FORMULA: "E = mc�" (parent: S-THEORY1)
N-LINK001: "https://docs.python.org/3/tutorial/" (parent: S-LESSON1)
How They Work Together
The Search & Context Flow
- User asks a question in chat
- System searches Main and Sub cards only
- Scoring happens:
- Main cards scored by relevance
- Sub cards scored independently
- Sub cards get 20% boost if parent Main card scores well - Top 10 cards selected for context
- Nano cards automatically included with any selected Sub cards
- AI receives complete context with all relevant details
Auto-Promotion Intelligence
When the #1 highest-scoring card is auto-promoted:
- If it's a Main card: Just the Main card is promoted
- If it's a Sub card: The Sub card AND its Nano children are promoted
- Nano cards alone: Never promoted (they can't be found in search)
Example Scenario
User asks: "How do I handle errors in Python?"
System finds:
1. Main card "Python Programming" (M-LEARN01) - Score: 8.5
2. Sub card "Exception handling" (S-ERROR01) - Score: 10.2 (boosted)
3. Sub card "Debugging techniques" (S-DEBUG01) - Score: 7.1
Context includes:
- S-ERROR01 and its 3 Nano cards with code examples
- M-LEARN01 (parent of top Sub card)
- S-DEBUG01 and its 2 Nano cards with tips
- Total: 3 searchable cards + 5 auto-included Nano cards
Best Practices
Organizing Your Knowledge
Start with Main Cards
- Create Main cards for major topics first
- Don't create too many - aim for 10-20 core topics
- Use clear, searchable titles
- Add comprehensive tags
Build Out with Sub Cards
- Add Sub cards as you dive into details
- Keep Sub cards focused on one aspect
- Use action-oriented titles for tasks
- Maintain consistent terminology
Enhance with Nano Cards
- Add Nano cards liberally - they don't affect search
- Use for anything you might need to reference
- Keep them concise and specific
- Perfect for copy-paste content
Tagging Strategy
Main Card: #python #programming #backend
Sub Card: #python #errors #exceptions (inherits + adds)
Nano: (inherits all parent tags automatically)
Relationship Management
Good Structure:
M-RECIPE: "Chocolate Cake Recipe"
S-INGRED: "Ingredients list"
N-DRY: "2 cups flour, 1 cup sugar..."
N-WET: "3 eggs, 1 cup milk..."
S-STEPS: "Baking instructions"
N-STEP1: "Preheat oven to 350�F"
N-STEP2: "Mix dry ingredients"
N-STEP3: "Combine and bake 30 min"
Poor Structure:
M-NOTES: "All my notes" � Too broad
S-STUFF: "Various things" � Too vague
N-RANDOM: "Some information" � Not specific
Technical Details
ID Generation
- Format:
[PREFIX]-[6 RANDOM CHARS] - Prefixes: M (Main), S (Sub), N (Nano)
- Characters: Uppercase letters and numbers
- Uniqueness: Checked against existing cards
- Example: M-7K9B2L, S-X3N7P4, N-5Q8M1R
Storage Structure
{
"version": "1.4",
"cards": [
{
"id": "M-ABC123",
"type": "main",
"title": "Machine Learning",
"content": "...",
"tags": ["ml", "ai"],
"subCards": ["S-XYZ456", "S-XYZ789"],
"createdAt": "2024-01-15T10:30:00Z",
"x": 250,
"y": 150
},
{
"id": "S-XYZ456",
"type": "sub",
"parentId": "M-ABC123",
"title": "Neural Networks",
"nanoCards": ["N-123DEF"],
// ...
}
]
}
Scoring Algorithm
Main Cards:
- Base NLP score: 0-10 points
- Keyword boost: 1-3 points
- Tag bonus: 0-2 points
- Total: 0-15 points
Sub Cards:
- Same base scoring as Main cards
- +20% boost if parent Main card scores > 0.3
- Example: Base score 8.0 � Boosted to 9.6
Nano Cards:
- Not scored directly (invisible to search)
- When included, inherit 50% of parent Sub card's score
- Used for context ranking but not selection
Real-World Examples
Learning a New Technology
M-DOCKER: "Docker Containerization"
S-BASICS: "Docker fundamentals"
N-INSTALL: "apt-get install docker.io"
N-CONCEPT: "Containers vs VMs: containers share kernel..."
N-TERMS: "Image: template, Container: running instance"
S-COMMANDS: "Essential Docker commands"
N-BUILD: "docker build -t myapp ."
N-RUN: "docker run -p 8080:80 myapp"
N-DEBUG: "docker logs <container-id>"
S-COMPOSE: "Docker Compose multi-container"
N-SYNTAX: "version: '3.8' services: web: image:..."
N-NETWORK: "Networks allow container communication"
Project Management
M-LAUNCH: "Product Launch Q1 2024"
S-RESEARCH: "Market research phase"
N-COMP1: "Competitor A: $99/mo, 10k users"
N-COMP2: "Competitor B: $149/mo, better UI"
N-SURVEY: "User survey: 73% want feature X"
S-DEVELOP: "Development milestones"
N-SPRINT1: "Auth system, user profiles"
N-SPRINT2: "Core features, API"
N-SPRINT3: "Polish, performance"
S-MARKETING: "Marketing campaign"
N-BUDGET: "$15k total: $5k ads, $10k content"
N-CHANNELS: "Focus: Twitter, LinkedIn, Reddit"
N-MESSAGE: "Tagline: 'Simplify your workflow'"
Personal Knowledge Base
M-HEALTH: "Health & Fitness"
S-NUTRITION: "Diet and meal planning"
N-MACROS: "Protein: 150g, Carbs: 200g, Fat: 65g"
N-MEAL1: "Breakfast: Oats, eggs, berries"
N-SUPP: "Vitamin D: 2000IU, Omega-3: 1g"
S-WORKOUT: "Exercise routines"
N-MONDAY: "Chest & Triceps: Bench 3x8, Dips 3x10..."
N-CARDIO: "Zone 2: 150-min/week, HR 120-140"
N-STRETCH: "Daily: Hip flexors, hamstrings, shoulders"
S-TRACKING: "Progress metrics"
N-WEIGHT: "Jan: 180lb, Feb: 178lb, Target: 175lb"
N-STRENGTH: "Bench: 185lb�205lb, Squat: 225lb�255lb"
Tips for Success
Do's
Create Main cards for distinct topics
Use Sub cards to break down complexity
Add Nano cards liberally for details
Maintain clear parent-child relationships
Use descriptive, searchable titles
Tag consistently across levels
Review and refactor your hierarchy periodically
Don'ts
L Don't create Main cards for everything
L Don't skip Sub cards and go straight to Nano
L Don't worry about too many Nano cards
L Don't create circular relationships
L Don't use vague titles like "Notes" or "Misc"
L Don't forget to link cards properly
Conclusion
The Three Core Cards system provides a powerful yet simple way to organize any type of knowledge. By understanding how Main, Sub, and Nano cards work together, you can build a personal knowledge base that grows with you and makes your information instantly accessible through AI-powered search and context.
Remember: Main cards for concepts, Sub cards for details, Nano cards for specifics. Let the hierarchy work for you, not against you.