Dqz00116 Layered First Principles Teaching
Overview
Transform complex concepts into progressive, first-principles explanations that build understanding layer by layer.
When to Use
Use when:
- You need to explain a complex concept to someone with less domain knowledge
- You're designing training materials, tutorials, or educational content
- You're preparing a technical presentation and need progressive disclosure
- A topic has multiple abstraction layers that require cognitive scaffolding
- You need to bridge the gap between intuitive understanding and technical depth
Don't use when:
- The concept is simple and doesn't benefit from layered decomposition
- You need quick reference documentation or a terse answer
- The audience already has deep expertise and only needs edge cases or implementation details
Quick Start
# Explain a concept progressively
kimi layered-first-principles-teaching "Explain blockchain"
# Target specific audience
kimi layered-first-principles-teaching "Explain transformers" --audience beginner
# Output to file
kimi layered-first-principles-teaching "Explain consensus algorithms" --output ./tutorial.md
Output Structure
Generated explanations contain 6 standard sections:
| Section | Content | Purpose |
|---|---|---|
| Opening | One-sentence essence + intuitive analogy | Immediate understanding |
| First Principles | Problem essence, why existing solutions fail | Foundation building |
| Progressive Layers | 3-4 layers from intuition to technical detail | Scaffolding learning |
| Analogies | Cross-domain comparisons | Relating to known concepts |
| Visualizations | ASCII diagrams, mental models | Spatial understanding |
| Summary | Key takeaways + further reading | Retention & next steps |
Audience Levels
| Level | Characteristics | Approach |
|---|---|---|
| Beginner | No prior knowledge | Heavy analogies, minimal jargon, focus on "why" |
| Intermediate | Some domain knowledge | Balance of intuition and technical detail |
| Expert | Deep domain knowledge | Focus on nuances, edge cases, implementation |
Teaching Patterns
This skill uses progressive disclosure patterns from prompts/:
- First Principles Analysis (
first_principles.txt): Strip away abstractions, find root causes - Layered Decomposition (
layered_decomposition.txt): Break into 3-4 cognitive layers - Analogy Generation (
analogy_generation.txt): Find relatable comparisons - Visualization Design (
visualization_design.txt): Create mental models and diagrams
Templates
Output templates in templates/ provide structure for:
concept.md: General concept explanationalgorithm.md: Algorithm walkthroughsystem.md: System architecture explanation
Examples
See examples/ for completed explanations:
blockchain_explained.md: From "digital ledger" to Byzantine fault tolerancetransformers_explained.md: From "pattern matching" to attention mechanisms
Workflow
When explaining a concept:
- Load first principles prompt → Identify core problem and breakthrough insight
- Load layered decomposition prompt → Structure into 3-4 cognitive layers
- Load analogy generation prompt → Find 2-3 cross-domain analogies
- Load visualization design prompt → Create ASCII diagrams and mental models
- Apply appropriate template → Generate final explanation
Constraints
- Maximum 4 layers to avoid cognitive overload
- Each layer must build on previous without introducing new prerequisites
- Analogies must be familiar to target audience
- Visualizations should work in plain text (ASCII)