Context Engineering Guide - Syllabus
Version: 1.0
Date: 2025-11-28
Status: Draft
Content Inventory
Main Articles (9 articles)
| # | File | Title | Main Topics | Est. Words | |---|------|-------|-------------|------------| | 1 | 1_prompt_is_not_enough.md | Prompt Is Not Enough | Prompt vs Context, AI as colleague metaphor, "giving little expecting much", 1-2 min rule, car seat example | 1,200 | | 2 | 2_what_good_context_looks_like.md | What Good Context Looks Like | Perplexity workflow, task template structure, Zed+Claude usage, iteration, context contamination | 1,100 | | 3 | 3_why_you_give_much_get_little.md | Why You Give Much and Get Little | Common mistakes, fixing vs starting fresh, context contamination, feedback with role play, junior test | 1,300 | | 4 | 4_think_like_engineer.md | Think Like an Engineer | Atomization principle, parallel agents trap, clear goals flexible path, vibe coding 2-min rule, workflow for complex tasks | 1,200 | | 5 | 5_ai_wont_steal_your_job.md | AI Won't Steal Your Job | Imposter syndrome, wow moments (garden, electricity), tool division, human irreplaceability, AI in schools | 1,100 | | 6 | 6_tools_i_use.md | Tools I Use | Perplexity/Zed+Claude/ChatGPT division, tool combinations, cost breakdown, practical settings | 1,500 | | 7 | 7_context_engineering_for_teams.md | Context Engineering for Teams | Why teams fail, documentation focus, team task template, implementation timeline, roles, success metrics, cultural change | 1,800 | | 8 | 8_practical_examples.md | Practical Examples | 6 detailed examples: debugging, tech selection, SQL optimization, documentation, legacy refactoring, consumption analysis | 2,500 | | 9 | 9_vibe_coding_vs_context_engineering.md | Vibe Coding vs Context Engineering | Vibe coding critique, parallel agents problem, context as fuel, QA bottleneck, how articles were written | 1,400 |
Total estimated: ~13,100 words in main articles
Supporting Materials
| File | Type | Key Content | |------|------|-------------| | otazky.md | Interview Q&A | 25+ questions with detailed answers, raw insights, personal examples | | clanky.md | Article Plan | Original series structure with topic assignments | | start.md | Research Notes | LinkedIn posts, external citations, raw thoughts | | presentation_structure.md | Presentation | 10 advices condensed, visual format, hooks | | linkedin_post_*.md | Social Posts | 9 teaser posts with key highlights |
Raw Topic List
Extracted from all source materials:
- Prompt vs Context definition
- AI as colleague metaphor ("knows nothing but can learn everything")
- "Giving little, expecting much" pattern
- 1-2 minute rule for AI response
- Car seat selection example
- Expense tracker example
- Perplexity workflow (research → context → AI)
- Task template structure (Problem, Context, Goal, Solutions, Tests)
- Zed + Claude usage patterns
- Context contamination problem
- When to fix vs when to start fresh
- Iteration and prompt editing
- Atomization principle (breaking down tasks)
- Parallel agents trap
- Clear goals, flexible path
- Feedback with role play
- "Is my plan good?" anti-pattern
- Junior developer test
- Imposter syndrome with AI
- Wow moments (garden, electricity analysis)
- Tool division by task type
- Tool combinations that work
- Team implementation challenges
- Documentation as context
- Team task templates
- Roles in AI-enabled teams
- Cultural change requirements
- Success metrics
- Debugging example (profile edit bug)
- Technology selection example
- SQL optimization example
- Documentation generation example
- Legacy refactoring example
- Consumption analysis example
- Vibe coding critique
- QA as bottleneck
- Model Context Protocol (MCP)
- How to write articles with AI
- AI in schools vision
- Human irreplaceability (situational awareness)
- Karpathy quote on context engineering
- Chroma Research on context rot
- Senior vs Junior (context is the difference)
Topic Clusters
Cluster A: Foundations - What is Context Engineering?
- Prompt vs Context definition
- AI as colleague metaphor
- Why context matters more than prompts
- Karpathy quote and industry validation
- Senior vs Junior = context quality
Cluster B: The Core Problem
- "Giving little, expecting much" pattern
- Why AI output seems unusable
- Common mistakes people make
- "Is my plan good?" anti-pattern
Cluster C: How Good Context Looks
- Components of good context
- Task template structure
- Examples of bad vs good context
- The junior developer test
Cluster D: The Process - Before You Prompt
- Give the WHY not just WHAT
- Break into atomic parts
- Show examples (one example > 1000 words)
- Say what to EXCLUDE
- Define what DONE looks like
Cluster E: The Process - During and After
- The 1-2 minute rule
- Explain WHY it's wrong
- Context contamination and when to start fresh
- Ask for opinion, not validation
- Iteration and prompt editing
Cluster F: Tools and Workflows
- Perplexity for research
- Zed + Claude for coding
- ChatGPT for non-technical
- Tool combinations
- Practical settings
Cluster G: Real Examples
- Debugging (profile edit bug)
- Technology selection
- SQL optimization
- Documentation generation
- Legacy refactoring
- Consumption analysis (solar panels)
- Car seat selection
- Garden advice
Cluster H: Teams and Scaling
- Why teams fail with AI
- Documentation as context
- Team task templates
- Implementation timeline
- Roles (PO, Dev, QA, PM)
- Success metrics
- Cultural change
Cluster I: Bigger Picture
- Vibe coding critique
- Parallel agents trap
- QA as bottleneck
- AI won't replace you
- Human irreplaceability
- AI in education
- Future of context engineering
Hierarchical Structure
Part 1: Understanding Context Engineering
Chapter 1: Why Your AI Prompts Fail
1.1 The "Giving Little, Expecting Much" Trap
- Source: Article 1, Interview Q20
- Key points: Personal evolution story, most people's mistake
1.2 Prompt vs Context - The Real Difference
- Source: Article 1, Interview Q19
- Key points: Prompt = task, Context = everything else
1.3 AI as Your New Colleague
- Source: Article 1, Interview Q8-9
- Key points: Knows nothing but can learn everything, 24/7 availability, trade-offs
1.4 Industry Validation
- Source: Article 9, External citations
- Key points: Karpathy quote, Tobi Lutke, Harrison Chase
Chapter 2: What Good Context Looks Like
2.1 The Five Components of Context
- Source: Article 2, Presentation
- Key points: Task, Constraints, Background, Examples, Success Criteria
2.2 The Universal Task Template
- Source: Article 2, Article 7
- Key points: Problem, Context, Goal, Solutions, Tests
2.3 Before and After: Real Examples
- Source: Article 1, Article 3
- Key points: Expense tracker, article writing, bad vs good prompts
2.4 The Junior Developer Test
- Source: Article 3, Interview Q24
- Key points: "Could a new colleague complete this?", universal applicability
Part 2: The Practice
Chapter 3: Before You Prompt - Preparation
3.1 Give the WHY, Not Just WHAT
- Source: Presentation, Interview Q1-3
- Key points: AI doesn't know priorities, Perplexity car seat example
3.2 Break Into Atomic Parts
- Source: Article 4, Interview Q5
- Key points: Atomization principle, smaller = better results
3.3 One Example Beats 1000 Words
- Source: Article 2, Article 3
- Key points: Show don't tell, attach previous work
3.4 Say What to EXCLUDE
- Source: Article 8, Interview examples
- Key points: AI includes everything unless told not to
3.5 Define What DONE Looks Like
- Source: Article 3, Interview Q22
- Key points: Clear success criteria, verifiable outcomes
Chapter 4: During and After - Iteration
4.1 The 2-Minute Rule
- Source: Article 1, Article 4
- Key points: Signal not failure, when to stop and adjust
4.2 Explain WHY It's Wrong
- Source: Article 3, Interview Q7
- Key points: AI learns from "wrong because X", give the lesson
4.3 Context Contamination
- Source: Article 2, Article 3, Article 9
- Key points: Bad output stays in context, when to start fresh
4.4 Ask for Opinion, Not Validation
- Source: Article 3, Interview Q18
- Key points: Models tuned to praise, better questions
4.5 Editing Prompts Mid-Work
- Source: Article 2, Interview Q11
- Key points: Zed approach, don't send correction messages
Chapter 5: Tools and Workflows
5.1 Choosing the Right Tool
- Source: Article 6, Interview Q12-13
- Key points: Perplexity/Claude/ChatGPT division by task type
5.2 Research Workflow
- Source: Article 2, Article 6
- Key points: Perplexity → extract → insert into next tool
5.3 Coding Workflow
- Source: Article 6, Interview Q3-4
- Key points: Zed + Claude, context management, Linear integration
5.4 Tool Combinations
- Source: Article 6
- Key points: What works together, what doesn't
Part 3: Real-World Application
Chapter 6: Practical Examples
6.1 Debugging Production Bug
- Source: Article 8
- Key points: Full bad/good comparison, profile edit bug
6.2 Technology Selection
- Source: Article 8
- Key points: Framework for admin panel, structured decision
6.3 SQL Optimization
- Source: Article 8
- Key points: From 8s to 0.3s with right context
6.4 Documentation Generation
- Source: Article 8
- Key points: OpenAPI spec, style matching
6.5 Legacy Code Refactoring
- Source: Article 8
- Key points: Step-by-step, don't change constraints
6.6 Personal Projects
- Source: Article 5, Article 8, Interview Q15
- Key points: Garden, solar panel analysis, car seat selection
Chapter 7: Context Engineering for Teams
7.1 Why Teams Fail with AI
- Source: Article 7
- Key points: Licenses without process, 2-hour training myth
7.2 Documentation as Context
- Source: Article 7
- Key points: "Everyone knows" doesn't work, what to document
7.3 Implementation Roadmap
- Source: Article 7
- Key points: 6-week plan, pilot → standardize → scale
7.4 Roles and Responsibilities
- Source: Article 7
- Key points: PO, Developer, QA, PM in AI workflow
7.5 Measuring Success
- Source: Article 7
- Key points: What to track, what NOT to track
7.6 Cultural Change
- Source: Article 7
- Key points: Mindset shifts, how to achieve them
Part 4: The Bigger Picture
Chapter 8: Vibe Coding vs Context Engineering
8.1 The Vibe Coding Reality
- Source: Article 9
- Key points: Lost projects, hacked apps, no backups
8.2 When Vibe Coding Works
- Source: Article 9
- Key points: Prototypes, scripts, experimentation
8.3 The Parallel Agents Trap
- Source: Article 4, Article 9
- Key points: Logical fallacy, dependent tasks
8.4 QA as Bottleneck
- Source: Article 7, Article 9
- Key points: Fast generation, slow verification
Chapter 9: AI Won't Replace You
9.1 The Imposter Syndrome
- Source: Article 5, Interview Q22
- Key points: Feeling like fraud, right approach
9.2 What AI Still Can't Do
- Source: Article 5, Interview Q16
- Key points: Situational awareness, human context
9.3 AI as Amplifier
- Source: Article 5, Interview Q25
- Key points: "Overcome yourself", not replacement
9.4 AI in Education
- Source: Article 5, Interview Q10, Q24
- Key points: Not for homework, for thinking, custom GPT idea
Appendix
A: Quick Reference - The 10 Advices
- Source: Presentation
- Condensed actionable tips with examples
B: Templates
- Task template
- Bug report template
- Feature request template
C: Tool Comparison
- Perplexity vs Claude vs ChatGPT
- When to use which
D: Citations and Resources
- Karpathy quotes
- Chroma Research
- External links
Coverage Map
| Syllabus Section | Primary Source | Supporting Sources | |------------------|----------------|-------------------| | 1.1 Giving little | Article 1 | Interview Q20 | | 1.2 Prompt vs Context | Article 1 | Interview Q19 | | 1.3 AI as colleague | Article 1 | Interview Q8-9 | | 1.4 Industry validation | Article 9 | External citations | | 2.1 Five components | Article 2 | Presentation | | 2.2 Task template | Article 2, 7 | Interview Q4 | | 2.3 Before/after | Article 1, 3 | Article 8 | | 2.4 Junior test | Article 3 | Interview Q24 | | 3.1-3.5 Before prompt | Presentation | Articles 1-4 | | 4.1-4.5 During/after | Articles 2-4 | Interview Q6-7 | | 5.1-5.4 Tools | Article 6 | Interview Q12-13 | | 6.1-6.6 Examples | Article 8 | Article 5, Interview Q15 | | 7.1-7.6 Teams | Article 7 | - | | 8.1-8.4 Vibe coding | Article 9 | Article 4 | | 9.1-9.4 Won't replace | Article 5 | Interview Q16, Q22, Q25 | | Appendix A | Presentation | - |
Coverage Summary
| Part | Chapters | Sections | Primary Sources | |------|----------|----------|-----------------| | 1: Understanding | 2 | 8 | Articles 1, 9 | | 2: Practice | 3 | 14 | Articles 2, 3, 4, 6 | | 3: Application | 2 | 12 | Articles 7, 8 | | 4: Bigger Picture | 2 | 8 | Articles 5, 9 | | Appendix | - | 4 | Presentation | | Total | 9 | 46 | All 9 articles + Interview |
Estimated Final Length
| Section | Est. Words | Est. Pages | |---------|------------|------------| | Introduction | 800 | 2 | | Part 1 (2 chapters) | 4,000 | 8 | | Part 2 (3 chapters) | 6,000 | 12 | | Part 3 (2 chapters) | 5,000 | 10 | | Part 4 (2 chapters) | 3,500 | 7 | | Appendix | 2,000 | 4 | | Total | ~21,300 | ~43 pages |
Unmapped Content
Content from sources that could be added but not essential:
- LinkedIn-specific posting advice from linkedin_post_*.md - consider for bonus material
- How the articles were written from Article 9 - could be appendix on meta-process
- Tool costs breakdown from Article 6 - nice-to-have but may date quickly
Open Questions
-
Should "Tools and Workflows" (Chapter 5) come before or after "Practical Examples" (Chapter 6)?
- Recommendation: Keep as-is. Tools first gives context for examples.
-
Should Vibe Coding critique be a full chapter or merged into another?
- Recommendation: Keep as chapter - it's a strong industry commentary.
-
Is 9 chapters too many? Could merge:
- Chapters 1+2 into "Foundations"
- Chapters 8+9 into "Industry Perspective"
- Recommendation: Keep separate for clearer navigation. Each chapter ~3,000 words is readable.
-
Should Quick Reference (10 Advices) be at beginning or end?
- Recommendation: End (Appendix). Readers need context first to understand the advices.
Quality Criteria Check
| Criterion | Status | |-----------|--------| | All source files reviewed | ✅ | | No major topic missing | ✅ | | Logical flow from basic to advanced | ✅ | | Each chapter has clear purpose | ✅ | | No orphan topics (uncategorized) | ✅ | | Coverage map complete | ✅ | | Estimated length reasonable (40-60 pages) | ✅ (~43 pages) |
Next Step
→ Proceed to Step 2: Topic Analysis → Create detailed analysis file for each chapter