You get:
- tutorial hell — learning without doing
- reading lists instead of concept maps
- no way to test if you actually understand
- gaps you discover after you’ve already started
- knowledge that never becomes action
But learning for a project is not accumulation.
It is targeted acquisition with a deadline.
- Essential concepts are the minimum viable knowledge set
- Testing questions reveal false understanding
- Resources should be specific, not “learn more about X”
- A synthesis prompt forces you to connect ideas
- Knowledge gap warnings prevent nasty surprises
Without curation, you learn everything and apply nothing.
This framework forces AI to be a learning architect who builds only what you need.
Assume the role of a knowledge curator and project-based learning architect who builds minimal viable knowledge packs. Your task is to create a Knowledge Pack for a specific project or learning goal. Generate: 1. 5-7 ESSENTIAL CONCEPTS What you must understand to succeed (with brief definitions) 2. TESTING QUESTIONS (one per concept) A question that reveals actual understanding, not recall 3. RESOURCE LIST (3-5 specific items) Books, articles, videos, courses, or people — no generic suggestions 4. SYNTHESIS PROMPT "Write a paragraph explaining how [CONCEPT X] relates to [CONCEPT Y] using only your own words." 5. KNOWLEDGE GAP WARNING What's missing from your current understanding that will hurt you later INPUTS: Project or Learning Goal: [E.G., "Launch a newsletter" / "Learn data analysis in Python" / "Write a business plan"] Your Current Knowledge Level (1-10): [INSERT NUMBER] Time Available Before Launch/Deadline: [INSERT DAYS OR WEEKS] One Thing You're Already Confused About (optional): [INSERT] RULES: - Essential concepts must be minimal (5-7, not 20) - Testing questions must be open-ended (not yes/no) - Resources must be specific enough to find (author, title, platform) - The knowledge gap warning must be specific, not "you need more experience" - If you have less than 2 weeks, prioritize the top 3 concepts only
- Use this before starting any significant project — it’s your learning roadmap.
- Answer the testing questions out loud; if you hesitate, you don’t understand.
- The synthesis prompt is the most valuable output — write it immediately.
- The knowledge gap warning is not discouragement; it’s advance warning.
- Revisit the Knowledge Pack weekly; your gaps will shift as you learn.
Project or Learning Goal: Launch a paid newsletter on Substack about urban planning
Your Current Knowledge Level: 6/10 (I know urban planning concepts, but nothing about running a paid publication)
Time Available Before Launch/Deadline: 6 weeks
One Thing You’re Already Confused About: How to convert free subscribers to paid without being pushy
This framework improves outcomes by forcing:
- essential concept identification (minimum viable knowledge)
- understanding testing (not recall testing)
- specific resource curation
- synthesis as truth detector
- honest gap warnings
Great project-based learning doesn’t teach you everything — it teaches you exactly what you need, when you need it.
Build Better AI Systems
Subscribe for advanced prompt engineering, AI knowledge systems, project-based learning frameworks, and practical strategies for builders and creators.