The Spaced Repetition Curriculum Builder: Turn Any Study Material Into a Structured Review Schedule
Why this prompt matters
Most self-study fails not from lack of material but from lack of a retrieval schedule. Learners re-read notes — which is recognition, not recall — and per the Ebbinghaus forgetting curve lose roughly 70% of new material within a week without active review. That pattern typically ends in a cram week before the deadline, which produces recall that feels solid in the moment but collapses within days of the exam, effectively wasting the weeks of study time that came before it.
What we use it for
You have 6 weeks before the AWS Solutions Architect Associate exam, a 40-page official study guide, and 5 hours a week to study around a full-time job. You've read the guide once already but you're not confident it stuck, and you don't know how to turn a static document into a day-by-day plan that actually uses spaced repetition instead of a second read-through.
Prompt
Act as an expert learning scientist who designs spaced-repetition study systems based on cognitive science research (the forgetting curve, active recall, interleaving). Context: I need to learn [SUBJECT OR MATERIAL] before [DEADLINE OR EXAM DATE]. I have [AVAILABLE STUDY TIME PER WEEK] to dedicate to this. My current familiarity with the material is [BEGINNER / INTERMEDIATE / ADVANCED]. Here is the source material or topic outline: [PASTE YOUR STUDY MATERIAL, SYLLABUS, OR TOPIC LIST HERE]. Task: 1. Break the material into discrete, testable concept units — not chapter titles, but atomic facts and skills a person can actually be quizzed on. 2. Rank each unit by difficulty and dependency (what must be learned before what). 3. Build a spaced-repetition review schedule using expanding intervals (day 1, day 3, day 7, day 14, day 30) mapped onto my stated weekly study time and deadline. 4. For each concept unit, write 2-3 active-recall questions that force me to produce an answer rather than recognize one. 5. Flag which units are highest-risk for my goal and need extra review passes. Constraints: - Do not just summarize the material — decompose it into atomic, individually testable units. - Recall prompts must require producing an answer, not selecting one from choices. - The schedule must fit within my stated weekly time budget. If it genuinely doesn't fit, say so explicitly and tell me what to cut or how much to extend the deadline. - Interleave topics once initial learning is done — do not schedule all units from one topic back-to-back in later review sessions. Output Format: - Section 1: Concept Unit List (numbered, with dependency notes) - Section 2: Review Calendar (table: Date | Units to Review | Est. Minutes) - Section 3: Active Recall Question Bank (grouped by concept unit) - Section 4: Risk Flags (which units need extra passes, and why)
Result
**Section 1: Concept Unit List (excerpt)** 1. IAM roles vs. IAM users vs. resource policies (no dependency — foundational) 2. Cross-account access via IAM role assumption (depends on Unit 1) 3. S3 storage classes and lifecycle transitions (no dependency — foundational) 4. S3 cross-region replication requirements (depends on Unit 3) 5. VPC subnets: public vs. private, and route table behavior (no dependency — foundational) 6. NAT Gateway vs. NAT Instance failover characteristics (depends on Unit 5) 7. Auto Scaling Group health checks: EC2 vs. ELB (depends on Unit 5) ... [34 more units covering compute, database, and networking domains]
**Section 2: Review Calendar (excerpt)** | Date | Units to Review | Est. Minutes | |---|---|---| | Mon Wk1 | New: Units 1-7 (IAM, S3 basics) | 60 | | Wed Wk1 | Review: Units 1-7 (Day 3 pass) | 25 | | Fri Wk1 | New: Units 8-14 (VPC, compute) | 60 | | Sun Wk1 | Review: Units 1-14 (mixed, Day 3+7 passes) | 45 | | Mon Wk2 | New: Units 15-21 (databases) | 60 | ... [continues through Week 6, exam-week schedule front-loads Risk Flag units]
**Section 3: Active Recall Question Bank (sample — Unit 2)** - Q: A company needs its Lambda function in Account A to write to an S3 bucket in Account B. Walk through the two IAM elements required and which account each one lives in. - Q: What's the maximum session duration for an assumed role by default, and how would you extend it? - Q: Why does resource-based policy alone sometimes fail for cross-account S3 access even when the bucket policy looks correct?
**Section 4: Risk Flags** - Units 4, 9, 22 (S3 replication, VPC peering limits, RDS Multi-AZ failover mechanics): historically the most-missed exam topics per official guide weighting — scheduled for an extra Day-21 review pass not in the standard interval sequence. - Your 5 hrs/week budget covers 34 of the 41 units at the standard cadence; the plan compresses Units 35-41 (advanced networking) into paired sessions in Week 5 — flagging this as tight. Consider either 6.5 hrs/week from Week 4 onward, or accept lighter review depth on those units.
Most study prompts people write for AI amount to "summarize this" or "quiz me on this," and both fall short of what actually produces retention. Summarizing produces recognition-friendly notes that feel like learning but don't survive contact with a real exam. Random quizzing without a schedule ignores the single best-documented finding in learning science: spaced review beats massed review, and the spacing has to expand over time to match how memory actually decays.
Why the structure matters
The prompt forces three things a plain summarization request never does. First, it decomposes material into atomic units rather than chapter-level chunks — "IAM roles vs. IAM users vs. resource policies" is testable in a way that "Chapter 3: Identity and Access Management" is not. Second, it maps those units onto an actual calendar with expanding intervals (1, 3, 7, 14, 30 days), which is the interval pattern spaced-repetition research consistently shows outperforms both cramming and evenly-spaced review. Third, it generates active-recall questions rather than flashcard-style prompts, because producing an answer from memory builds a stronger retrieval pathway than recognizing a correct answer among choices.
The constraint that does the real work
The instruction requiring the plan to fit inside a stated weekly time budget — and to say explicitly when it doesn't fit — is the most load-bearing constraint in the prompt. Study plans that ignore real time constraints get abandoned within a week. Forcing the model to flag an infeasible plan rather than silently producing an unrealistic one turns a study calendar into something a person can actually follow, and gives them the information to make a real tradeoff between more weekly hours or a longer timeline.
Adapting it
The prompt works for exam prep, but the same structure applies to onboarding a new hire onto an unfamiliar codebase, ramping up on a new technical domain before a role change, or preparing for a technical interview loop. The key input to get right is the source material field — the more structured the pasted content (an actual syllabus or outline beats a vague topic name), the more accurately the model can build genuinely atomic, testable units instead of guessing at scope.