When: Friday afternoon, or any consistent five-minute window.
What to do: 1. Pick any three Compression Blocks from recent reading. (If you are reading this book for the first time, use the chapters you read this week.) 2. For each Compression Block: cover the right column with your hand or a card. 3. Read the Concept field. 4. Try to produce: Core Idea, the Tradeoff, the Failure Mode, the Signal. 5. Uncover each field only after attempting recall. 6. For fields you could not produce: read, cover, repeat once.
What you are training: The retrieval pathway from concept name to full framework description. This is the recall you need in a design review or architecture conversation.
Why it works in 5 minutes: Three Compression Blocks, 30–60 seconds each for the attempted recall, plus correction time if needed. Under 5 minutes is achievable even on a busy Friday.
When: The first Monday of each month, or any consistent 30-minute window.
What to do: 1. Pick one framework chapter (F1 through F9). Rotate through them — 9 months to cover all nine. 2. Close the book. On a blank page or whiteboard, try to write: the full list of items in the framework (all 12 mental models, all 15 principles, etc.). 3. Open the book and check. Circle any items you missed or mis-remembered. 4. For each missed item: read the Layer 2 paragraph. Close the book. Recall the item in your own words. 5. That is the session.
What you are training: The ability to enumerate a framework completely under pressure. Once that recall is reliable, the principles get used — without fear of forgetting one mid-conversation.
The indicator: If you can produce fewer than 70% of the items in any framework without looking, that framework needs two monthly sessions before you move to the next.
When: Once per quarter, scheduled as a calendar block like any other important work.
What to do: 1. Pick a real system — either one you currently work on, or a well-known public system. 2. Apply the four-step method from Chapter 14 in full: classify archetype → run review questions → name tradeoffs → map failure modes. 3. Then apply the five architecture diagrams. Draw all five for the system you chose. Use the actual architecture where available; use a plausible architecture otherwise. 4. After completing the exercise, note which frameworks felt sharp (you applied them fluently) and which felt rusty (you had to look things up). 5. The rusty frameworks become the focus of the next two monthly sessions.
What you are training: The integration of all nine frameworks in a single applied exercise. Reading about frameworks is different from using them. The quarterly session is the practice round.
The outcome: A personal calibration of which frameworks are strongest and which need reinforcement. This is more valuable than a uniform review of all nine, because it directs effort toward what needs it.
When: Once per year, ideally at a natural reflection point (year-end, job anniversary, or returning from a conference).
What to do: 1. Read the Reference Book cover to cover. This should take 3–4 hours at normal reading speed. 2. Annotate as you go: mark concepts that hit differently with a year more of experience. Mark frameworks that feel outdated given what you have seen this year. 3. After reading: revisit your previous annotations (if you annotated the prior year). What has changed? What is more true than before? What is less? 4. Write one page: the three things you understand differently now than you did a year ago, and why.
What you are training: The metacognitive layer — your understanding of your own understanding. The frameworks will land differently after a year at a new company, after a difficult incident, after a significant architecture decision. The annual read is how you integrate experience into the framework rather than just accumulating experience alongside it.
The value of the one page: The writing is not for publishing. It is for the next year’s annual review. The person who read the frameworks fresh will not see what the person with a year of new context sees. Capturing this creates a log of your evolving technical judgment.