Dan Wang's writing process operates on three distinct temporal layers, each with different purposes:
The Capture Layer (year-long, immediate): When something interesting happens — a great sentence drifts into mind, an observation lands, a phrase captures something important — he immediately records it in his iPhone notes app. The capture is unpolished. It is not meant to be written prose yet. It is raw material.1 The discipline is immediate recording. The moment you think "I'll remember that," you've usually lost it.
The Refinement Layer (months-long, return visits): Because the annual deadline is January 1st, Wang has months to return to captured material. He can spend an afternoon on a single sentence, testing variations, discovering what it can carry, removing unnecessary words, discovering how to make it do more work.1 This is where the stylistic deliberation from studying Mozart and Stendhal translates into practice.
The Assembly Layer (final 10 days, hasty sprint): When December arrives, all the cognitive work is done. The hard decisions about which sentence is best, which observation belongs where, what structure contains all this material — those decisions are complete. The final sprint is execution of already-made decisions.1
Most writers collapse these layers. They try to capture, refine, and assemble simultaneously. This prevents any of the layers from working. Capture becomes slow because you're worrying about polish. Refinement never happens because you're already moving to the next project. Assembly becomes panicked because you're building from unrefined material.
Wang's system works because the layers are separated in time.
Not everything gets captured. Wang is selective. Three categories go in:
Sentences: Exact phrasings that arrive as complete units. Not paraphrases or rough thoughts. The exact sentence that says the thing. These might be sentences he hears, sentences he thinks of, or sentences that appear in something he's reading and strike him as worth remembering.1
Observations: Specific concrete moments that seemed important. Not generalizations. Not reactions. The actual moment. "I heard someone say X" not "people seem to believe Y." The specificity is crucial because observation without detail is abstraction.
Ideas: Structural arguments or questions that seem worth pursuing. But captured at the level of precision, not vague inspiration. "What does the relationship between official directives and actual behavior reveal about information control?" not "interesting thing about China."1
What doesn't get captured: reactions, feelings, judgments about whether something matters. The capture system is not a journal. It is not reflective. It is a bucket for material that might matter for writing.
This selectivity is important. An undisciplined capture system becomes overwhelming. Everything that seems interesting goes in. By the end of the year, you have thousands of unprocessed notes and no way to synthesize them. A disciplined system captures only material that is specific, precise, and generative.
Having raw material is only useful if you return to it. Wang's approach requires that you set aside time throughout the year to revisit captured sentences and observations. This is not daily review. It is periodic engagement. Months in, you can return to a sentence captured months ago and discover you can improve it. You understand now what you didn't when you first captured it.
This is where sentence-level craft improves. Each return visit to a sentence is an opportunity to ask: Is this the right word? Can this be more precise? Does this carry the weight I want it to carry? Can the rhythm improve? You are refining without committing — there is no pressure because the final essay does not exist yet.
The refinement layer is where the study of stylistic models becomes practical. The variations you're testing are not random. They are informed by your study of how Stendhal moves between scales, how Mozart creates rhythm through repetition with variation, how the best writers economize language.
Sentence as Structural Anchor — Note capture is specifically designed to accumulate material excellent enough to become structural anchors. The refinement layer ensures that captured sentences reach the level of excellence required.
Annual Publishing Discipline — The three-layer system only works within the annual timeline. Without the boundary of January 1st, refinement has no endpoint and assembly never begins.
Voice Cultivation Through Stylistic Models — Refinement is where the study of models becomes embodied practice. You are not imitating. You are solving craft problems using solutions studied from masters.
Choose your capture location: iPhone notes, dedicated notes app, or physical notebook. Single location only. The discipline depends on immediate capture. If you have to decide where to put something, you've already slowed it down.
Establish capturing discipline: Anytime a sentence or observation or precise idea arrives, capture it immediately. No filtering. No "is this good enough?" Just capture.
Monthly refinement session (30 minutes): Once monthly, read all captured material. Don't organize or edit. Just review. This primes your mind to recognize patterns.
Quarterly deep review (2 hours): Every three months, go through all captured material slowly. Test variations on captured sentences. Try shortening them. Try removing a word. See what emerges. Make notes on promising directions.
Pre-assembly audit (early December): Before the final sprint, review all captured material. What clusters have emerged? Which sentences are genuinely excellent? Which observations have surprising connections? This is planning without writing.
The key is separation of layers. Capture does not refinement. Refinement does not assembly. Each layer has its own time and discipline.
The Sharpest Implication: If writing quality depends on separation of capture, refinement, and assembly across different time scales, then writing quickly from current ideas produces inferior work not because speed is bad but because speed collapses necessary layers. The constraint of time separation is more important than the constraint of deadline pressure. You could write one essay per year and still produce poor work if you capture, refine, and assemble simultaneously. You could write one essay per month and produce excellent work if you properly separate the layers.
Generative Questions: