Write permanent notes for your six-months-from-now self without source access — no shorthand, no "see above," fully self-contained

Why This Is a Rule

The value of permanent notes is realized at retrieval time, not creation time. At creation, you're saturated with context: you just read the source, you're thinking about the topic, the shorthand makes sense, and the note seems perfectly clear. Six months later, that context has evaporated. "The mechanism is the same as Ch.3 — see notes on boundary conditions" is meaningless when you don't remember Chapter 3, can't find the referenced notes, and don't recall what "boundary conditions" referred to in this context.

Self-contained notes survive context loss because they carry their own context. Instead of "Same as Ch.3" → "The feedback mechanism here operates through the same loop as [linked note: Negative Feedback Loops in Temperature Regulation]: deviation from setpoint triggers corrective action proportional to the deviation magnitude." The note is comprehensible in isolation, six months from now, without the source, without your current mental state, without any external reference.

The "six months without source access" standard is deliberately conservative. If a note survives this test, it survives any realistic retrieval scenario. The test forces you to externalize every assumption, spell out every abbreviation, and replace every relative reference ("above," "here," "this") with an absolute one (the specific concept, note, or fact).

When This Fires

• When writing any note intended to persist in your knowledge system
• When reviewing old notes and finding them incomprehensible
• When your note system contains notes you wrote but can no longer understand
• Complements Every new permanent note must link to at least one existing note — identify relationships at creation, not during some future review (link at creation) with the quality standard for the note itself

Common Failure Mode

Context-dependent shorthand: "Same mechanism as before, but with X instead of Y — implications are obvious." Nothing is obvious six months later. Every "obvious" implication should be spelled out. Every "same as before" should be replaced with the specific mechanism. Every shorthand should be expanded.

The Protocol

(1) After writing a permanent note, apply the six-month test: "If I read this note in six months, having forgotten everything about the source and context, would I understand it?" (2) Check for: Relative references ("above," "this," "the author") → replace with specific names and concepts. Shorthand ("RLHF," "ToC," "the usual mechanism") → expand or define. Context-dependent claims ("obviously this means X") → explain why it means X. Source-dependent logic ("as shown in the paper") → state the conclusion in the note itself. (3) Each permanent note should contain: the claim or idea (stated explicitly), the reasoning (why this is true or important), and enough context to understand both without any external reference. (4) Use the note's title as an additional comprehension test: does the title alone convey the core idea? (5) Err on the side of over-explanation. A note that's slightly verbose but fully comprehensible is vastly more valuable than a concise note that's opaque.