Methods · History, human story & context

Note architecture

Keep notes you can actually recombine: atomic, linked, written in your own words, instead of a linear transcript you'll never reread.

When to use this: While you practise
Learning impact: Strong change
Learning evidence: Credible with caveats
Where you are with the topic: Strengthening memory
How often to use it: Build and keep as a habit

When to use it

Your notes are unreadable copies of the source and you cannot recombine them.
You have lots of highlights and no atomic claims in your own words.
Links between notes are decorative arrows with no relationship named.

What to ask AI

Type something like: "Here are two of my notes. What's the relationship between them: instance, contradiction, mechanism, precondition, analogy? If there isn't one, say so."

Ask it to refuse inventing connections.

Keep the hard part on your side: you write notes in your own words before asking for links.

How it helps

Notes worth keeping are generative: atomic, in your words, recombinable.

Architecture beats volume, a pile with arrows is not a system.

Naming the relationship between two notes is the thinking; "related" is not a relationship.

How to run it

Write notes as small claims in your own words.
Keep them atomic enough to link.
Ask what relationship holds between two notes, or none.
Reject invented connections.
Recombine for an essay outline only after relationships are named.

Example exchange

You

Here are two of my notes. First: spacing beats cramming for long-term retention. Second: interleaving can feel harder but improve discrimination. What relationship holds between them, and if there isn't one, say so.

They look like sibling mechanisms under desirable difficulty, not a contradiction. Analogy is a weak label here. Want a third note that would be a precondition for applying both?

You

Maybe that you need a set of items to schedule?

Yes, that is a precondition. Write it as its own atomic note in your own words.

Copyable prompt

Here are two of my notes. What's the relationship between them: instance,
contradiction, mechanism, precondition, analogy? If there isn't one, say so.

The Tell

Here is how you know this method has flipped: notes are still in the source's words, or links have no named relationship.

A beautiful graph of empty nodes means the AI connected things you never understood.

Principle evidence

Strength of the underlying learning idea, not a claim about AI products.

The underlying learning idea is rated moderate. Evidence is stronger for generative note-taking in one's own words than for any branded personal-knowledge system as such.

AI delivery evidence

Whether an AI tutor delivers this method well is a separate question.

AI will happily invent links. Force "none obvious" as an allowed answer; delivery quality hinges on your atomic notes.