Feed a language model a few years of your own voice memos, journal, or notebook. Ask it nine questions in a row, each one following the pull of the last answer. What surfaces is not new information — it's the shape of the inquiry you've been running without noticing. These are the nine moves.
Indexed by topic, an archive flattens — same categories year after year. Indexed by energy, it surfaces the live questions. The first move is to ask the model what's currently unresolved, not what's most frequent.
You'll usually get three to five threads. Some will be obvious to you; one or two will name something you've been circling without admitting.
These four are common shapes. Yours will be more specific. The point of the opening turn is to let the archive choose the gravity, not you.
Archives overflow with mystical or aspirational framings — you, speaking to future-you, in the register of an inner talk. Ask the model to collapse the framing into mechanics, and watch how much of the mysticism was already specification in disguise.
This works for ideas, relationships, and practices. The move is the same: whatever's wrapped in grand language, unwrap it. What does it do?
The answer often comes back surprisingly executable. Concepts you thought were vague turn out to have been drafted with quiet precision over many memos. You just hadn't collated them.
Many of us think in voices — older and younger self, disciplined and playful, the builder and the one who rests. The archive contains those negotiations almost verbatim. Ask for the transcript, not the summary; the shape of the bargain is almost always lopsided.
The mission is what matters. Everything else is indulgence.
There's a time and a place for play. This isn't it.
You will follow my lead.
If I hold you to the mission, you hold me to curiosity. That's the deal.
I'm not the enemy. I'm what keeps us from becoming the thing we built.
I'm asking for reciprocity, not permission.
The illustrative quotes above are hypothetical — any archive that contains a sustained self-dialogue will have its own cast. What stays constant is the asymmetry: one voice demands, the other offers a trade, the trade gets dismissed, and the losing side becomes the subject of a later counter-movement.
The first resolution is rarely the final one. Archives written over years contain self-corrections — quiet ones, usually not flagged as revisions. Ask specifically: where does the later voice complicate the earlier one?
In most archives, whatever gets rejected as indulgence in one voice note returns, six to eighteen months later, promoted from indulgence to mechanism. Play becomes the thing that lets the work land. Rest becomes the discipline. The earlier binary is not refuted — it's simply transcended.
This is why the archive is more honest than any single memo. It contains the walk-back. The dialogue continues after the memo ends.
Across enough notes, the same spatial and physical metaphors reappear. Containers, gateways, conduits, axes, gardens. Ask the model which one organizes the others; the answer is usually a concept you've been circling for years without naming.
Ask the archive to be unkind. Where do you directly contradict yourself — not with small evolutions of view, but with opposing beliefs held simultaneously? The common archetypal contradictions look something like these.
The contradictions above are shapes anyone could hold. Yours will rhyme with some of them. What matters isn't that you resolve them — it's that you see, by name, which pole you actually act from.
Some ideas show up across several recent notes, circling something, not yet landed. These are the most generative threads in any archive — because they haven't collapsed into a fixed position yet.
Ask for them by shape, not name. The model is good at detecting what's outlining itself without a word for it.
These emerging threads often hit the same wall: the act of naming the thing changes it. An object, once defined, loses the verb-like quality that made it worth attending to. The thread is alive because it hasn't been captured yet — and to capture it is to end it.
Archives span time. Over enough of it, beliefs flip. Not revisions — reversals. Ask the model for the biggest one: the position you held strongly that you now reject with equal force.
The common reversal shape is external validation → internal sufficiency, but the specifics vary. Some archives flip the other way — a self-sufficient stance later opens to receiving. Either direction teaches you something the single-voice present-tense can't.
What is the single live question this whole archive is circling?
After eight turns, the archive has absorbed the shape of the inquiry itself. Now ask for the one question the whole corpus is orbiting — one sentence, no answer. The question that comes back often has a characteristic shape: how do I honor X without the very act of holding X destroying the thing about X that matters? Preservation and articulation in tension. Most archives are circling some version of this.