Precision as a transferable stance

Get the reference right. Build on it without loss.

A workbench of tools that carry precision, worked to a high finish in fields that rarely talk to each other, into work that usually goes without it: coordination, evaluation, and meaning.

These tools establish what is actually there, the verified floor. They do not tell you what it is worth. That judgment stays yours.

Gauge blocksA gauge block is a certified length, accurate to millionths of an inch. Stacked up from a reference surface, blocks build any dimension without accumulating error. They were precision's original reference standard in machining, and a literal picture of the idea here: get the reference right, then build on it without loss.
Start with your problem

What are you trying to do?

No problem to pick yet?

Just make your AI work this way.

One standing instruction that changes how your assistant works: reach for missing context, build the case that breaks its own answer, keep the verdict yours.

Pick the one that fits. It opens the tool, the fit test, and a prompt you can run.
Tool
The three-plate method. Take three or more genuinely independent reads, grind them against each other (test each against the others) across the whole object; the convergence is the verified floor, the divergence is the contested map.
Fits when
A verified floor would settle your question, you can get three reads that differ at the root, and you do not need the tool to tell you what the thing is worth. If a standard already exists, inherit it instead.
Inside AI
The grind runs to the floor, with one hard guard: the three reads must be genuinely independent, not one model's restatements. The worth-judgment stays with you.
Here is something I need to evaluate, and there is no standard to measure it against: [PASTE].

Give me three genuinely independent reads of it, each from a different starting frame, not three restatements of one view. Grind them against each other across the whole thing. Report only what all three converge on as the verified floor, and separately map where they diverge as the contested territory. Do not tell me what it is worth. Leave that judgment to me.
Tool
The sharper Occam's razor, riding on the frame-primacy grind. A thing is excess only if a sharp rival frame exposes it as a preference and a lossless-reconstruction test forecloses over-cutting.
Fits when
You can hold a genuine rival frame, sharper than a generic objection, and you can state what lossless reconstruction means for your object so the razor does not amputate real structure.
Inside AI
The grind and the cut run to the minimal basis. Stop at the first round that brings no new independent surface; past the basis it manufactures false divergences. The stop is a rule, not a feeling.
Here is a claim or spec I want cut to what is load-bearing: [PASTE].

Hold a genuine rival frame against it, sharper than a generic objection. Cut only what that rival exposes as my preference rather than real structure. Stop the moment a lossless-reconstruction test says any further cut would lose something real. Show me what is load-bearing, what was excess, and where you stopped and why.
Tool
The Dimensional Frame Language diagnostics. Place each candidate frame at its level, grind the rival decompositions to their shared basis, then name that basis in terms that carry their meaning.
Fits when
The terms are absent or contested and the decomposition matters.
Inside AI
The placement, the grind, and the vocabulary check run inside AI. The check that you are not imposing structure the domain does not have stays with you.
I need to name or decompose a domain that has no agreed vocabulary: [DOMAIN].

Place each candidate frame at its level, so we are not arguing a surface claim against a structural one. Grind the rival decompositions down to the basis they all reduce into. Name that basis in terms that do not smuggle in unearned weight. Flag anywhere we might be imposing structure the domain does not actually have, and leave that call to me.
Tool
The Dimensional Frame Language vocabulary discipline. It flags terms that import unearned weight, latent claims of authority, deference patterns, and quiet substitutions, and offers the operational replacement.
Fits when
There are terms in a specification, a claim, or a document whose wording matters.
Inside AI
Fully. This is the most automatable tool in the kit, and it already runs live in the public Dimensional Frame Language analyzer and the open-source language server.
Check this text for terms that import unearned weight: latent claims of authority, deference patterns, quiet substitutions, or words borrowing meaning they have not earned: [PASTE].

For each flagged term, name what the term is quietly claiming, and give the operational replacement that states plainly what it does.
Tool
The three-plate method, with independent frames as the plates. Three genuinely different frames, ideally different base models, grind the result; the floor is what survives all of them.
Fits when
You can assemble three genuinely independent checks. It fails worst when the three share a base model or a training, because then they are one plate and their agreement is an artifact.
Inside AI
This is the case that runs inside AI most, and is most often gotten wrong, because independence is the whole game and the hardest thing to get from one system. The acceptance decision stays with you.
I have an AI output I cannot check against any oracle: [PASTE].

Assemble three genuinely independent checks of it, ideally different base models or genuinely different approaches, not one model restated three times. Report only what survives all three as the floor, and where they disagree. Tell me if the three are not truly independent. The accept-or-reject call is mine.
Tool
The gauge-block principle and composition from a minimal basis. Precision does not transfer between layers for free, so you carry it deliberately, building every form as a configuration of declared primitives.
Fits when
A standard already exists. This is the companion to the three-plate method: where that one bootstraps a reference because none exists, this one carries a reference you already have.
Inside AI
The composition is mechanizable.
I already have a standard and need to carry it into new forms without losing precision: [STANDARD].

Build every form as a configuration of declared primitives from a minimal basis, so precision does not leak between layers. Flag anywhere a form is not fully reducible to the declared primitives, because that is where the precision is being lost.
Tool
The disconfirmation practice construct. The standing instrument that wraps the grind with an append-only registry, a revert when a round breaks nothing, and the dry-round stop as a gate a machine can read.
Fits when
The grind is recurring and its auditability matters. It is built for the AI context, because it routes its own residue, the part it cannot settle, out to a human or a different model.
Status
This is a design and build plan, not yet a built tool. The mechanism is specified; the instrument is not yet standing.
The thinking behind it

Not a method to adopt whole. A workbench to reach into.

It is not a theory you have to believe. The tools are built to combine, so reaching for one tends to set up the next. They come from a single move: taking precision, the discipline of getting the reference right and then building on it without loss, and reinstating it where the material is softer. That shared source is why the tools fit together rather than sitting as a pile.

01 · Precision first

Get the reference right before you measure.

Most failures are not measurement errors. They are missing or merely local references.

02 · Disconfirmation

Grind it against something built differently.

Convergence between independent accounts is the signal; a single confident account is the thing to distrust.

03 · Frame awareness

Know which frame you are in.

Surface, structure, and whole are three different stances. Most confusion is a claim made in one and read in another.

04 · Honesty about where it stops

Establish the floor. Hand back the verdict.

A tool that pretends to deliver the verdict is the failure these are built against.

See how the pieces fit together
REACH THE FLOOR CONFORMANCE · MACHINE KEEP THE VERDICT JUDGMENT · PERSON the hand-off WHERE THE FLOOR IS SET, NOT THE WORTH
The bound pair: let the machine reach the floor, keep the verdict human
EVALUATE WITHOUT A STANDARD Grind to floor Cut the excess Hold the terms Trace the chain NAME A DOMAIN WITH NO VOCABULARY Place by level Grind to basis Name the basis WORTH- JUDGMENT STAYS HUMAN
The value is the chain: both worked chains end where the verdict stays yours
Adopt the vein, not the fingerprint

These are the tools this way of working was built with. They are one person's fingerprint. What transfers is the obligation they discharge: getting the reference right, grinding against independence, keeping the verdict human. You can discharge that obligation with these tools or with your own.

The released toolkit is the structural showing. The personal account of why this way of working came to be is held privately and is not part of the release, named where it is referenced so the structure is honest about its own edges.