feat(pmu): phase 2 of the ShortLex registry — name the 132 deep sub-categories, no LLM

Operator: "the sub categories need to be named... use on-chip semantic-ish
to pick a word... the most meaningful word... without an LLM... use
statistical measures... with the gzip sensor." The registry
(data/pmu/shortlex-144-registry.json) has always had this documented as
"phase 1" -- 144 entries (3 top-level, 9 mid-level, 132 depth-3 children
like A1A..C3N), every `semantic` field empty, with an explicit comment
that phase 2 was still owed.

Found the running code first rather than reinventing: the 132 children
already have real, mined corpus text (`dump`, ~800-900 chars each) in
data/pmu/shortlex-children-candidate.json, produced by an existing
bottom-up derivation pipeline (derive-children.mjs) that already computes
a self-vs-other-family distinctiveness gate at the whole-dump level
(meanSib vs maxNonFam). And the canonical gzip-NCD sensor already exists,
exported, as gzipLen/ncdSim in src/lib/pmu/unified-drift.mjs -- reused
directly, not reimplemented.

New: scripts/pmu/shortlex-name-children.mjs applies the SAME self-vs-
other-family comparison one level deeper, per candidate WORD instead of
per whole document: for each child, tokenize its dump (stopword-filtered,
camelCase-split so concatenated identifiers resolve to real words), score
each candidate by (compression-similarity to its OWN parent's dump) minus
(max compression-similarity to any of the 8 OTHER parents' dumps), and
keep the most distinctive word. Deterministic, reproducible, zero LLM
calls anywhere on this path -- confirmed by construction, not by grep,
since the script only imports node:zlib's gzip and the two pure-function
sensor exports.

Ran with --write: 132/132 children now carry a real word (e.g. A1A ->
"professional", A1I -> "metaphorical", A1L -> "architecture") sitting
next to their gibberish-looking codes -- the code anchors the coordinate,
the word anchors the meaning.

Guard: tests/pmu-simulator/shortlex-children-named.test.mjs -- asserts
all 132 are non-empty, asserts the words aren't degenerately identical
across children (a real distinctiveness check, not just "non-empty").

Originating-Terminal: 🎤 voice
Relevant-Rooms: 🎤 voice, 🧪 laboratory, 🔨 builder
Story: Direct build-out of a documented, long-standing "phase 2 TODO" in
the registry's own comment -- the candidate corpus and the sensor both
already existed; the missing piece was the word-selection method itself,
built to reuse the pipeline's own established evaluative pattern
(self-vs-other-family) rather than inventing an unrelated scoring scheme.
Persona-Intent: A future session reading a region label like "A1I ·
metaphorical" should understand that word was picked because it's
compression-central to A1I's own mined content and NOT to any of the 8
sibling-family territories -- not an LLM's guess, a measured distinctiveness
score they can recompute from the same two JSON files.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017NvY8cDHg42UqCsWmqnMBB

commit 54a13efbc — 144×144 lattice triptych (on-chip)

◎ the ingest chose A2,C3 — Strategy·Goal × Operations·Flow — acting on A1,A1 — Strategy·Law × Strategy·Law (grip 0.813) · the walk starts at the ACTOR's row · 3 file(s)

⏱ ingest 3823.3ms (corpus→lattice) · definer walk 3713ms (row→transpose→row, ballistic XOR per hop) · total 9509ms
ingest here = commit-scoped SENSING only (this commit's message + changed files → lit anchors) — deep seed authoring lives in the reef-self-loop, off the commit path

rows = ACTOR (lens) · cols = PATIENT (object) · diagonal = self-reference; ◎ at actor-row × patient-col — every panel, same orientation. Heat above the diagonal = actors ranking ShortLex-earlier than their patients (the walk follows definers uphill in rank).
The 12 canonical lanes (rows top→bottom, cols left→right): A·Strategy · B·Tactics · C·Operations · A1·Law · A2·Goal · A3·Fund · B1·Speed · B2·Deal · B3·Signal · C1·Grid · C2·Loop · C3·Flow

PAIR LATTICE (the map we normally see)

PRE-WALK SENSE · INTENT grid — the snapshot, no point of view

the no-point-of-view snapshot — what the sensor lit before any walk; the leaf walk below is what bridges the two grids

	ABC A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N A1–C3
ABC A1–C3 A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N

PRE-WALK SENSE · REALITY grid — the snapshot, no point of view

the no-point-of-view snapshot, reality side — same sensor law

	ABC A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N A1–C3
ABC A1–C3 A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N

PRE-WALK Δ · straight comparison of the two ingested grids (green both · cyan intent-only · amber reality-only) · 4% overlap · ◎ = where the competence pixel sits in the raw overlap

good = green overlap without being a copy — near-100% is self-similarity, not alignment

	ABC A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N A1–C3
ABC A1–C3 A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N

◎ the pixel this commit belongs to

A2,C3 — Strategy·Goal × Operations·Flow (actor) acting on A1,A1 — Strategy·Law × Strategy·Law (patient) · grip 0.813

actor seed A2,C3 — Aiming for a quarterly goal target establishes the position that flow throughput and delivery must reach. Flow is the rate at which committed work crosses the finish line. Inventory accumulates upstream; flow is the only thing the customer pays for, and the only thing the operator can measure honestly. Movement is the irreversible delivery of value across the boundary. Anything stationary is cost; anything moving is either revenue, deferred revenue, or the exhaust of revenue not yet realized. Throughput dominates utilization. A ninety-nine-percent utilized system at low throughput pays the same wages as a fifty-percent utilized system at twice the throughput, but the second one ships the customer's order. The target is the cell coordinate the system MUST occupy by end-of-quarter. Anything that does not drive toward that coordinate is exhaust, and exhaust does not become substrate by being labelled progress. Vision is the coordinate the lattice rotates around — the fixed bearing every iteration is measured against. Without the fixed point, motion is movement; with the fixed point, motion is convergence. Define the goal as a verifiable predicate, not a feeling. The system terminates when the predicate evaluates true, and the predicate is what the next agent inherits when this one hands off. A goal that cannot be tested is a wish. Wishes have no termination condition, so they consume budget forever; goals have a finish line and therefore a payable price.

patient seed A1,A1 — EU AI Act Article 14 requires human-on-the-loop oversight, with the supervisor able to interrupt the system at any moment. Compliance attaches to the artifact the artifact cannot escape — provenance is a regulatory primitive, not a feature. Constraints precede objectives. The cap-table, the license text, the consent envelope — each draws the inviolate line the action must respect, and the line is enforced before the optimization runs. A rule names what the system MAY NOT do. If the agent can do X, then 'forbid X' is not the constraint — it is wishful thinking dressed as policy. Constraints live in the gate, not in the prompt. Regulation is not advisory. The fine is the price tag stapled to the omitted control; insurance is the option pricing on the failure-mode the regulation already named.

INTENT · leaf-walk clouds · BLUE→VIOLET by ply (◎ = the competence pixel: actor∩patient)

good = the clouds concentrate where the commit SAYS it works

	ABC A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N A1–C3
ABC A1–C3 A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N

REALITY · leaf-walk clouds · ORANGE→RED by ply (◎ = the same pixel — the shared perspective)

good = the same clouds as intent — shipped where declared

	ABC A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N A1–C3
ABC A1–C3 A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N

DELTA · graded cloud divergence (green agree · magenta declared-not-done · amber done-not-declared) · 3% agree

good = mostly green; magenta = said-not-done (chase it in the code), amber = done-not-said (chase it in the docs)

	ABC A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N A1–C3
ABC A1–C3 A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N

TOLERANCE · cloud topology (dim green in-shape · hot amber bleed · red orthogonal concentration)

red only matters above the flip — a few amber is normal bleed

	ABC A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N A1–C3
ABC A1–C3 A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N

SHORTLEX-3 PROJECTION (ABC · A1..C3 · the 132 children — candidate, pre-ratchet)

PROJECTED INTENT · the commit’s claims, head→axis_i tail→axis_j (cyan)

the three-length view — good = claims land inside their own zone

	ABC A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N A1–C3
ABC A1–C3 A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N

PROJECTED REALITY · same projection law (amber)

same projection law — good = reality occupies the zones intent does

	ABC A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N A1–C3
ABC A1–C3 A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N

STRAIGHT COMPARISON · green both · cyan intent-only · amber reality-only · 4% overlap

good = the zones agree; cross-zone scatter is the drift to read

	ABC A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N A1–C3
ABC A1–C3 A1A–O A2A–O A3A–O B1A–O B2A–O B3A–O C1A–N C2A–N C3A–N

shortlex-3 zone occupancy (lit cells per diagonal square)
INTENT: 1 in the 3×3 ABC corner · 2 in the 9×9 A1..C3 square · 829 in the 132×132 children square · 137 cross-zone
REALITY: 1 in the 3×3 ABC corner · 7 in the 9×9 A1..C3 square · 1193 in the 132×132 children square · 318 cross-zone
axes: identical 144 ShortLex names both sides (symmetric); weights asymmetric
zone 3 = candidate children, pre-ratchet (repo-derived dumps not yet past the perturbation-probe gate) · intent 45 claims @ θ 0.6406 · reality 33 claims @ θ 0.625 · projection 6964ms

SHAPE-MATCH σ = -2.29 — the real intent↔reality definer walk vs random-reality (actual 0.0171 vs random 0.094). >0 = the walk distinguishes real alignment from noise; aggregate over independent walks → the divergent series.
σ_drift · noise · below zero — this read is noise, not a measurement

Most out-of-lane row: B3 · Signal — that lane is drifting most (read the red band in that row).

Three-colour tolerance (out-of-lane = reality fires where intent is weak, in orthogonal blocks): green=in-lane agreement · amber=a few out-of-lane (tolerated) · red=armed — orthogonal out-of-lane 5% vs tolerance 25%.
215 green · 248 amber · 0 red · 300 XOR-friction cells.

Both axes = the 144 ShortLex anchors, cell=row⊕col, on the chip. INTENT=docs+rules · REALITY=code · DELTA=friction.

tile dump — what the SimHash matched on the diagonal (inspect the sensor; tap to expand)

A,A The lattice we build today is what next year's underwriter inherits.

intent 75% (no distinctive match)

reality 77% // pipeline already uses at the whole-dump level (meanSib vs maxNonFam in the candidate file's

B,B Tactics is the choice of which leverage to pull at which beat.

intent 69% (no distinctive match)

reality 73% // split camelCase BEFORE lowercasing (e.g. "fullyApproved" -> "fully Approved") so a concatena

C,C Operations is the loop that runs whether anyone watches.

intent 78% (no distinctive match)

reality 69% (no distinctive match)

A1,A1 EU AI Act Article 14 requires human-on-the-loop oversight, with the supervisor able to interrupt the system at

intent 77% (no distinctive match)

reality 78% // slug that leaked into the dump, not a word; excluding hyphens lets the tokenizer split it in

A2,A2 The target is the cell coordinate the system MUST occupy by end-of-quarter.

intent 81% Guard: (1) the 132 depth-3 ShortLex children have a non-empty `semantic` word in the registry —

reality 73% // Operator (2026-07-06): "the sub categories need to be named... use on-chip semantic-ish to p

A3,A3 Every dollar builds either the floor (substrate that lifts future bets) or the ceiling (cap on the next round)

intent 67% (no distinctive match)

reality 69% import { gzipLen, ncdSim } from '../../src/lib/pmu/unified-drift.mjs';

B1,B1 Latency is the price of a missed beat.

intent 70% (no distinctive match)

reality 73% (no distinctive match)

B2,B2 The deal sets the exchange rate between two value substrates.

intent 70% (no distinctive match)

reality 72% (no distinctive match)

B3,B3 Signal-to-noise is the inverse cube of bandwidth — narrow the channel, the message arrives; broaden it, recipi

intent 69% Guard: tests/pmu-simulator/shortlex-children-named.test.mjs -- asserts all 132 are non-empty, a

reality 75% const ownDump = candidates.parents[parentKey]?.dump || '';

C1,C1 Power flows where the grid lets it.

intent 73% (no distinctive match)

reality 70% // THE METHOD (statistical, not an LLM):

C2,C2 Each iteration is a hypothesis tested against the ground truth of the world.

intent 69% (no distinctive match)

reality 70% (no distinctive match)

C3,C3 Flow is the rate at which committed work crosses the finish line.

intent 77% (no distinctive match)

reality 69% (no distinctive match)

commit context — 3 file(s)

feat(pmu): phase 2 of the ShortLex registry — name the 132 deep sub-categories, no LLM

Operator: "the sub categories need to be named... use on-chip semantic-ish
to pick a word... the most meaningful word... without an LLM... use
statistical measures... with the gzip sensor." The registry
(data/pmu/shortlex-144-registry.json) has always had this documented as
"phase 1" -- 144 entries (3 top-level, 9 mid-level, 132 depth-3 children
like A1A..C3N), every `semantic` field empty, with an explicit comment
that phase 2 was still owed.

Found the running code first rather than reinventing: the 132 children
already have real, mined corpus text (`dump`, ~800-900 chars each) in
data/pmu/shortlex-children-candidate.json, produced by an existing
bottom-up derivation pipeline (derive-children.mjs) that already computes
a self-vs-other-family distinctiveness gate at the whole-dump level
(meanSib vs maxNonFam). And the canonical gzip-NCD sensor already exists,
exported, as gzipLen/ncdSim in src/lib/pmu/unified-drift.mjs -- reused
directly, not reimplemented.

New: scripts/pmu/shortlex-name-children.mjs applies the SAME self-vs-
other-family comparison one level deeper, per candidate WORD instead

data/pmu/shortlex-144-registry.json
scripts/pmu/shortlex-name-children.mjs
tests/pmu-simulator/shortlex-children-named.test.mjs

Produced on the chip by the running pipeline (scripts/pmu/pipeline.mjs → runPipeline → walk + xor), commit-scoped · on-chip per-walk 2866.05ns (this commit). Recompute: node scripts/pmu/commit-triptych.mjs --commit 54a13efbc

Commit artifacts (open on a computer — Gmail mobile won't follow file:// links; the inline images above are the phone-viewable form): reef grid · this triptych · the HDL — all three are also attached to this email.