feat(bookclub): daily LLM composer with hard subject-dedup guard — fix the 5× repeated subject

The picker dedups the SECTION against the sent-ledger (never repeats), but the
SUBJECT came from a fixed 13-entry theme list matched first-hit that NEVER checked
the ledger — so "the drift the green dashboard hides" shipped FIVE times
(07-03/07/08/12/16) and read as the same email.

compose-email.mjs (operator 2026-07-16: "the selector should be a prompt that
writes out all the parts, including picking the section"): LOCAL qwen ($0, not
gated by .thetacog/llm-pause — that flag blocks only the paid Claude lane, so this
respects the 07-14 cost cut) picks among pick-passage's dedup-GUARANTEED unsent
candidates (chosen by NUMBER, not by copying a long id) and writes a UNIQUE subject
+ why-today opener. HARD DEDUP GUARD: chosen section ∉ ledger AND chosen subject ∉
ledger, else retry once, else a DETERMINISTIC fallback (top unsent candidate +
title-derived unique subject). The send NEVER blocks on the model and NEVER repeats
— model up or down. Wired into the controller preview; send-daily already carries
pinned.subject_line to the blast, so preview-subject == send-subject.

render-email subjectFor is a second dedup SAFETY NET: even on the rare no-composed-
pin send, the theme scan skips any hook already in the ledger and the fallback is
made unique from the passage title. So NO path can repeat a subject.

Guard: tests/ops/bookclub-compose.test.js — wired · $0/local (no claude) · dedup
guard + deterministic fallback · behavioral: render-email never re-emits a used hook
(hermetic temp ledger via BOOKCLUB_LEDGER override).

Persona-Intent: the daily reader — should feel the email is freshly written for today, never a rerun; a repeated subject silently taxes that trust, so the fix is the never-repeat guarantee, not just a nicer hook.
Originating-Terminal: 🎤 voice
Relevant-Rooms: 🔨 builder, 🧭 navigator, ☕ network
Story: Operator got what read as a repeat book-club email from the voice room; traced it to the subject-generation (navigator/tactics surface) not the section picker, and fixed it at the compose step with a guard so the audience (network) never sees a rerun.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01U2SmLaA9twdunCNda5Rjf6

commit 1ff397309 — 144×144 lattice triptych (on-chip)

◎ the ingest chose A1,A1 — Strategy·Law × Strategy·Law — acting on B2,C1 — Tactics·Deal × Operations·Grid (grip 0.734) · the walk starts at the ACTOR's row · 4 file(s)

⏱ ingest 446.3ms (corpus→lattice) · definer walk 822ms (row→transpose→row, ballistic XOR per hop) · total 1407ms
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) · 6% 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

A1,A1 — Strategy·Law × Strategy·Law (actor) acting on B2,C1 — Tactics·Deal × Operations·Grid (patient) · grip 0.734

actor 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.

patient seed B2,C1 — A deal negotiation sets the exchange rate for power flow through the infrastructure grid and route. Power flows where the grid lets it. Choose connections before topologies; the topology is what the connections compose, and a connection added later is a topology silently rewritten. Latency is geography expressed in milliseconds. The grid IS the map of where compute meets storage meets bandwidth; the map is the territory at the scale the request actually traverses. Infrastructure decisions are reversible at ten-times the cost of being right the first time. Buy the irreversibility down before you provision the easy decisions, because the easy decisions ride on the irreversible ones. A connection between two systems is a contract that outlasts the systems on either end. Wire-protocols outlive the processes they connect; the wire is the substrate, the endpoints are the exhaust. The deal sets the exchange rate between two value substrates. Mis-priced exchange is a slow-bleed; correctly priced exchange compounds both sides, because each cycle of trade enriches the bench on both ends. Terms are the API between two organizations — every clause is a contract method that survives the handshake. Renegotiation is the breaking change; the deal is the stable interface either side can build against. Negotiation closes at the moment both parties' BATNAs cross. Push past that point and you've bought regret at retail; stop short and you've left margin on the table that the next deal will not return to you.

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) · 5% 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 · 6% 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: 0 in the 3×3 ABC corner · 3 in the 9×9 A1..C3 square · 1713 in the 132×132 children square · 222 cross-zone
REALITY: 0 in the 3×3 ABC corner · 9 in the 9×9 A1..C3 square · 1882 in the 132×132 children square · 259 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 21 claims @ θ 0.6094 · reality 69 claims @ θ 0.625 · projection 256ms

SHAPE-MATCH σ = -0.74 — the real intent↔reality definer walk vs random-reality (actual 0.0625 vs random 0.083). >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: C1 · Grid — 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 25% vs tolerance 25%.
624 green · 1073 amber · 0 red · 168 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 72% (no distinctive match)

reality 75% .replace(/^.*?—\s*/, '') // everything up to and including the em-dash prefix

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

intent 67% (no distinctive match)

reality 70% const used = usedHooks();

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

intent 72% feat(bookclub): daily LLM composer with hard subject-dedup guard — fix the 5× repeated subject

reality 78% process.stderr.write('compose: qwen returned no usable JSON/choice — retry/fallback\n'); contin

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

intent 78% (no distinctive match)

reality 75% async function qwen(prompt, { model, temperature = 0.7, num = 260, timeoutMs = 60_000 } = {}) {

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

intent 77% compose-email.mjs (operator 2026-07-16: "the selector should be a prompt that writes out all th

reality 78% # falls back to a deterministic unique subject if qwen is down, so this NEVER blocks the send.

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

intent 73% (no distinctive match)

reality 70% // 1. pick-passage --top N gives the deterministic, dedup-GUARANTEED candidate set (real book t

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

intent 72% (no distinctive match)

reality 78% function parseJsonish(s) {

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

intent 77% (no distinctive match)

reality 77% (no distinctive match)

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

intent 73% (no distinctive match)

reality 80% if (process.env.IM_MODEL) return process.env.IM_MODEL;

C1,C1 Power flows where the grid lets it.

intent 73% (no distinctive match)

reality 73% // normalize a subject to its HOOK for comparison: drop the "📖 Book Club · <date> — " prefix,

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

intent 73% (no distinctive match)

reality 73% // derived from its title/hooks, ledger-checked for uniqueness). So the send NEVER blocks on th

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

intent 77% (no distinctive match)

reality 77% (no distinctive match)

commit context — 4 file(s)

feat(bookclub): daily LLM composer with hard subject-dedup guard — fix the 5× repeated subject

The picker dedups the SECTION against the sent-ledger (never repeats), but the
SUBJECT came from a fixed 13-entry theme list matched first-hit that NEVER checked
the ledger — so "the drift the green dashboard hides" shipped FIVE times
(07-03/07/08/12/16) and read as the same email.

compose-email.mjs (operator 2026-07-16: "the selector should be a prompt that
writes out all the parts, including picking the section"): LOCAL qwen ($0, not
gated by .thetacog/llm-pause — that flag blocks only the paid Claude lane, so this
respects the 07-14 cost cut) picks among pick-passage's dedup-GUARANTEED unsent
candidates (chosen by NUMBER, not by copying a long id) and writes a UNIQUE subject
+ why-today opener. HARD DEDUP GUARD: chosen section ∉ ledger AND chosen subject ∉
ledger, else retry once, else a DETERMINISTIC fallback (top unsent candidate +
title-derived unique subject). The send NEVER blocks on the model and NEVER repeats
— model up or down. Wired into the controller preview; send-daily already carries
pinned.subject_line to the blast, so preview-subject == send-subject.

render-email subj

scripts/bookclub/compose-email.mjs
scripts/bookclub/render-email.mjs
scripts/controller/bookclub-daily.sh
tests/ops/bookclub-compose.test.js

Produced on the chip by the running pipeline (scripts/pmu/pipeline.mjs → runPipeline → walk + xor), commit-scoped · on-chip per-walk 632.97ns (this commit). Recompute: node scripts/pmu/commit-triptych.mjs --commit 1ff397309

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.