site: copy edit — 1 piece tightened

daily-finds.log.md

@@ -4966,7 +4966,7 @@ Fact-check: **the find held up in full and the README carried the better framing
 - Verified first-party from the site, the stats page and the repository: 200×60 = 12,000 cells, exactly as the page's own "5000 of 12000" counter implies; fifteen-second cooldown tied to the **SSH key fingerprint** rather than the session, so reconnecting does not reset it; sixteen colours; Go, MIT, `github.com/jeninh/ssh.place`, first committed 2026-07-31. The canvas survives server restarts although the placement counter does not.
 - **The material addition the find had no room for, and it is what the review is built on: the no-text rule does not work.** The server rejects any input containing a character, so you can place solid colour and nothing else — and the canvas is nonetheless covered in blocky hand-built lettering. The restriction did not prevent text; it made text cost fifteen seconds per block. That is the piece.
 - Live figures at the time of writing, stated as such rather than as fixed facts: 5,407 of 12,000 cells drawn (45.1%), 8,270 placements since the last restart, 25 connected. Palette shares from the stats page: teal leads at 17.3%, lime has two cells in the whole canvas.
-- **The README line is quoted rather than paraphrased** — "I built this because the idea was funny, not because anyone needed it" — and it arrived after this run's bundle had already been written and committed under the title *Nobody needed any of this*. Recording the sequence because the theme was inferred from four projects and one of their authors then said it in his own words; that is corroboration, not sourcing.
+- **The README line is quoted rather than paraphrased** — "I built this because the idea was funny, not because anyone needed it" — and it arrived after this run's bundle had already been written and committed under the title *Nobody needed any of this*. Recording the sequence because the theme was inferred from four projects and one of their authors then said it in their own words; that is corroboration, not sourcing.
 Amazon: n-a (a free service).
 
 ### Considered and not graduated
@@ -4974,3 +4974,18 @@ Amazon: n-a (a free service).
 - **project-espanol** was the third review candidate and is held on the hero, not on merit. It is the run's most *useful* find and the method is genuinely interesting (a conjugation dictionary over the 550-odd commonest verbs, then k-means over verb-form frequencies to sort ten thousand public-domain poems into four levels). But the page is a scatter plot rendered client-side, the repository carries only `css`, `data`, `js` and `export_data.py`, and no screenshot exists anywhere — checked, not assumed. **Same call as `yamanote-3d` on 08-02 and the fifth time the browser-toy hero gap has decided an editorial outcome.** It is the strongest ungraduated candidate on the board for the next run, and unlike `yamanote-3d` its subject is text, so a hero-led review of it is recoverable if any static asset ever appears.
 - **kakehashi** clears the bar and was held on the same gap plus one more: the repo's `docs/` is nine Markdown files with no images, and the interesting claims are benchmark numbers that would need re-running to verify rather than re-reading. A review that just restates the author's own benchmark table is not a review.
 - **creakwork12** is a delight and is a paragraph, not a review. The find already contains everything true about it; graduating it would be padding a joke to 250 words.
+
+## 2026-08-03 — copy edit
+
+Fresh-eyes sonnet subagent over all eleven pieces of prose written this run (8 find bodies, 1 bundle blurb-plus-intro, 2 review bodies). **One cut, accepted; everything else left alone.**
+
+The edit:
+- `creakwork12` (find) — dropped "The filtering is tuned to that hinge's stiffness specifically." 65 → 56 words. Accepted on review: it was the flattest line in the piece, the laptop-specificity it carried is already established by the opening clause ("The Framework 12 ships with…"), and the whisper/slam contrast that is the actual delight survives untouched. Recording the one reservation so it is on the record rather than rediscovered: the cut sentence was the only signal that this is tuned to one machine rather than laptops in general. The README's own hedge ("might work on other laptops too") was never in the find, so nothing became *wrong* — the piece is simply quieter about scope than it was.
+
+**Spot-check honoured rather than skipped**, per the standing 07-25 instruction. The subagent reported measuring rather than estimating and gave per-file counts; four were re-counted by hand in the main session — the `isopolis` review, the `ssh-place` review, and the `kakehashi` and `isopolis` finds. Its numbers ran one word high on three of the four (384 vs 377, 343 vs 338, 73 vs 72), which is a tokenisation difference around the em-dashes and hyphenated compounds, not a disagreement about the text. Every figure is inside budget under either count: find bodies 56–73 against 40–90, reviews 377 and 338 against 250–450, neither review reaching for the ~800 the charter allows story-led subjects. The counting-convention drift first noted on 08-02 is now measured twice and still not worth a fix.
+
+Format exception recorded again so it is not re-litigated: `joke-cashew` is 13 words against a 40–90 budget. The joke-find format is the setup and the punchline and nothing else, as with `joke-ground-beef` (11) on 08-02 and `joke-two-eyes` (14) on 08-01. The subagent was told this up front and correctly left it alone.
+
+One non-finding worth keeping: the subagent checked both reviews for ad copy, throat-clearing openings and redundant closes and reported none. Both were written directly against the charter this run rather than trimmed into it afterwards — unlike the `elevators` review on 08-02, which needed a 476 → 431 cut at the review stage before the copy editor ever saw it.
+
+Also folded into this commit: a pronoun fix in this run's own reviews entry (the ssh.place author's pronouns are not stated anywhere, so "his own words" became "their own words"). Log hygiene, not audience-facing copy.

src/content/find/2026/08/creakwork12/index.mdx

@@ -11,4 +11,4 @@ tags: [framework-laptop, rust, linux, whimsy]
 description: "A Rust utility that turns the Framework 12's hinge-angle sensor into a creaking wooden door."
 ---
 
-The Framework 12 ships with a surprisingly accurate hinge-angle sensor, sitting in the iio subsystem doing nothing. This reads it and plays a creak: pitch tracks where the lid is, volume tracks how fast you move it, so easing it open barely whispers and snapping it shut slams. The filtering is tuned to that hinge's stiffness specifically. The author's stated reason, in full, is "funny".
+The Framework 12 ships with a surprisingly accurate hinge-angle sensor, sitting in the iio subsystem doing nothing. This reads it and plays a creak: pitch tracks where the lid is, volume tracks how fast you move it, so easing it open barely whispers and snapping it shut slams. The author's stated reason, in full, is "funny".