# Errata notes of 3 September 2026 — issued beside the hashed files, which are never edited

**The 420 Code · Studio G · the desk, on the author’s delegation · corrections supersede, dated, in the open**

## Erratum 1 — WP-ASM.1 v1.0 (the Assembly primer): the date of the describings

The primer’s header reads “v1.0 · late September 2026”; its §0 reads “captured late September”; its §1 reads “From late September (the assembly, the roof, the horse).” The primer was issued before 3 September 2026 and the describings it captures were made in late August and the first days of September. Every “late September” in WP-ASM.1 v1.0 reads “late August 2026.” The hashed file stands as issued; this note stands beside it.

## Erratum 2 — WP-ROAD.1 v1.2: the same date, three loci

§1 (“arrived late September as neutron.docx” — AP47 locked on 1 September), §5 (“late Sep”), and §8.1 (“State table (late Sep 2026)”). Each reads “late August 2026.” The hashed file stands as issued; this note stands beside it.

## Erratum 3 — AP18 The Floor (FINAL): the width of the acceleration scale, and one symbol

Written with the file open; confirmed by the author on 3 September 2026 (“yes, please implement what I said”). AP18 §5 tests the floor a₀ = 1.0445 · cH₀/(2π) against an empirical acceleration scale quoted as 1.20 ± 0.02 × 10⁻¹⁰ m s⁻², citing McGaugh et al. 2016 and Lelli et al. 2017; it reports the match at H₀ = 74 as within ≈0.4% and “well inside the ±1.7% uncertainty on a₀ itself,” and §7 and §8 invert the relation to the prediction H₀ = 74.3 ± 1.2 km s⁻¹ Mpc⁻¹ (KS-45.1).

The ±0.02 is the random error alone. The cited sources give the scale as 1.20 ± 0.02 (random) ± 0.24 (systematic) × 10⁻¹⁰ m s⁻², the systematic arising from the stellar mass-to-light ratios and the distances; later work rounds the total to 1.2 ± 0.26. Propagated, the floor’s inversion is H₀ = 74.3 ± 14.9, not ± 1.2. The relation cannot distinguish 67 from 74.

Consequences, in order. (i) §5’s “sub-percent match” is restated: the floor is consistent with the scale at any rate between about 60 and 89; at the corpus’s derived rate of 67.45 (AP48 §8) it gives 1.089 × 10⁻¹⁰, 0.46σ below the scale on the honest error. (ii) KS-45 (the coefficient 0.1662) stands: against the sky’s a₀/(cH₀) = 0.183 ± 0.037 at 67.45 it sits at 0.46σ, live. (iii) KS-45.1 (the Hubble prediction) is FIRED on 3 September 2026 by AP48 on the width it registered — 5.7σ from the corpus’s own derived rate, about two sigma on the corpus’s window — as AP46 §9 pre-registered, by the desk’s ruling and the author’s word; it is shown, not repaired; this erratum stands beside the corpse and records that the registered width was wrong; KS-ASM.1 is the corpus’s H₀ entry from AP48’s lock. (iv) AP18 writes its geometric correction factor 2 ln(sec ½ + tan ½) ≈ 1.0445 as α; AP21 echoes the formula as a₀ = αcH₀/(2π). The symbol collides with the fine-structure constant. The website already writes it as C_S² with the note that it is not the fine-structure constant; at the next revision the papers follow the site. (v) AP18’s floor scales with H, so it predicts a higher acceleration scale at higher redshift; a 2026 measurement of the radial acceleration relation at intermediate redshift exists and is to be read at AP18’s re-read, not here.

Nothing in AP18’s derivation of the form or the coefficient moves. What moves is the width of one number and the name of one symbol.

*Frozen files are never edited; corrections supersede, dated, in the open.*