AP48 — The Assembly, locked · Frozen Predictions

Locked — 2026-09-04

This is the state of record for Artist's Proof 48, The Assembly. It was verified independently (manifest, two blind computational passes, external-citation spot-checks) before being locked, and locked only once the author confirmed directly — not merely delivered — that he stood behind it, including the part that costs the corpus its own H₀ prediction. Anchored on the420code-proof at tag ap48-locked-2026-09-04. Kill switches KS-ASM.1–3 are armed from this lock.

KS-45.1 fired · 2026-09-03

This paper fires a kill switch on an already-public paper. AP18 ("The Floor") registered H₀ = 74.3 ± 1.2 km/s/Mpc as KS-45.1. AP48 derives H₀ = 67.45 km/s/Mpc from the corpus's own structure — 5.7σ from AP18's figure, on its registered width. An erratum corrects that width (it omitted a systematic uncertainty; honestly stated it is ±14.9) but the correction stands beside the fired switch, not instead of it — re-sizing a live switch's bar after the corpus's own derivation has already contradicted it, then not counting the fire, is exactly the repair this house refuses. In the paper's own words: “The distinctive prediction the corpus used to carry on H₀ — a value above both camps — is gone.”

Under the standard equation, an age of 13.83 needs a rate of 67.45, and a rate of 74.3 gives an age two lane-times below the cycle. AP46 pre-registered that AP48 would close this — by deriving a non-standard relation, or by one edge firing. It closed by adjudication, not reconciliation. This paper does not save both of AP46's numbers.

What's being closed

AP46 held two things at once: a one-cycle age of the universe (13.830 Gyr, from the corpus's own α⁻¹⁸ construction) and the corpus's own Hubble constant (74.3 ± 1.2, from AP18's floor). Under the standard age–rate relation, a rate of 74.3 gives an age of only 12.55 Gyr — two lane-times short of the cycle. Something had to give. AP48's finding: the age is right, the rate was wrong.

The rate (once, from CODATA 2022)

0.954H₀t₀, closed form, Ωₘ = 0.3115

67.45km/s/Mpc, the derived rate

64.4–70.8the window, one lane-time either side

5.7σKS-45.1's firing, on its registered width

H₀t₀ = (2 / (3·√Ω_Λ)) · arcsinh(√(Ω_Λ/Ω_m))
0.954 × 70.70 km/s/Mpc (= 1/cycle) = 67.45 km/s/Mpc

No fitted parameters: the cycle (13.830 Gyr) is the same construction AP46 already locked; the standard H₀t₀ relation is textbook flat-ΛCDM; Ωₘ = 0.3115 is not this paper's to choose — it is named from the corpus's own inputs. One integral, one number.

Against measurement: the relic glow (Planck 2018, 67.4 ± 0.5; with BAO, 67.66 ± 0.42), the Chicago–Carnegie red-giant ladders (Freedman et al. 2025, TRGB-alone 70.39 ± 1.22 ± 1.33 ± 0.70; JWST-only TRGB 68.81, JAGB 67.80) all sit inside or at the edge of the window. The Cepheid ladder (Riess et al. 2025, 73.49 ± 0.93) and the SH0ES red-giant values (Li et al. 2026, 72.1–73.3) sit outside it. This paper's stance: the Cepheid ladder's ~73 is not the expansion rate. Either its systematics are found, or this paper dies on KS-ASM.1 — possibly within a few years of JWST data.

The floor, re-read

AP18's own coefficient — a₀ = 1.0445 × cH₀/(2π) — is untouched. What was wrong was its registered width: AP18 quoted only the random error (±0.02×10⁻¹⁰) on the empirical acceleration scale, when the sources it cites also carry a systematic (±0.24×10⁻¹⁰) from stellar mass-to-light ratios and distances. Honestly propagated, AP18's own inversion is H₀ = 74.3 ± 14.9, not ± 1.2 — wide enough that the floor's relation survives comfortably (0.46σ) even as its narrow registered bar does not.

This is the distinction the whole finding turns on: the relation lives: the registered bar it was tested against does not, and a switch fired on the bar it was actually registered at stays fired even when a wider, more honest bar would have let it live.

Kill switches

Switch	Claim	Status
KS-ASM.1	The rate: 67.45 km/s/Mpc, window 64.4–70.8	LIVE — EMPIRICAL
KS-ASM.2	The balance: flat, thinnings of four and three, no additive ε term	LIVE — STRUCTURAL/EMPIRICAL
KS-ASM.3	The roof before the nuclei: constants steady before the first second	LIVE — EMPIRICAL
KS-STRETCH.4	The era surface (AP46's, registered for this paper)	MET · conditional on KS-ASM.2
KS-45.1	AP18's H₀ = 74.3 ± 1.2 (already public)	FIRED 2026-09-03

If the rate lands above the window (the Cepheid ladder confirmed), KS-ASM.1 fires and flat ΛCDM at the sky's partition fails with it — the corpus claims no advantage there either way. If below, KS-STRETCH.3 (AP46's) fires instead and AP46's chain reverts to a candidate.

What this paper does not claim

Does not claim the rate to better than its window, the awareness window, the wobble's spectrum, or the eye loop.
Does not claim the storm's length or the count that completes the assembly — both open, named as the author's own drawer (D-ASM.1), not blocking this lock.
Adds no energy-budget term of any size. Does not touch α's value.
Does not repair or reweigh anything fired, and does not repair KS-45.1.

Run it yourself

python3 verify_ap48.py

Reproduces the H₀t₀ closed form, the AP46 cycle, the derived 67.45 km/s/Mpc rate and its window, the 5.7σ KS-45.1 firing, and the corrected AP18 floor check — from CODATA 2022 alone, stdlib only. It reports KS-45.1 as fired every time it runs; that's the check working, not a bug.