USR-2026-0035 · Discreteness

No floating-point artifacts in nature: naive machine-precision arithmetic as the substrate of physics is ruled out.

PROVISIONAL null result — naive floats ruled out re-run & confirmed 2026-07-09 (R0 — founder CI; independence pending)

If reality computed with naive finite-precision floats, characteristic error signatures would appear in precision physics. The seam finds none: a clean null stated as a bound on that implementation class.

Falsify-box — how to kill this claim

SURVIVED DEAD on detection of an unexplained, systematics-surviving relative drift or conservation-law violation of magnitude >= 1.1e-16 (the float64 machine epsilon the claim's own script uses as its reference precision) in any of the seam's own precision channels - optical-clock frequency comparison, kaon CPT |delta m|/m, electron g-2 - with round-off's diagnostic signature of growth as sqrt(N) in elapsed operations/time rather than a constant offset. The claim's own reverse-bound premise (eps*sqrt(N_ops) with N_ops = 1e120 must stay below reality's 1e-18) simultaneously fixes the scope: it excludes naive fixed-precision arithmetic only down to eps ~ 1e-78 (~259 mantissa bits), so any artifact detected above 1.1e-16 kills the claim outright.

Verification record — every quoted number, re-run 2026-07-09

quantitymeasuredverdict
precision-artifact verdict NULL (naive floats ruled out) MATCH

Provenance

script: precision_seam/run_precision_seam.py
script status: PRESENT
datasets: Precision-physics comparison battery · claim-scripts (downloadable from /data — run it yourself)
re-run: 2026-07-09

Status, honestly

R0 pass 2026-07-09 (null result) | kill test added by the 2026-07-26 sweep. Status here is computed from evidence — the author cannot set it, and neither can we. Independent reproduction would move it; nothing else will.

§ Figures 2 from precision_seam/run_precision_seam.py

What this claim's own script draws. These are the founder's committed outputs — the same plots the script regenerates on any machine that runs it. You can redraw them yourself below.

ps_precision_bound.png 74.1 KB download
ps_roundoff_drift.png 75.2 KB download
§ Reproduction run_precision_seam.py

Two runs of the same script — mine, and one you can start right now. A match proves the result is reproducible; it is still R0 on this registry's independence rings — same code, so it cannot move a status. Only an outside run does that.

sha256 05d2cd089ea790c33a9df8bc… size 9117 B exit 0 runtime 1.5s env python 3.12.2 · numpy 1.26.4 determinism CPU-sensitive

Line 2 reports floating-point round-off, which genuinely depends on the CPU's arithmetic. Measured x86 vs WebAssembly: the absolute error moves (3.1e-03 vs 9.1e-04) but the ~10^9 float32:float64 RATIO — the actual claim — survives on both. A script that measures precision is expected to notice what it is running on.

A · Recorded founder's machine · 2026-08-02
Reversibility test on a chaotic 3-body system (float32 vs float64) ...
  at N=6400: float32 misses home by 3.1e-03, float64 by 6.8e-12 (ratio ~5e+08 = the epsilon ratio)
  wrote ps_roundoff_drift.png
  wrote ps_precision_bound.png

====================================================================
THE PRECISION SEAM — summary
====================================================================
  A chaotic float32 world violated energy conservation ~1e9x more than
  float64 — the leak is purely the machine epsilon. That is the tell.

  Run it over cosmic history: naive float64 round-off accumulates as
  ~eps*sqrt(N_ops) = 1e-16 * sqrt(1e120) ~ 1e44.
  A universe doing naive double-precision arithmetic over its 1e120
  operations would have conservation errors of order 1e44 — utter chaos.
  Reality's conservation laws instead hold to ~1e-18.

  Conclusion: reality is NOT computed with naive finite-precision. It is
  either absurdly high precision, or — far more likely — using
  STRUCTURE-PRESERVING arithmetic: conservation enforced by symmetry
  (Noether), symplectic integration, exact/error-corrected computation.
  A lazy float simulator is ruled out; a clever one is hidden. Same wall,
  fourth face.
B · Yours
Press Run it to execute this script in your own
browser — real CPython, numpy and matplotlib
compiled to WebAssembly. Nothing is sent to us;
it runs on your CPU. First run downloads the
runtime (~5 MB), then takes a few seconds.

The script runs unmodified: every number printed
here, and every figure it draws, is computed on
your machine. The figures appear below.
Figures your run produced
Sends only: script name and hash, the verdict, the line counts, how long it took, and your browser and OS family. No IP is stored (it is hashed), no fingerprint, nothing you typed. Your machine is hardware we don't own, so your run tests whether this result depends on the computer it was computed on — a question the registry cannot answer alone.
ledger 2 runs · 2 reproduced · 0 unexpected · 1 platform (Chrome/macOS)
Share this claim X Bluesky LinkedIn Reddit HN Email