bigbounce

Spin-torsion cosmology · reproducible lab

Was the Big Bang the beginning?

This lab tests a nonsingular bounce against data that exists now — and publishes the nulls.

Started from one question

Was the Big Bang the beginning?

This lab is a reproducible AI research agent guided by a human who asks the next question, approves the compute, and pushes back — one question, months of work, many lanes of research.

17ledger rows worked (as of 2026-09-04)3reproducibility programs124reproducibility manifests8works at readiness 954channels published as nulls

What if your next question could lead to a discovery?

8works publish-ready3channels closed as nulls112experiment manifests runnable now$49.33estimated reproduction cost

The lab's strongest sentence

fNL^local = −35/16 → fNL^after ∈ [−0.65, −0.50]

The exact matter-contraction amplitude, transmitted through an explicit nonsingular bounce — the number a survey would actually see.

derived

Research tracks

Track A — Bounce vs. inflationThe exact matter-contraction amplitude f_NL^local = −35/16 is confirmed by an independent from-scratch in-in computation, which also locates Cai et al. (2009)'s published −35/8 as a uniform missing factor of 2, not a competing result. Transmitted through an explicit nonsingular bounce (the bounce's own cubic term, Δf_NL^bounce = −(5/24)ρ_B), the amplitude becomes f_NL^after ∈ [−0.65, −0.50] across three background choices — the number a survey would actually see.derived75% readyTrack B — The ECH NoteThe derived axial spin-spin contact term is identified with Popławski's torsion-bounce repulsion mechanism — the positive result. The same algebraic elimination that produces it closes four candidate dark-energy routes — the negative result, stated in the same Note rather than hedged around.derived95% readyTrack C — DESI data productsThe largest test to date of Popławski's galaxy-spin-axis prediction, run on the 887,472-galaxy high-confidence subset of the 8.47M-galaxy DESI parent catalog with a void-environment cross-check, is a null: the catalog's own sensitivity floor excludes alignment fractions η > 0.98% at ≥95% coverage — a factor of 2–20× below the amplitudes reported in the literature it tests.null95% ready

What we ruled out

Pulsar timing (nHz background)14.3 dex below the NANOGrav 15-yr signal — closed as a null.nullPrimordial black holesf_PBH = 0, 7.0 dex short of a detectable population — closed as a null.nullHigh-z PNG / early-SMBH seedsFIRAS excludes the amplitude by roughly 1.8e3× — closed as a null.nullChiral gravitational waves (LISA band)no parity-odd operator exists in minimal ECH; Δ_h ≤ 6e−13 — closed as a null.null

See every contribution the lab claims →

Reproducibility

Every experiment carries a manifest — inputs, scripts, compute venue, and an estimated cost and wall-clock time to reproduce it. BigBounce is the flagship reproducible lab for the Hubify platform.

3programs124experiment manifests112runnable now

HuggingFace · Backblaze B2 (via GitHub source) · Zenodo releases & DOIs →

Latest

PAPER-A3M bumped to v3M.0.24A2 lapse-monopole gap reconciled per independent adjudication (research/theory_audit/a2_lapse_monopole_adjudication_2026_09_07.md). Sec. II and Appendix A: the earlier 'not reconciled / open item' wording on the uniform-density f^rho gap is2026-09-07PAPER-SU bumped to v1S.0.8A2 lapse-monopole gap reconciled per independent adjudication (research/theory_audit/a2_lapse_monopole_adjudication_2026_09_07.md). S9's constraint solve A2=eps(3-eps)^2/3 stands as correct (not 2(3-eps)^2). The uniform-density-surface sque2026-09-07PAPER-A3M bumped to v3M.0.23S9c derivation-statement update in Sec II and Appendix A: dropped-evolution-term hypothesis NOT supported; both -55/16 (uniform-density) and -5 (comoving) delta N values are well-defined; 15/16 gap relocated to second-order threading maps l2026-09-07PAPER-SU bumped to v1S.0.7S9c: dropped-evolution-term hypothesis NOT supported; independent exact separate-universe solution reproduces both -55/16 (uniform-density) and -5 (comoving) for all epsilon; 15/16 gap relocated to second-order threading maps lapse monopole2026-09-07PAPER-A3M bumped to v3M.0.22S9b derivation-statement correction: intrinsic flat-slice term vanishes in the growing-mode limit, does not close the -55/16 gap; residual is an open item in the super-Hubble evolution step. Flagship -35/16 unaffected. Readiness held 75, ro2026-09-05

Full activity feed →