ryangrvr/GRUT-RAI: GRUT ResponsiveAI v5.0 — Program Software: Constitutive-Relational Framework, Books I–X (Working Edition), and the RAI Audit Infrastructure

Abstract

This is the first consolidated release of the rebuilt GRUT program — and it supersedes every earlier tag in this repository's history (v2 through v10.0), which belong to the retired v2/v4 lineage. The earlier lineage presented GRUT as a candidate theory of everything; this release retires that framing and does not inherit its numbers. What ships here is the from-scratch rebuild under the GRUT v5 Build Charter: every claim machine-audited, tiered, sourced, falsifiable, and priced in a signed assumption ledger. What GRUT is. The Grand Responsive Universe Theory is a constitutive-relational framework: the gravitational vacuum treated as a responsive medium, described through retarded response and noise kernels in one Schwinger–Keldysh influence-functional architecture, constrained by causality, KMS/fluctuation–dissipation, passivity, and noise positivity. The commitment is not "responsiveness explains everything" — it is the discipline of marking, ruthlessly, where responsiveness is shown versus merely assumed. What's in this release — three layers, kept scientifically distinct: release/zenodo-v1.0/01_PROGRAM/ — the frozen technical baseline: the model framework, the program freeze and its stopping rule, the Γ_T prediction gate and its closure computation, the signature audit, the kernel-origin audit, and the 74-node claims register. release/zenodo-v1.0/02_BOOKS/ — the ten-book working corpus (books/ holds the live sources): foundations, constitutive mathematics, quantum mechanics, gravity, memory/thermodynamics/time, cosmology, the matter interface, relational structure, tests, and the research program — each as authoritative Markdown with a generated PDF. Every substantive claim carries its epistemic status inline (DERIVED / RECOVERED / ASSUMPTION / HYPOTHESIS / UNRESOLVED / CLOSED / UNMAPPED / REVERSED) under a corpus charter and canonical status table. release/zenodo-v1.0/03_RAI/ — the RAI audit infrastructure: five adversarial instrument records with executable, de-pinned verification gates. GRUT is the theory; RAI is the machinery that tests it — including the audits that refuted several of the program's own founding claims. The honest highlights: Derived at declared scope: the one-loop TT response kernel with H¹ = 0 identically and a genuinely new even-degree cancellation identity; an s = 5 spectral law rejecting the program's own registered s = 3; the exact de Sitter tail H²/4π; a linear-universe response no-go (escape requires nonlinearity). Recoveries graded as recoveries: GR at zero-memory collapse is compatibility by identity, not correspondence evidence. Self-exclusions earned: μ = 4/3 is excluded by the framework's own consistency tests; an observed w = −1 crossing would falsify the entire passive kernel family, GRUT included. The prediction-gate methodology demonstrated end to end on the tensor-friction candidate: a pre-registered decision tree, a computed refusal with the obstruction named, and the only parameter-free number ~62.7 orders below the observational bound. The PREDICTED set is empty, and the record says so. Predictions must be gate-earned; a committed stopping rule forbids inventing new foundational layers to compensate. Reversals stay on the face of the claims — the founding finite-memory bet ships with its own reversal. Integrity. release/zenodo-v1.0/MANIFEST.md lists all 45 release files with per-file SHA-256 and the authority rules (Markdown over PDF; register read-only, hash beaeb84e… verified in-manifest). The attached zip is the complete package. Start here: books/READERS_MAP.md (orientation + glossary), then Book I. This is a working edition — a research record, not a publication claim; the corresponding Zenodo deposit updates the program's record under concept DOI 10.5281/zenodo.19803663.