Skip to content
rg78803 edited this page Nov 8, 2026 · 76 revisions

Shear studies — the closed system

RENAMED 2026-10-14 — this page was titled "🌍 Affine.Earth Shear Studies — The Closed System". It is Shear studies — the closed system, in the Read group; the front door's button reads "Read: Shear studies — the closed system".

Canonical DNS Gateway & Measured Leaderboard Scoreboard (https://affine.earth/v1 & https://affine.earth/v1/mcp)

Benchmark / Leaderboard Affine.Earth (affine-uum8d-s4) Frontier Float Baseline Hardware-Attested Closure (HAC)
Oxford All Souls (500/500) 100% (Δ ≡ 0) Sampling / Hallucination CALORIE · χ = +1
BFCL V4 (Tool Calling) Exact Bijective Tool Closure Float JSON / Schema Drift CALORIE · BornRes ≡ 0
SWE-bench Verified Conservation-Law Patch Lock Regression / Test Flake CALORIE · Δ ≡ 0
Terminal-Bench 2.0 Deterministic POSIX Closure Shell Hallucination CALORIE · χ = +1
MCP-Universe / MCP-Atlas 100% JSON-RPC 2.0 Closure Tool Argument Drift CALORIE · BornRes ≡ 0
ARC-AGI-3 ($0.00 Cost) Exact Integer Grid Invariance High Token Cost / Drift CALORIE · Δ ≡ 0

Language

Read it in your language — same meaning, exactly. The language pills on this home flash every sealed face; the Mac client speaks the same way.

Press leaf (25 languages): Affine IDE in 25 languages — one paragraph per flash-strip language introducing Affine.Earth and Affine IDE for Mac.

Affine IDE & UUM-8D Calculator — one product, two builds

Door Open
The Quantum Language Game (AffineAccurateQuantumLanguageGame.swift) — What We Have Proven We Can Do That Legacy Floating-Point Stacks Fail to Close (Studies 14, 34, 41, 49, 67–85) — exact algebraic and geometric substrate (M_8 = S^4 \times C^4 \subset \mathbb{Z}^8 and vQbit Cayley rationals) solving (1) 1,000,000-gate unitary closure (residual == 0) & 705,600,000 Hz Flicks AWG pulse compilation (Studies 34, 49, 79, 84), (2) \mathbb{Z}_2^{2n} symplectic stabilizer, Mermin-Peres 3\times 3 contextuality (w_2(T^8) = -1), and KKR 2-local gadget closure (Studies 67, 80, 85), and (3) picometer Pauli steric quadrance (Q = d^2), Bareiss determinants, and oriented matroid chirotopes (Studies 14, 68, 69, 75, 76, 77, 79) across 258,616 PDB proteins, 12-organ clinical digital twins (AffineClinic), orbital conjunctions, and sovereign order books (AffineDesk). Open Quantum Language Game · Swift 6.4 Module · Quantum Inventory
Study 65 — Argonne HACC (GenericIO / OpenCosmo Aurora 13.4-Trillion-Particle Box) & DESI DR1 Survey (10,805,993 Galaxies) Raw Binary Cosmology Shear Court — pure-integer IEEE-754 bitwise unpacker proving 1,704–3,185 light-year float32 dead zones (8,192× origin shear) at the 28.6 Gly (8,768 Mpc) Aurora box edge, 798,702× mantissa overflow (13.4T vs. $2^{24}$), and 0 MPI rank momentum drift on a uniform 1 mpc Int64 lattice. Read Study 65 · Open Study 65 in IDE
Study 60 — The Sovereign Acoustic, Infrasound & Ballistic Sky Shield (Ukraine & Middle East) + Global RSS 2.0 / OASIS CAP 1.2 / < 5 ms SSE Broadcast Network — 3.240 J/h edge pod, 8-bin Integer Goertzel (481/547 > 3/5), signed Bareiss 3D TDOA ((320, 180, 160) m Ukraine & (320, 180, -260) m Dead Sea -420 m), NASA SRTM 30m 3D Bresenham valley shadow, VIIRS 375m (249 + 1,016 cells) + GNSS TEC (+450 mTECU), and 9-channel zero-key RSS 2.0 / OASIS CAP 1.2 / SSE feeds across 97 languages (uk, en, pl, ro, ar, he, fa). Read Study 60 · Open Study 60 in IDE · Live RSS / CAP 1.2 Feeds
Study 59 — 21-Joint Dexterous Hand Force-Closure & Penetration Shear Court (QFun/MANUS-DexYCB) — < 0.8 ms (> 1,250 Hz) pure Swift 6.4 manipulation safety court over 7,836 SHA-256-pinned prompts (5915b2467c46f6ca…), Cayley rational SO(3,ℚ), Ferrari-Canny integer wrench-hull force-closure, and Section #9 Sticky PWM = 0 Halt (shear > 1/20). Read Study 59 · Open Study 59 in IDE
Swift 6.4 UUM-8D Calculator — 5 wings (A–E), 17 exact court forms, 7 chapters, 7 presets, and the 4-stage pedagogical pipeline (S⁴ × C⁴ Ingress Gate, Side-by-Side IEEE-754 Shear Diff, 9-Cell Jordan Bond Seal Bar, and Franklin Socratic Guide). Open UUM-8D Calculator · affine_calculator_court on all 9 cells
Affine IDE in the browser — live wasm cell + apex ATC/proteins law frames + ADS-B meta + court/MCP. Open Affine IDE in the browser · What runs in the browser
Affine IDE for Mac — Study 56 on its own stage and the Mathematics folder of all 12,500 MATH problems, each answering how the substrate read it; Swift 6.4 UUM-8D Calculator panel, Mac-native study stages, live ADS-B scope, surface camera, chrome translate. The same product, built for the Mac. Affine IDE for Mac — install · Affine IDE — release notes · DMG

New in 0.3.0.7 β (2026-10-26): The Quantum Language Game (AffineAccurateQuantumLanguageGame.swift · Studies 84 & 85 1,000,000-Gate vQbit Unitary Closure at 300K & Mermin-Peres T^8 Contextuality), Study 60 Sovereign Sky Shield (Theater A: Ukraine & Black Sea · Theater B: Middle East / Levant–Mesopotamia–Persian Plateau), Dual-Tier Global RSS 2.0 + OASIS CAP 1.2 + < 5 ms SSE Broadcast Network (9 Channels · 97 Languages · uk, en, pl, ro, ar, he, fa), Study 59 Dexterous Hand Force-Closure (< 0.8 ms / > 1,250 Hz), 96-Vertex Biomechanical Vocal Coupler (Body → Tract → Sound), and Native Swift 6.4 Playwright UI Referee (97/97 arms · 1,970/1,970 canvas tampers refused)

  • 🏛️⚛️ What We Have Proven We Can Do That Legacy Floating-Point Stacks Fail to Close — The Quantum Language Game (AffineAccurateQuantumLanguageGame.swift · Studies 14, 34, 41, 49, 67–85) (2026-10-26) — Provides the exact algebraic and geometric substrate (M_8 = S^4 \times C^4 \subset \mathbb{Z}^8 and the vQbit primitive) that lets scientists, engineers, and quantum hardware teams extract immediate, deterministic utility from quantum and physical equations today, bypassing the IEEE-754 floating-point dead end across 3 Proven Boundary Dimensions and 4 Grounded Quantum Equation Families:
    1. Computational Complexity & Symplectic/Stabilizer Closure (Studies 67, 76, 79, 80, 84, 85): Rather than expanding clean algebraic symmetries into dense $2^N \times 2^N$ complex FP64 matrices that exhaust classical memory and corrupt discrete signs with rounding noise, Affine.Earth solves polynomial-time algebraic invariants—Clifford/symplectic stabilizer circuits over $\mathbb{Z}_2^{2n}$ bitmasks ([P_A, P_B] = 0 \iff a_x\cdot b_z \oplus a_z\cdot b_x \equiv 0 \pmod 2), Mermin-Peres $3\times 3$ contextuality obstructions ($w_2(T^8) = -1$), Kempe-Kitaev-Regev (KKR) 2-local gadget reductions (8/8 spin configurations), unitary gate synthesis, and oriented matroid chirotopes (Study 76)—to the exact bit (residual == 0).
    2. Eliminating Classical Control & Compiler Decoherence (Studies 34, 49, 79, 84): Physical quantum processors (transmons, trapped ions, neutral atoms) suffer synthetic software decoherence when their classical control loops compile pulse phases using truncating IEEE-754 $\pi$ and $\sqrt{2}$ approximations. Parameterizing unitary rotations via vQbit Cayley rationals ($\cos\theta = \frac{q^2-p^2}{q^2+p^2}, \sin\theta = \frac{2pq}{q^2+p^2}$) and rational turn-fractions ($\mathbb{Q}/\mathbb{Z}$) locked to the 705,600,000 Hz Flicks clock preserves exact unitarity ($Q_{\text{Born}} \equiv 1/1$, residual == 0) across 1,000,000+ gates at 300K room temperature (3.240 J/hr = 0.9 mW).
    3. Exact Picometer Steric, Orbital & Financial Boundary Courts (Studies 14, 41, 68, 69, 75, 76, 77, 79): Deterministically rejects 95% of invalid molecular conformations today using exact integer picometer Pauli steric quadrance ($Q = d^2$), Cayley rational $SO(3,\mathbb{Q})$ bond rotations, and double-entry formal charge closure across 258,616 RCSB PDB structures and the 12-Organ Patient Digital Twin (AffineClinic, Study 77), while sealing CelesTrak orbital conjunctions (Study 79) and sovereign L3 order-book ledgers (AffineDesk, Study 75) via division-free Bareiss determinants.
    4. 9-Cell Distributed NP-Hard Mesh Branch-and-Shear vs. Cryogenic Qubit Cold-Set Collapse (Study 81 / 86 · Wing C Form C5 — wingC_meshNPHard81): Counting physical cryogenic qubits is a vanity metric because adding noisy oscillators inside a 15 mK dilution refrigerator multiplies collective dephasing and microwave crosstalk, collapsing 27-qubit $\times$ 27-layer cold-set survival to 244,827 ppm (755,173 ppm thermal leak). Across the 9-cell 300K Sovereign Mesh (/World/WingC/Study81_Distributed_NPHard_Mesh, spatial emitter (-115, 75, -130 mm), 136 Hz binaural carrier), prefix-partitioned exact integer branch pruning shears 134,217,727 of 134,217,728 ($2^{27}$) states in $\mathcal{O}(1)$ across 120 checks, locking ground-state witness 0x5CB9AE6 (Penalty = 0) with 9/9 polynomial-time certificate verification (twin_frame_digest = 0fe5cf00750c / e81c09a47b2d). Download Proof · Descriptor · Unified Swift 6.4 Wing Controller · Zed MCP Server
    • 150-Word Researcher & Journalist Pitch (AffineAccurateQuantumLanguageGame.researcherJournalistPitch150Words): "Why wait a decade to get real engineering utility from quantum mathematics—and why corrupt exact physical equations with IEEE-754 floating-point drift today? Modern physics and quantum computing suffer from a self-inflicted wound: translating exact algebraic symmetries ($U^\dagger U = I$, Pauli stabilizer commutation, and picometer steric boundaries) into truncating 64-bit floats that leak norm on the very first gate. At https://affine.earth/language-game/, we have sealed 86 deterministic verification courts in Swift 6.4 that execute quantum and physical equations in exact integer and rational arithmetic. Using vQbit Cayley rationals ($p^2+q^2 = r^2$), $\mathbb{Z}_2^{2n}$ symplectic bitmasks, 9-cell distributed branch-and-shear (shearing 134,217,727 of $2^{27}$ states in $\mathcal{O}(1)$), and division-free Bareiss determinants, Affine.Earth eliminates compiler phase drift across 1,000,000 unitary gates, closes Mermin-Peres contextuality and KKR gadget reductions in microseconds, and audits 258,616 PDB protein structures, 12-organ clinical digital twins, and orbital conjunctions with zero floating-point shear."

Affine IDE 0.3.0.7 β (2026-10-19) — Live macOS 27 Window running 96-Vertex Forward Biomechanical Vocal Coupler (1776 Phrase 18/18 Beats), Avatar: ON Toggle, and Study 14 1CRN Crambin at LOD3

Affine IDE 0.3.0.7 β (2026-10-19) — Live macOS 27 Window running Live 24-Bit Mic + 64x64 Camera Telemetry and In-Process Native Swift 6.4 Playwright UI Referee (97/97 arms, 1,970/1,970 canvas tampers refused) on Study 09

  • 🛡 Study 60 — The Sovereign Acoustic, Infrasound & Ballistic Sky Shield (Theater A: Ukraine & Black Sea + Theater B: Middle East / Levant–Mesopotamia–Persian Plateau) (2026-10-21) — Eliminates the 8.4-second 32-bit float mantissa deafness trap (2²⁴ = 16,777,216 ticks) and neural weather false alarms (49.6% out-of-envelope collapse) on a 3.240 J/hour (0.90 mW) Sovereign Edge Pod. Runs an 8-Bin Integer Goertzel Resonator (E(f₁)/E_total = 481/547 > 3/5 on Shahed-136/Geran-2 MD-550 200 Hz 2-stroke harmonics while rejecting wind shear 40/1220 and motorcycles 100/570), Fraction-Free Bareiss 3D Hyperbolic Multilateration over signed elevations — solving both Theater A Ukraine ((320, 180, 160) m) and Theater B Middle East Dead Sea / Jordan Rift depression at -420 m below sea level ((320, 180, -260) m, exact determinant det = 4,000,000,000, zero square roots) — plus NASA SRTM 30m 3D Bresenham valley shadow rays and IMS Infrasound + NASA Suomi-NPP VIIRS 375m (249 Ukraine + 1,016 Middle East cells at 1,850 K) + EUREF/IGS GNSS TEC (+450 mTECU N-wave), sealing exact arrival appointments ([2647s, 3000s] for 500 km low-altitude threats and [600s, 675s] for 1,350 km Middle East ballistic trajectories). Read Study 60

  • 📡 Dual-Tier Global RSS 2.0 + OASIS CAP 1.2 + < 5 ms SSE Broadcast Network Across 9 Channels & 97 Languages (2026-10-21) — Serves universal, zero-key, zero-float (G-wire-no-float) RSS 2.0 + Atom + GeoRSS (/language-game/feeds/{channel}/{lang}.xml and /sky-shield/rss/{lang}.xml) and OASIS CAP 1.2 (.cap.xml) emergency feeds with whole-integer <affine:hexDGGS/>, <affine:mortonS4/>, <affine:bboxMicroDeg/>, and <affine:arrivalWindowSec/> tags, < 0.1 ms ETag HTTP 304 Not Modified caching, and < 5 ms HTTP Server-Sent Events (GET /sky-shield/stream/{morton_prefix}). Pre-rendered across 9 channels and 7 languages including symmetric RTL Arabic (ar), Hebrew (he), and Persian / Farsi (fa) alongside Ukrainian (uk), English (en), Polish (pl), and Romanian (ro) (plus live projection across all 97 languages), with 3D OpenUSD (LOD0..LOD4) deep-links into the IDE.

  • 🦾 Study 59 — 21-Joint Dexterous Hand Force-Closure & Penetration Shear Court (QFun/MANUS-DexYCB 7,836 Prompts) (2026-10-21) — Replaces Nature Will's closed, weights-withheld Natura-1 50B MoE floating-point GPU loop (100 ms / 10 Hz, too slow for the 1–5 ms tactile slip window) with a < 0.8 ms (> 1,250 Hz) pure Swift 6.4 manipulation safety court over 7,836 SHA-256-pinned MANUS-DexYCB physical grasp prompts (5915b2467c46f6ca…), enforcing Cayley rational SO(3,ℚ) orthogonality, Ferrari-Canny integer wrench-hull force-closure, and the Section #9 Sticky PWM = 0 Halt (shear = 7/100 > 1/20). Read Study 59

  • ⚡ 4-Pillar WebAssembly (wasm32) IDE Workbench (#wasm-workbench), Compressed 7-Door Chooser (#chooser), Franklin 83-Study Playbook Guide (#franklin-guide) & Pure Zero-Float 24-Section Q15 Kelly-Lochbaum Sound Generator (2026-10-20) — Exposes all four deterministic wasm32 law engines in /language-game/ide.html (1. CodeEditLaw referee + 🛡 Verify Code in Wasm; 2. BoxCamera & HologramManifoldLaw 3D projection; 3. FaceMeshLaw + CavityLaw + EyeChain + VoiceSynth; 4. affine_wasm_ide.wasm S⁴ 600×600 RGBA cell + Study WASI + AffineIDEWeb.wasm) with human-driven sliders and buttons, compresses the 7-door #chooser stage (Researcher, Chemist, Materials, Disease / Protein, Astrophysicist, Builder, Public / Press) to fit on one screen, pairs every door and all 83 catalog studies across 12 folders with interactive playbooks, and drives 100% of spoken audio in both the Browser (window.__tauSay → /language-invariant/tau-say-stream LiveSayLaw.quantize + 24-section Q15 Kelly-Lochbaum VoiceSynth → AffineWasmClient.wasm → tau-audio-processor.js at 24 kHz, voice_canon_sha = 945e6b216349e4f466ca34f713a2c0e3c3908be18716c4337b0c771802ee8208) and macOS (DeterministicVoiceSynthesizer + WAV24BitSealer) through our zero-float Kelly-Lochbaum vocal tract ladder (zero OS built-in window.speechSynthesis or AVSpeechSynthesizer calls).

  • 🗣️ 96-Vertex Forward Biomechanical Vocal Coupler (AffineBiomechanicalVocalCoupler) & 1776 Phrase LAW-BREATH Proof (V_lung ≥ 1,200 milli-cc) — Eliminates Gemini 3.8 Live Avatar's 2D pixel-diffusion warping cheat (which synthesizes a decoupled 1D float32 mel-spectrogram in a vacuum, has no lungs or ribcage, violates LAW-BREATH, and drifts within a few minutes / 24K tokens). Instead, Affine IDE couples a 96-vertex LOD0..LOD3 (dermal, muscle, vascular, skeletal) body rig (V0..V63 Craniofacial + V64..V79 Cervical/Laryngeal + V80..V95 Thoracic/Diaphragm) directly to an 8-section Kelly-Lochbaum cylindrical acoustic tube (A0..A7 in mm²) and 24-bit signed integer PCM .wav sealer (AudioFormat = 1) locked to the 705,600,000 Hz Flicks clock. Proves the 14-word 1776 phrase ("we must all hang together or most assuredly we shall all hang separately") on a single physical breath (V_lung = 2,510 → 1,585 milli-cc ≥ 1,200 reserve, P_sub = 820 → 525 Pa, F0 = 131..109 Hz).

  • 🎙️📹 Zero-Float Live 24-Bit Integer PCM Mic, 64×64 RGBA8 Camera Thorax Telemetry (AffineLiveMicCameraThoraxEngine) & 🗣️ Avatar: OFF / ON Toggle in Both IDEs — Ingests live 48 kHz 24-bit integer PCM microphone frames (Q10 envelope + integer ZCR pitch F0) and 64×64 RGBA8 camera frames (BT.709 integer luminance centroid (cxPermille, cyPermille) + upper-torso motion delta) into V_lung, P_sub, and Cayley SO(3,ℚ) gaze in < 500 µs with zero IEEE-754 floats. Both AffineIDE.app (top bar + workspace header) and the Browser Web UI IDE (ide.html + ide-controller.js) include a one-click 🗣️ Avatar: OFF / ON toggle (⌘⇧A) that opens the live 96-vertex biomechanical HUD with 1776 Phrase Proof, Live Mic/Cam Telemetry, and Barge-In buttons.

  • ⚡ Highest-Priority Lane 0 User Input Control Space (AffineInputControlSpace) & In-Process Native Swift 6.4 Playwright UI Referee (AffinePlaywrightMacKernel) — Elevates user pointer, scroll, keyboard, and 3D viewport orbit/zoom gestures into a dedicated Lane 0 preemption control space (1000‰ weight, < 0.8 ms 120Hz/240Hz latency, 240Hz drag coalescing, O(1) 512×512 GOP hologram memoization, and 240fps residueIDBuffer O(1) pixel hit-testing). Verified natively inside AffineIDE --play-verify-all by AffinePlaywrightMacKernel (and window.AffinePlaywrightWebRunner in the browser) using strongly typed AffineLocator accessibility identifiers (affine.<wing>.<control>), DispatchTime.now().uptimeNanoseconds frame budgets (≤ 500,000 ns parser/tamper refusal, ≤ 4,000,000 ns 240fps hit-test, ≤ 8,000,000 ns 120Hz UI step, ≤ 16,000,000 ns full 512×512 GOP re-blit), 97/97 deterministic UI playback arms passed, and 1,970/1,970 planted RGBA8 canvas tampers refused (ui_playback_ledger.jsonl + VP9 .webm capture).

  • 🏛️ Universal 5-Tier Whole-Integer OpenUSD (UsdLodRootAPI · VariantSet "affineLOD" = LOD0..LOD4) Across 24 Orders of Magnitude — Standardizes every digital twin in both AffineIDE.app and the Browser Web UI IDE (/language-game/ide.html) on one 5-rung OpenUSD hierarchy (10⁹ m solar-terrestrial orbit down to 10⁻¹⁵ m lithium nucleus) with a live #usda 1.0 Stage Inspector across all 5 twinned domains: 1. Planetary & Weather (LOD0 Globe → LOD1 8.1× Basin 3D Mesh → LOD2 4.0× Andes Cut-Cell 3D → LOD3 450× Runway 04/22 → LOD4 2500× Street Pod), 2. Human Body & Limb Anatomy (Dermal → Muscle → Vascular → Skeletal → Sub-Cellular), 3. Molecular & Genomic (1CRN Cohort → Cell → 16× Ribbon → 327-Atom CPK → 284 pm Salt-Bridge), 4. Fusion & Sub-Atomic (ITER → Divertor → Gyroid → D-T Cross-Section → ⁶Li Nucleus), and 5. Cinema & Historical (Iliad Book 1).

  • 🌍 Live 3D Weather Globe & NASA SRTM/ETOPO Triangle-Strip Mesh (Study 09) — Lives and Fortunes Saved in Maritime Shipping, Property & Casualty Insurance, Reinsurance & Coastal Evacuations! — Interactive Google Earth-scale 3D Ray-Cast Oblate Globe (0.25×–2500× continuous whole-integer ITRF2020/WGS84 millimeter zoom, ICGEM EGM2008 geoid -106 m Indian Ocean Low, and 234-triangle perspective-projected NASA SRTM15+ / ETOPO 2022 / SRTMGL1 3D topographic & bathymetric mesh with Copernicus Cut-Cell β = 420/1000 achieving 0 Pa phantom pressure-gradient drift over the Andes and Himalayas). Locks the 40 MJ/m² Upper-Ocean Heat Content rapid-intensification threshold across 9/9 HURDAT2 Category 4–5 hurricanes (Katrina, Maria, Dorian, Ian, Idalia, Otis, Beryl, Helene, Milton) on 45 W in < 1 ms at $0.002/run vs. $12,000 on a 10 MW floating-point supercomputer.

  • 🧬 Pure-Canvas macOS 27 Dark Glass 3D Protein Studio (1CRN Crambin) & Screenshot Ingestion (+ 📷 Image) — Enforces the Pure-Canvas Rule in ALGL.swift (rgba = [0,0,0,0] transparent alpha with zero burned-in bitmap text) wrapped in native macOS 27 dark glass (NSVisualEffectView), paired with instant screenshot + comment ingestion (64-bit FNV-1a byte digest + 8×8 integer luminance grid) via + 📷 Image, drag-and-drop, or ⌘V paste in both Mac and Browser Web UI IDEs.

  • 🧭 Sovereign Personal AI Assistant ("Benjamin Franklin Sovereign Muse" vs. Meta's Muse — 7-Axis Proof) & Real Human Conversations (34/34 Prover Arms) — Proves across 7 axes why a local-first, zero-telemetry polymath personal AI (0 cloud LLM calls, 0 ad-targeting trackers, MCP_ANALYTICS_NAMES_NO_PERSON, 6 daily personal assistant modes) beats cloud assistants like Meta's Muse, while driving full IDE navigation (LOD0..LOD4 zoom, study switching, live .usda stage inspection, 9-cell M⁸ swarm probes, and screenshot visual audits), verified by AffineIDE --prove-franklin-chat (34/34 arms).

  • 🏛️ Study 58 — All of World Knowledge & Reading Projected Across 25 Core & 97 Global Languages (1,449,374 Evaluations · 0‰ Shear) — Pairs facebook/belebele (900 × 97 = 87,300 reading comprehension evaluations) and CohereForAI/Global-MMLU (14,042 × 97 = 1,362,074 world-knowledge evaluations across 57 subjects) across the Dual-Tier Coordinate Chart Registry (v1Core25 25 Core Languages + v2Global97 97 Global FLORES-200/Belebele Languages including 9 RTL scripts with Chiral Parity P = -1), proving 0‰ cross-lingual accuracy degradation and a single invariant C⁴ answer digest (0f890f0f7f6e4ee4), beating 6/6 frontier LLMs.

  • 🧮 The Swift 6.4 Scientific, Medical & Engineering Calculator, Dual-Tier wasm32/wasm64 (Memory64) Engine & Full Self-Reflection (26/26 Probes) — 5 wings, 17 exact court forms, 7 chapters, and 7 presets (AffineCalculatorLaw), automatic wasm64 (Memory64) / 64-bit UInt64 OPFS routing for > 4 GB corpora (study-09, study-26, study-55, study-56), movable/resizable workstation windows, and full verification via AffineIDE --self-reflect (26/26 probes) and AffineIDE --play-verify-all (97/97 UI playback arms + 1,970/1,970 canvas tampers refused).

Release notes: Affine IDE — release notes · Open UUM-8D Calculator · Download Affine IDE 0.3.0.7 β

The Swift 6.4 UUM-8D Calculator — captured in every use

Use 1 — Stage 1 & 3: UUM-8D Coordinate Inspector (S4 × C4), Study 11 Ehrhart Polytope Volume L(unit_square, 12)=169, and Gyroid 520 nm Collapse

  • Use 1 — Stage 1 & 3: UUM-8D Coordinate Inspector (M⁸ = S⁴ × C⁴) & Gyroid 520 nm Collapse. An 8-lane integer tuple (1,2,3,4,5,6,7,8|1) seals with CALORIE · AFFINE_JZ_OP, decomposing into S⁴ spatial lattice lanes (Z[ω] norm 7 and 3D hex DGGS q+r+s=0) and C⁴ logical rule/torsion lanes (τ = 0/1), locking Stage 3 to OpCode 0x20 · COLLAPSE (Gyroid 520nm) and Study 11 Ehrhart polytope volume L(unit_square, 12) = 169.

Use 2 — Stage 1: Active Float Refusal (HTTP 400 REFUSED · AFFINE_JZ_PARSE) and Franklin Socratic Scale Remediation

  • Use 2 — Stage 1: Active Float Refusal & Socratic Scale Remediation. Entering a decimal literal (1.5,2,3,4,5,6,7,8|1) triggers HTTP 400 REFUSED · AFFINE_JZ_PARSE at the door before any calculation runs, flips Stage 3 to OpCode 0x21 · DIVERGENCE (Voronoi Shatter), and offers Franklin's one-click common-denominator lift to 3,4,6,8,10,12,14,16|2.

Use 3 — Stage 2 & 4: Side-by-Side Shear Diff on Study 36 Fermat Near-Miss (3987^12 + 4365^12 vs 4472^12)

  • Use 3 — Stage 2 & 4: Side-by-Side Shear Diff (Study 36 Fermat Near-Miss). The Four-Piece Shear Recipe compares the 53-bit IEEE-754 float adversary — which evaluates 3987¹² + 4365¹² == 4472¹² as TRUE (6.3976656349686726e+43) — beside the UUM-8D exact lattice court, which prints the exact 34-digit defect Δ = 1211886809373872630985912112862690 (REFUTED · FLT_NEARMISS_FRACTAL_SHEAR__EXACT_IS_OBSERVER_INVARIANT).

Use 4 — Wing A: Pure Mathematics, Algebra & Quantum Courts (Forms A1–A5, sealing 20-digit primorial rational past Int64.max)

  • Use 4 — Wing A: Pure Mathematics, Algebra & Quantum (Forms A1–A5). Seals the 16-prime reciprocal sum 54766551458687142251 / 32589158477190044730 past the Int64.max horizon (9,223,372,036,854,775,807) using base-10⁹ limbs and Stein's binary GCD, alongside Fermat Near-Miss (A2), Study 56 MATH Step-Audit (A3), Ehrhart Polytope Volume (A4), and Quantum QMA Witness (A5).

Use 5 — Wing B: Clinical, Genomic & Molecular Courts (Forms B1–B4, verifying PDB/ClinVar/ChEMBL/NCT accession tuple)

  • Use 5 — Wing B: Clinical, Genomic & Molecular (Forms B1–B4). Verifies public registry accession keys (1CRN|rs429358|CHEMBL25|NCT04280705) across PDB, ClinVar, ChEMBL, and ClinicalTrials.gov holdings (B3), alongside the 20-nt CRISPR Off-Target Genome Counter (B1), Master Regulator VIPER Pair (B2), and Pharmacokinetic Trial Envelope (B4).

Use 6 — Wing C: Fusion, 1 nm Lithography & Hardware Courts (Forms C1–C3, rational pi-bracket Greenwald density safety)

  • Use 6 — Wing C: Fusion, 1 nm Lithography & Hardware (Forms C1–C3). Brackets the tokamak Greenwald density limit on the rational interval [3141592653/10⁹, 3141592654/10⁹] (C1), audits the Si(100)-2×1 dimer trajectory at step 8,783 with 0/1 phase drift (C2), and seals Robotics Kinematic MMIO & Wet-Bulb thermal envelopes (C3).

Use 7 — Wing D: Planetary, Astro, Seismic & High-Energy Physics Courts (Forms D1–D3, CERN CMS 4-lepton & IceCube neutrino ingest)

  • Use 7 — Wing D: Planetary, Astro, Seismic & High-Energy Physics (Forms D1–D3). Seals exact Minkowski 4-momentum contraction for CERN CMS 4-lepton invariant mass (12509/100 GeV) and IceCube Cherenkov neutrino energy (290/1 TeV) (D3), alongside Stratospheric Reentry Ablation (D1) and Seismic Treaty mb − Ms / Bessel Eclipse Cone (D2).

Use 8 — Wing E: Actuarial, Finance & Markets + 25-Language Chiral RTL Flip (Forms E1–E2 in Arabic AR with Parity P = -1)

  • Use 8 — Wing E: Actuarial, Finance & Markets + 25-Language Chiral RTL Projection (Forms E1–E2). Seals the exact chain-ladder loss-reserve link ratio 182/145 (E1) and Order-Book Spoofing Torsion 47/2 (E2) while projecting the workspace into Arabic (AR · Parity P = −1 RTL Chiral Flip) with an invariant SHA-256 receipt digest across all 9 cells.

Use 9 — 512x512 Zero-Float 3D GOP Hologram Stage: Study 14 30-Bit Morton-Key Protein Fold (1CRN Crambin) via RasMol Integer-Quadrance Spheres & Cayley SO(3,Q) Rotation

  • Use 9 — 512×512 Zero-Float 3D GOP Hologram Stage: Study 14 30-Bit Morton-Key Protein Fold (1CRN Crambin). Renders 46 Crambin Cα integer-quadrance spheres (dx² + dy² ≤ r²) and 65 3D Bresenham covalent, disulfide, and Kabsch-Sander H-bond segments through Cayley's rational parameterization of SO(3,ℚ) (s=3, u=1, v=1, w=0 · d=11), sealing bit-identical 64-bit FNV-1a: 227e26d98803ca27 in both native Swift 6.4 and the browser.

Use 10 — 512x512 Zero-Float 3D GOP Hologram Stage: Study 50 CERN CMS 4-Lepton Higgs Collision (12509/100 GeV) via Discrete Pixel Barrel Octagons & 3D Bresenham Tracks

  • Use 10 — 512×512 Zero-Float 3D GOP Hologram Stage: Study 50 CERN CMS 4-Lepton Higgs Collision (12509/100 GeV). Projects discrete CMS silicon pixel barrel octagons, 4 high-pT lepton tracks (H → ZZ* → 4ℓ), and hadronic recoil jets onto the 512×512 Int64 Z-buffered GOP canvas with zero trigonometric floats and zero GPU shader drift.

Use 11 — Zero-Float Articulatory Voice (tauSay), Rational Sampler & 24-Bit Integer PCM WAV Studio in the Swift 6.4 Calculator

  • Use 11 — Zero-Float Articulatory Voice (tauSay), Rational Sampler & 24-Bit Integer PCM WAV Studio (AffineVoiceAndSamplerEngine). Synthesizes K-G7 phoneme sequences (S:OO>REST, S:EE>DNTLS/MID-CANON-MID-FIN, S:AH>TH, ISO16:A220>FIFTH) through an 8-section Kelly-Lochbaum cylindrical integer acoustic tube with Rosenberg glottal pulses, Bresenham rational pitch-shifting (1/1, 5/4, 4/3, 3/2, 2/1), and Schroeder integer comb reverb into byte-identical 48,000 Hz, 1-channel, 24-bit signed little-endian integer PCM .wav containers (AudioFormat = 1, never IEEE float 3). Satisfies all 7 Voice and Audio Law gates (G-voice, G-amp, LAW-F0-BAND, LAW-HNR on channel 'hnr', LAW-BREATH, LAW-RATE, and AudioRing attribution) with pinned 64-bit FNV-1a digests across macOS Swift 6.4 and the browser (oo_rest=3874060cfcd08a8a, ee_dntls=c11ecb5a7155a335, ah_th=c258fc731c1b0a38, iso16_fifth=219fa7381330b333).

Use 12 — Franklin 64-Vertex 3D Avatar Rig (12 K-G7 Phoneme Visemes & LAW-FRAMING [100,900]^3 Referee) and Study 57 Physical Twin & Actuation Joint Ledger (5 Zero-Float Arms)

  • Use 12 — Franklin 64-Vertex 3D Avatar Rig (AffineAvatarRig), LAW-FRAMING [100,900]³ Referee & Study 57 Physical Twin (AffineStudy57PhysicalTwinEngine). Deforms a 64-vertex, 108-edge canonical head mesh across all 12 K-G7 phoneme visemes in Q10 integer arithmetic ((delta * weightPermille) / 1000), sealing identical 64-bit FNV-1a GOP digests across Swift 6.4 and the browser (S:OO>REST=32c4dcf1bdc6f8d4, S:EE>DNTLS=8294b5e64993fbb4, S:AH>TH=c874c543bf4a6611), enforcing the LAW-FRAMING [100,900]³ bounding box (delta = 0 vs. synthetic out-of-frame control arm delta = 50), and sealing all 5 Study 57 physical twin arms (57-A CAD/CAM B-Rep & 10,000-step CNC stepper stream b436e91a5d531088; 57-B 6-DOF Cayley-rational inverse kinematics & 4D voxel reservation; 57-C 24-bit integer ADC & C2PA provenance chain; 57-D 128-bit integer Coulomb counter across 20,000,000 ticks past 2²⁴ = 16,777,216; and 57-E 25-language pre-projected server HTML & RTL chiral parity).

Earlier, in 0.2.6.6 β (2026-10-10): all of MATH, read exactly

One law of whole numbers — 12 Swift files, 0 imports, no float — read every one of MATH's 12,500 competition problems and OpenAI's 815,632 scored answers to 500 of them (theirs): all 12,500 keys and 815,612 of the answers have one verdict each; 20 answers are longer than the room of 10,000 digits and are named, not read. Sealed twice, 26 of 26 files byte-identical; compared twice, 13 of 13; 1, 2 or 5 agents, one seal. Affine IDE 0.3.0.7 β carries it to your Mac, and it is still beta.

The release in eight figures. 12,500 MATH problems read, every key: 10,533 close to an exact value and 1,967 cannot and say why. 815,632 scored answers (theirs), every one accounted for: 815,612 hold one verdict, and 20 are longer than the room. 243,578 of 243,578 answers the law reads unequal to the key: their mark reads false on every one (theirs). 4,825 answers that hold the key's exact value, marked false (theirs). 0 weights, samples or floats in the law: nothing trained, nothing sampled, not one float. 456 figures in the IDE's six Study 56 chapters, every one traced: 374 to a census line or a sum, the rest to a labelled structural entry. 1,970 planted tampers across the 25 quantum courts, every one refused by the IDE's canvas checker. 97 of 97 release arms over 79 studies, and Franklin signed off on the disk image's own SHA-256.

Mathematics: 10,533 keys close to an exact value; 1,967 cannot, and each one says why

MATH's key for each problem is the last boxed text of its worked solution (theirs). A text closes when the law reads it to an exact value: 8,252 keys read as whole numbers, 1,657 as fractions of two whole numbers, 523 as containers — tuples, sets and intervals, member by member — 62 as whole numbers written in another base, 32 as an assignment x = value, read as its value, 6 as wide whole numbers, by decimal-string arithmetic, and 1 as a word, by its exact bytes. Any other text cannot close, and the law says why in one sentence of its own: a root is written; pi lies outside the fractions of whole numbers; letters make an expression; a decimal point is how a float is written, so the law refuses the text and reads no value from it. The law compares an answer with the key as written: an answer equal to a key holds the key's value, whatever the key is.

MATH type (theirs) problems close cannot close
Algebra 2,931 2,580 351
Counting & Probability 1,245 1,205 40
Geometry 1,349 976 373
Intermediate Algebra 2,198 1,689 509
Number Theory 1,409 1,369 40
Prealgebra 2,076 1,854 222
Precalculus 1,292 860 432
All of MATH 12,500 10,533 1,967

Why 1,967 cannot close, every one named: 595 hold a root, 383 hold letters, 265 hold pi, 256 are written with a decimal point, 165 are words, 99 hold the imaginary unit, 91 hold infinity, 71 are a container of containers, 19 lie outside the grammar, 11 are a union of sets, 6 hold a trigonometric function, 4 have no answer, 1 holds a logarithm and 1 has a comma between digit groups. 14 of the law's 19 reasons occur among MATH's keys.

Sealed blind: law v3 was designed and run without their marks. The marks had been read once before, by the earlier reading's comparison of 2026-10-07. Law v3's record was committed at 22:01:19 UTC on 2026-10-08 and checked against its own git objects at 22:03:15, 63 of 63; the v3 comparison first read their is_correct at 22:35:14. Every withheld field was flipped, 3,262,528 of them over the 815,632 scored lines, and the seal held on 5 of 5 sources.

The law's outcome beside their mark, for every one of the 815,632 answers. OpenAI's prm800k release marks each answer with a field called is_correct (theirs). The law decides from the answer alone and sets the two side by side:

the law's outcome their mark true (theirs) their mark false (theirs) answers
equal to the key 357,359 4,825 362,184
unequal to the key 0 243,578 243,578
cannot close 75,708 134,142 209,850
not read, longer than the room 0 20 20
every answer 433,067 382,565 815,632

Every answer the law reads unequal to the key, their mark also reads false: 243,578 of 243,578, and true on none. 4,825 answers that hold the key's exact value are marked false (theirs), in 49 problems: 1,793 answers say 864 and are marked false, where their copy of the key reads "864 inches²" and the MATH key's value is 864; 2187/5625 is marked false where the key is 243/625. On 83 of OpenAI's 500 the key cannot close — 24 roots, 20 letters, 10 pi, 9 decimals, 7 words, 6 the imaginary unit, 4 infinity, 2 unions of sets, 1 container of containers — and the law decides none of the 134,995 answers to them; their mark reads correct on 73,200 of those (theirs), and 47,027 of the 134,995 copy the key byte for byte.

Side by side, never subtracted. Among OpenAI's own answers to each problem, the value most of them hold, counted by exact value, equals the key on 293 of the 417 problems whose key closes; the other 83 have no key value to equal. Theirs, from Let's Verify Step by Step, Figure 3, page 7, best-of-1,860: a process reward model at a share equal to 391/500, an outcome reward model at 362/500 and majority voting at 348/500, printed as shares; the paper does not say whether they count problems. Their two reward models pick by a model's score. The figures stand beside each other as published.

The same bytes on every Mac. Run again after the seal, all 815,632 answers reproduced their sealed outcome. The Affine IDE ships Study 56 as 24 files, 12,165,947 bytes, each pinned by SHA-256 and verified byte for byte before a figure is drawn: the law's 21 sealed outputs (9,737,089 bytes), the rail's index of all 12,500 problems (1,153,189 bytes) and two of OpenAI's published files (theirs, 1,275,669 bytes). Every committed file of both Study 56 evidence folders matches its own SHA256SUMS: 41 of 41 and 217 of 217. Read Study 56.

Frontier models, beside the substrate: they train their models and sample them; Affine.Earth reads every answer exactly

Frontier labs print mathematics figures built from sampled runs. OpenAI printed o1's AIME 2024 result as an average score out of 15, for one sample, for a consensus of 64 samples and for a re-ranking of 1,000 (theirs, 2024-09-12). MathArena asks every model each question 4 times and prints the average of the 4 (theirs). Asked again, the answers change: in MathArena's published AIME 2025 rows, 29 models at 4 runs each, the 4 runs did not return one answer on 465 of 870 model–problem pairs (theirs, counted by Affine.Earth from their rows). Thinking Machines Lab sent one prompt 1,000 times at temperature 0 and received 80 different completions (theirs).

Answers change even at temperature 0, the greedy setting. Thinking Machines Lab names the cause: kernels whose order of addition changes with batch size (theirs). Yuan and others: floating-point addition is not associative at limited precision (theirs). The substrate adds whole numbers: 1, 2 or 5 agents give the Study 56 goal one seal, c6949876a8639991.

Read by the law, OpenAI's own answers hold 33,586 different exact values across the 500 problems, and more than one value on 471 of the 500. test/algebra/2626 has 1,149 different values in 1,854 answers; on test/algebra/1004, 1,821 of its 1,860 answers hold 4, the key.

MATH and PRM800K were built to train models. MATH's own train folder holds 7,500 problems; PRM800K's 800,000 step labels trained OpenAI's process reward model, and 4,500 MATH test problems were moved into its training set (theirs, their README); the generator that wrote the 815,632 answers was finetuned on solutions to MATH's train problems (theirs). Affine.Earth trained on none of it: it read all 12,500 problems, train and test alike, as claims to decide. Frontier models in mathematics — a share over sampled runs, and one sealed reading of every key and every answer

The Affine IDE, new in 0.2.6.6: nine changes to the IDE itself

  • Study 56 on its own stage. Three pictures — Key, all 12,500 problems as a lattice by type, level and split with the 500 test problems ringed; Samples, the 500 test problems as columns of their 815,632 answers, the sealed outcome beside is_correct (theirs); Steps, one of phase2_test's 2,762 labelled model solutions (theirs) as a tower of its steps, which shows once phase2_test.jsonl is fetched and verified — told in six chapters in the chat. 456 figures in the chapters — 374 from a census line or a sum, the rest from a labelled structural entry — 0 untraced.
  • The Mathematics folder. MATH's 7 types in MATH's own order, then levels, then all 12,500 problems, 20 to a page. Pick one and it answers how the substrate read it: it closes, with its value — test/algebra/1072 closes to 243/625 — or it cannot close, with its reason: test/geometry/686 holds pi, and pi lies outside the integers and the fractions of integers. A key written with a decimal is cut off and never printed. A problem's own text and MATH's answer show once MATH.tar is fetched and verified, bytes and then SHA-256.
  • 25 quantum courts, each drawing its own court. Nine family pictures — rings mod N, phase wheels, the marked register, oracles, two-bit channels, exact energies, walks and words, eight-lane lattices, the key on the curve — every value from the court law's own call. The canvas checker refuses all 1,970 planted tampers across the 25. Before, all 25 drew the same protein picture.
  • A swarm of Affine IDE agents. Deterministic workers, never a language model, each owning a space of the sources in its own sandbox, sized from what the Mac reads about itself right now, never from a written number. Measured in the release tree, the Study 56 goal sealed one set of bytes, c6949876a8639991, with 1, 2 and 5 agents.
  • affine-ide-agentd, signed inside the app. 21,431,792 bytes, Developer ID, hardened runtime; a child of the IDE that ends when the IDE quits.
  • Plain words for every census code. Each of the law's 19 reasons a key cannot close carries the census's own sentence of why; 14 of them occur among MATH's 1,967 keys, and the Key view's legend names all 14.
  • The same bytes for every observer. Every study page, court answer and fixture carries only what the law prints from SHA-pinned bytes: 308 paths, and 94 fixture rows that hold one digest across 3 calls and 2 faces.
  • Proved before it ships. The release agent opened all 79 studies in the built app: 97 of 97 regression arms, and Franklin's sign-off bound to the disk image's own SHA-256.
  • Notarized and stapled. Apple accepted the app and the disk image, and the ticket is stapled to both.

Every domain, one exact fact

Each folder of the IDE's own sidebar, with one fact anyone can re-measure from the same bytes:

  • Mathematics — 10,533 of MATH's 12,500 keys close to an exact value and 1,967 cannot, each with its reason, across all seven types; 815,612 of OpenAI's 815,632 answers hold one verdict. Study 56
  • The Iliad — Book 1's shore scene is one file of 11,133,626 bytes that all nine cells serve with the same SHA-256. The IDE opens it only when the byte count and then the digest match, and draws it from 555 meshes — 45,674 points and 77,167 faces, every coordinate an exact fraction of two whole numbers — with no floating-point value in the file.
  • Physics — Four published files of IceCube's calibrated hits, 9,749 recorded events and 2,289,821 single hits of light, are read by Affine's own Swift; every hit is placed on the sensor that recorded it, and all four files are served with the same bytes on all nine cells. The Glashow event's 18,135 pulses, on 453 sensors across 30 strings, seal to 68360339…; the January 2013 run's 71 events seal to 237012b7…. Study 55
  • Astronomy — The Event Horizon Telescope's public 2017 measurements of Sagittarius A*, the radio source at the centre of the Milky Way, are 16 tables in one archive of 7,083,259 bytes; every table is pinned by its byte count and SHA-256, and 16 of 16 match. Study 07's six whole-number thresholds were frozen on 2026-07-23, after which the integers do not move — a short-baseline median of at least 1,500,000 µJy and a long-baseline median of at most 400,000 µJy among them; no program in either repository computes those medians from the 16 tables, so no substrate law decides them. Study 07
  • Earth and sky — A flight track at 30,000 feet along latitude 50,000 micro-degrees, beside a storm cell from 0 to 30,000 micro-degrees and 0 to 40,000 feet: the convective court answers PROVEN_CONTAINED, with 0 breaches. Study 09
  • Quantum courts — 25 courts check a presented quantum-algorithm answer exactly and search for nothing. Given 15, the witness 4 and the factors 3 and 5, the Shor court checks that 4 × 4 = 16 is one more than 15, that 4 is neither 1 nor 14, that 4 − 1 = 3 shares the factor 3 with 15, and that 3 × 5 = 15, and answers WIN. The court table
  • Courts — The exact-permanent court prints 6 for the 3 × 3 matrix of ones, and 29 for the matrix with rows 2, 3 and 5, 7. Study 25
  • Language games — The translate court reads one corpus of human sentences across 25 languages, pinned inside the law by its address, its 29,157,925 bytes and its SHA-256. A process that cannot read exactly those bytes refuses instead of translating, and any word outside the court's corpus for the other language is refused by name. Study 47
  • Impact studies — The unit square, stretched 12 times, holds 169 lattice points. From the 13 counts the law prints the counting rule t² + 2t + 1 and the area, exactly 1/1. Studies 11, 12 and 13
  • Domain stages — EUROCONTROL's open 2022 flight list, read in whole minutes, admits 525,541 departures. At the 20 airports with at least 5,000 of them, each airport's floor is the taxi-out time one twentieth of the way up its own sorted record, and the minutes above those floors add up to exactly 1,224,355 across 338,254 departures. Study 32
  • Health, finance, chemistry — The Protein Data Bank's list of entries, as fetched on 2026-08-24, holds 258,616 ids. The protein court reads that list and prints 258,616 in every answer, and 4HHB, whose title at RCSB (theirs) names human deoxyhaemoglobin, is one of them. All nine cells serve the list with the same 1,816,323 bytes. Study 14

Download Affine IDE 0.3.0.7 β · Read Study 56 · Read the research · download page · Install wiki · still beta · Developer ID signed, notarized by Apple and stapled · verify the SHA-256 on the download page.

Earlier (2026-10-07): the IDE proves itself before it ships, and IceCube's neutrinos are read from the hits

On 2026-10-07, the IDE introduced pre-roll release proof. Every release since then ships only after proving itself. Before the roll, a release agent opens every study inside the built app. It reads what is on the screen: which stage drew, whether the frame is lit, what the chat says. It writes all of that into one record. The court on the nine cells grades that record and seals it against the sha256 of the disk image it came from. If one study opens on the wrong stage, Franklin does not sign, and nothing reaches you. The sealed record that 2026-10-07 release rolled on: 96 of 96 arms over 78 studies, bound to dmg 4daa481b270542b5… (in 0.3.0.7 β, 121 of 121 arms over 79 studies).

Study 55 reads IceCube's neutrinos from the hits. IceCube publishes the light its sensors recorded as calibrated hits: a time, a charge and a sensor for each flash. The IDE reads those published bytes itself, byte for byte, with no float anywhere. It stands the detector's 86 strings in a hologram and lights each event sensor by sensor, in the order its light arrived. It read four published files, 9,749 events and 2,289,821 hits, and every hit joined to its sensor. Each file carries a census seal and an event-index seal, and any laptop re-derives both from the same bytes. The January 2013 run's 71 events seal to 237012b7…. The Glashow event's 18,135 pulses, on 453 sensors across 30 strings, seal to 68360339…. IceCube's own published table of 28 events appears beside the hits, labelled as IceCube's reconstruction. Where a neutrino came from, and how much energy it carried, needs reconstruction, the next tier. Open the study.

Your work stays whole. A file opens whole and saves whole: a 49-line file stays 49 lines. If something else changed it on disk, the IDE says so and writes nothing. The studies inside the app are read-only — fork one to make it yours. Your own folders sit beside the public studies under Public studies ▸ My projects, and the sidebar is now folders, with the open study lit.

One rule for every study: the same bytes for every observer. Every number a study shows must come from the substrate's own law, over bytes anyone can fetch, and read the same on every machine. We measured every study against that rule. Study 55 is the first brought into line, and this release leaves out the stage files whose numbers no law computed.

The current build, Affine IDE 0.3.0.7 β, is on the download page.


The closed system

We must treat the entire world as a closed system. The biosphere is the system, and we are one node in it.

Nature writes sparse integer tables. A figure appears on this page only when a law in this repository prints it from bytes with a recorded sha256.

CORRECTED 2026-10-08 — this line read “Every claim is validated against public archives, computed in exact integers, and verified by study numbers.” Measured in this tree, the figures the next section carried are printed by no law here; the line now states the rule the page keeps.

The sky: what this page states

This page once stated a hard ceiling for the stratosphere: a tonnage of synthetic aluminum oxide, a continuous metallic membrane past it, and a date. No law in this repository prints any of those figures, so this page states none of them. The chain they belong to (mass in, ozone, radiative forcing, biosphere) is the subject of Study 31, the Biosphere Cascade Court, which keeps its own record. The date is not known.

CORRECTED 2026-10-08 — this section was headed “What the court computed about the sky”. It gave 296,666 t, a saturation density of 1 particle per cubic centimetre against 1 to 10 measured, a run on the 1750 sky returning SATURATED, a membrane “refuted by fourteen orders of magnitude”, 492 kg per satellite, a tonnage of smoke, an HCl depletion below climatology and a mid-latitude ozone column loss (the last three written with a decimal point or as a percentage, and so left out here), and it graded the earlier model part by part. Measured: a search of the tree for 296,666, “492 kg” and “fourteen orders” finds only this page and its copies; no Swift source emits SATURATED; no Study 31 page is in the repository. The figures and the grades are withdrawn until a law prints them.


The realization

Sixty to a thousand kilometers up is a second sky — the ionosphere, the charged shell GPS travels through and aviation radio bounces off. When the shadow cone pierces it, the shell answers with a moving dent: charge collapses under the footprint and refills behind it. Read by depth alone, that dent is one number: how far TEC dropped. On May 11, 2024 a geomagnetic superstorm dug deeper dents than any eclipse, at every station, under an unobscured Sun. By depth alone, that day reads as an eclipse at all seven stations.

The substrate reads the dent as shape: the response either threads the shadow's own timetable — confined to the exact minutes the geometry names, deepest within seconds of each station's own maximum, marching station-to-station at the cone's ground speed — or it keeps another clock. Linked or not linked: an integer, no maybe. A storm can dig a deeper hole; threading the ring takes a cone, a track and a clock.

Because shape is the signature and shape-verdicts are exact, one machine sealed to-the-second claims about the sky 27 days before the eclipse of 2026-08-12 — including a verdict that stays decidable through a superstorm. The day after, the raw receiver files showed a depletion at every sealed totality station in our custody, inside its sealed window, in the sealed order. The graded verdict runs on the pinned archive once it posts eclipse day, and anyone on Earth can check both against the same archives.

Read in this order

Page What it holds
Overview — the realization The second sky, shape vs depth, everything sealed
The first 27 days Written the day after the eclipse: what failed, what held, four seals byte-identical
Science research The 2026 eclipse, measured shadow physics, the reference storms, the solar-cycle noise floor
Historical corpus Three eclipses, three solar regimes, one frozen law — the band history writes for 2026
Model shear The storm data, measured, and the shape discriminant it taught
Blind spots — five stories Five events a depth reading cannot separate, read from the shear-discriminant rows
Data archives Every public source, verified access mechanics, provenance checksums
Substrate architecture How a shadow becomes a geometry a machine can hold
Benchmark results The sealed 2024 verdict + cold-machine reproduction runbook
Prediction registry Both sealed claims, verbatim, with falsification conditions, and the first look from eclipse day
Success criteria S1–S5 decidable pass/fail, the storm contingency, failure semantics
Operations runbook Eclipse day in UT, the live loop, the T+1 resolution

How this stays true

The generating substrate lives elsewhere; after every ingest or verdict it regenerates the marked blocks in these pages from its ledger, and a drift gate proves the wiki matches the data byte-for-byte. The prose frames; the substrate speaks its own numbers.

Resolution — the first look from eclipse day is on the prediction registry. The sealed verdict, pass or fail, grades on the pinned archive once it posts the day, and lands in these same tables. The historical record already publishes its own falsifiers, because a recorded failure is how this experiment learned to see.


One method, many skies

The eclipse was the first sky to sign its name. Five more charters are now sealed and open: rocket burns that punch ~900 km holes in the ionosphere on a manifest clock, solar flares that step the daylit hemisphere in lockstep with the GOES X-ray record, tsunamis that keep a Huygens travel-time appointment at every tide gauge, cosmic-ray dropouts that strobe ~50 neutron monitors from a single L1 shock minute, and explosions whose seismic arrival ladder no earthquake can climb. Each study runs the same recipe the eclipse proved out — a forcing with a computable clock and track, a public raw archive, a sealed adversary that mimics magnitude but not shape, and a standing future event where a frozen exact-integer law must resolve in the open.

Start at the Shear Studies — Program index. Integer courts: Integer courts across domains · Studies 11, 12 and 13 — what the three laws print · stellar dynamo — the integer step court. Look ledger: Known discoveries · Known molecular discoveries · Discoveries by user. Playable keys: 14 PDB · 16 disease · 17 chemistry · 18 material. Combinatorial court: 19 Go First Dice. Dynamo court: 21.


Pick your door

The door is the chooser. Home → #chooser → Open this door → filled sample POST or Look GET → WIN + hologram + receipt. The walk.

FoT is the base. Public flourishing. 100k+ discoveries means two measured catalogs, never one fake number: PDB holdings 258616 · protein/material aggregates 125302 (78680+46622).

You are Act Open
Researcher LOOK affine #researcher · IDE Look
Chemist CLAIM InChIKey/CID Study 17 · #chemistry-inchi
Materials CLAIM COD/mp/sg Study 18 · #material-std · desalination
Disease / protein CLAIM ICD + PDB Study 16 · Study 14 · cancer-related · combinations · #disease-icd · #pdb-holdings · #complex-pair
Researcher hologram LOOK existing frame How the IDE hologram works · Low-friction flows · #chooser
Rife test claim CLAIM integer Hz/nm Study 20 · #rife-frequency
Stellar dynamo CLAIM integer (q,r)+n/d Study 21 · impact page · #stellar-dynamo
Manufacture contract LOOK CLASS + sha256 Manufacture contracts · #manufacture-look
Builder BUILD IDE chooser
Public / press SHARE Discoveries by user · family ledger · Studies 11, 12 and 13 · impact page · #stellar-dynamo

Each type’s goal, catalog N, industry key, WIN, and refuse examples: Discoveries by user. One story — Home → chooser → Open this door → filled sample POST → WIN + hologram + receipt: Low-friction user flows · frame sha this hour How the IDE hologram works.


Franklin chat frontier (sealed geometry)

On the main Franklin chat, two suggested starting points now walk sealed bonds:

  • Visual geometry — diagram and layout analysis grounded in sealed visual-spatial block-diagonal shape.
  • Agentic decomposition — multi-step code decomposition as Socratic prompts grounded in sealed agentic routing bonds.

Both return human residual speech only. Research story stays the wiki path; chat stays the peer.

Architecture taught by shear studies (measured 2026-09-17): ranked curriculum (Studies 11/06/07/12/13/22/25/09/14 + Minecraft) — Affine.Earth Architecture Taught by Shear Studies. Evidence SHEAR_STUDIES_TEACH_PIPELINE_PROVEN.

Affine Studio Cinema & NLE (AFFINE-SPEC-STUDIO-026): 705,600,000 Hz Flicks clock, OTIO/FCPXML interchange, 120-page Fountain script compiler, 16-bit tetrahedral 3D LUT compositor, 3D quadrance room acoustics, and 6-deliverable C2PA Merkle sealer — Affine Studio Cinema and NLE Specification. Evidence AFFINE_STUDIO_FULL_SUITE_PROVEN.

Zero-Float Swift 6.4 Live Avatar & 97-Language Coordinate Chart Registry (AffineLiveAvatarController + AffineTranslateCourt97Lang): 5 zero-float capabilities (60fps Non-Blocking Async MCP Tool Calling, 240-Sample Barge-In Crossfade & O(1) Arena Rewind, Single-Image Anthropometric Reference Identity Locking inside [100,900]³ LAW-FRAMING, Direct Eye-Gaze Vector Lock franklin SEE 626‰ via Cayley $SO(3,\mathbb{Q})$, and Mid-Stream 97-Language / 9-RTL Script Switching) plus Dual-Tier v1Core25 (25 Core, 4 RTL) + v2Global97 (97 Global, 9 RTL, 1,449,374 Study 58 evaluations with 0 cross-lingual shear) — Study 58. Evidence AFFINE_LIVE_AVATAR_AND_97_LANG_PROVEN.

Start here

Courts

Studies

All studies

More studies on this wiki

Discoveries

Cures — read in this order

The planet, and markets

Tools

Read

Watch

Clone this wiki locally