-
Notifications
You must be signed in to change notification settings - Fork 0
Affine Math Court Glama
The sky writes geometry. Quantum complexity does too.
Open MCP: https://affine.earth/language-invariant/mcp · Registry: earth.affine/math-court
Public example: clients/math-court-mcp · Live proof: bash clients/math-court-mcp/prove-live-court.sh
Evidence: clients/math-court-mcp/evidence/tools-list-20260827T194820Z.raw · edge nbg-01 · HTTP 200
IonQ, D-Wave, and every VQE notebook sell the same story: evolve a continuous wavefunction, sample a probability, call the average the answer. Affine.Earth discarded the wavefunction. The Math Court maps those problems as rigid intersections on an 8-dimensional flat torus and returns decimal strings — exact rationals, exact integers — that any agent can replay.
You do not need a lab seat. You need a JSON-RPC post.
The public client in clients/math-court-mcp computes no verdict. It is a visitor: it lists the tools, walks every court domain, names the four QMA solvers, and asks the cells to lock a Jordan bond. Measured on the live 9-cell apex:
| Proof | What locked |
|---|---|
| The court is public | Live tools/list on 2026-08-27T21:51:49Z advertised 43 tools (HTTP 200) — the 20 QC verifiers rolled to the cells the same day. Receipt: clients/math-court-mcp/evidence/VERDICT-20260827T215149Z.json (34 WIN · 0 MISS of 34, every QC fixture sealing its PROVEN marker). The earlier 23-count and 21-count walks are history. |
| The court is the whole court | One walk covered geometry, chance, algebra, physics, QCD, health, finance, compilers, fluids, eclipse, EHT, seismic, disease, chemistry, materials, PDB, Rife, dynamo, and the four QMA domains. 29 LIVE · 0 LEARN. |
| Miner verbs stay off the public surface |
mine, isolate, sha256d, submitblock were absent. The win is the verified bond, not a wallet. |
| Floats cannot enter | The wire is decimal strings. A JSON Float64 is refused. Continuous-quantum agents that speak IEEE never reach a seal. |
| QMA fixtures return exact geometry | 2-local energy −1/1. Spin-glass energy −1. N-representability WIN. 3×3 permanent 6. |
| The bond is the win |
verify_jordan_bond returned AFFINE_JZ_SHEAR_ZERO on the uniform lock. |
The Mac app (AffineNPHard court example) is the same walk with a desktop surface. It measured 30 LIVE · 1 LEARN. The LEARN row is Go First Dice, which wants sixty-face arrays — not a court failure, a different payload shape.
These are wired fixtures. A presented configuration is verified. An unknown ground state is not searched. That is the honest claim, and it is already stronger than a VQE notebook: the answer is a string you can print, not a distribution you can average.
| Study | Legacy pitch | Affine lock | Call this tool |
|---|---|---|---|
| 22 — 2-local Hamiltonian | VQE / continuous (e^{-iHt}) | energy −1/1 | execute_2local_hamiltonian |
| 23 — Spin-glass / TIM | D-Wave analog / QMC sign problem | energy −1 | route_spin_glass_manifold |
| 24 — N-representability | Float SDP / wave amplitudes | WIN | verify_n_representability |
| 25 — Matrix permanent | Float boson-sampling “supremacy” | permanent 6 | execute_exact_permanent |
Same fixtures accept math_court with domains qma_2local, qma_spinglass, qma_nrep, qma_permanent.
Full algorithm map: Quantum algorithms inventory. vQbit law: UUM-8D + vQbit.
Live HTTP: POST https://affine.earth/language-invariant/mcp tools/list → HTTP 200, 43 names after the 2026-08-27 cell roll (the 23 below plus the 20 QC verifiers listed in the next section).
In-repo mirror: cells/xcode/Resources/language-game/mcp-tools.json · CapabilityRegistry.servableEntries.
If those two disagree, the live HTTP wins. Re-measure: bash clients/math-court-mcp/prove-live-court.sh.
| # | Tool | Role |
|---|---|---|
| 1 | atc.assert_4d_deconfliction |
Exact 4D separation |
| 2 | corpus_bonds |
Sealed corpus bonds |
| 3 | corpus_capability_map |
Corpus census |
| 4 | corpus_coverage |
Fleet KV coverage |
| 5 | critique_frame |
Deterministic frame critic |
| 6 | execute_2local_hamiltonian |
QMA 2-local verifier |
| 7 | execute_exact_permanent |
#P permanent (n\le 3) |
| 8 | execute_transition |
Membrane / Rational S4/C4 lanes |
| 9 | feeds_catalog |
Feed catalog |
| 10 | game_frame_meta |
Language-game frame |
| 11 | ide_rebuild_mesh |
Founder-cell rebuild |
| 12 | lattice_op |
UUM-8D lattice algebra |
| 13 | math_court |
Domain ingest + catalog |
| 14 | membrane_health |
Membrane identity |
| 15 | noaa_goes_r_weather |
Live NOAA / GOES-R |
| 16 | route_spin_glass_manifold |
QMA spin-glass verifier |
| 17 | twin.robotics.evaluate_exact_ik |
Exact integer IK |
| 18 | umc_direct |
UMC GAV direct |
| 19 | umc_resume |
UMC resume |
| 20 | umc_status |
UMC status |
| 21 | verify_jordan_bond |
(J_Z) shear-zero |
| 22 | verify_n_representability |
QMA 2-RDM verifier |
| 23 | weather.convective_containment |
Exact convective containment |
Five catalogued ABSENT rows (mcpTool: false, empty primitive) are not served: twin.quantum.stabilizer_tableau, twin.chem.qubo_ising_docking, twin.fluid.lbm_step, twin.chem.steric_knot_check, twin.materials.symmetry_burgers.
ON GLAMA (live 43). If a name is in the live tools/list, it is on affine.earth MCP and any Glama agent can call it. Studies 22–25 are on that court (execute_2local_hamiltonian, route_spin_glass_manifold, verify_n_representability, execute_exact_permanent, plus math_court domains qma_*). Studies 11–13 use math_court domains geometry, chance, algebra. Fixtures and the caller: clients/math-court-mcp/fixtures/live-court-studies.json · bash clients/math-court-mcp/prove-live-court.sh.
LIVE ON GLAMA — 20 QC verifiers, rolled 2026-08-27. Law: QCCourtLaw.swift (9/9 law tests, floats refused at the wire). Membrane: CapabilityRegistry + CertifiedUserMCPMembrane + AffineMathCourt.qc. Live tools/list is 43, and every one of the 20 answered its fixture WIN with its AFFINE_QC_* PROVEN marker on the public apex (receipt above). Names: verify_shor_witness, verify_period, verify_grover, verify_qft_phases, verify_qpe_phase, verify_hhl, verify_deutsch_jozsa, verify_bernstein_vazirani, verify_simon, verify_amplitude_amplification, verify_amplitude_estimation, verify_phase_kickback, verify_quantum_walk, verify_vqe_energy, verify_qaoa_energy, verify_quantum_counting, verify_teleport, verify_superdense, verify_bell_measurement, verify_topological_word. Call table: Conjecture Alignment. These grade a presented witness. They do not search an unknown period, marked item, or variational landscape. vmNotConnected catalog bodies are not this surface.
STILL NOT TOOLS. run_quantum_algorithm, miner verbs, ECDLP-from-Q, VQE optimizer, QAOA pulses. Lean / substrate stay the research engines. Typed AlgorithmCatalog shells still throw vmNotConnected — do not expose those.
| Research object | Where it actually runs |
|---|---|
| Shor / period / Grover / QFT / QPE court | Named MCP tools above (after deploy) · local swift test --package-path LatticeRender --filter QCCourtLaw
|
| Shor witness Lean |
cd proof/lean && lake build FirstRoars.ShorWitnessCertifier · Lion-Shor-Witness-Certifier
|
| Shor ECDLP substrate |
cells/xcode/Sources/ShorECDLPSubstrate/ · Lean ShorECDLP.lean
|
| Studies 12–13 local prover | gaiaftcl prove conjecture-workload |
| Miner verbs |
mine, isolate, sha256d, submitblock, nphard — unproven, off tools/list
|
Full map: Quantum algorithms inventory.
bash clients/math-court-mcp/prove-live-court.shThat script POSTs the live apex, writes clients/math-court-mcp/evidence/, prints WIN/MISS, and exits 2 if the required names drift or a study fixture misses. Count 23 (pre-roll) or 43 (post-roll) both pass; the live apex serves 43 as of 2026-08-27T21:51:49Z, and the DEPLOY_REQUIRED state is history — a re-run today prints 34 WIN · 0 MISS.
swift run --scratch-path /tmp/affine-math-court-mcp --package-path clients/math-court-mcp/swift-example affine-math-courtswift run --scratch-path /tmp/AffineNPHard-harness --package-path apps/AffineNPHard AffineNPHard court examplePayloads are decimal strings. If you send 0.5, the court refuses you. If you send 1/2, it grades you.
Legacy quantum computing needs a cryostat, a microwave schedule, and a press release about “advantage.” Affine.Earth put the same problem class on a public MCP that any Glama agent can call from a laptop. The cells do not anneal. They do not descend a gradient. They project the presented string onto the torus and either lock or refuse.
That is the demonstration: the hard problems they advertise are already a court, and the court is already open.
- ⚛️ QUANTUM-ALIVE — 24 courts LIVE
- The lattice holds
- Impact study — continuum dead
- Death of continuous shear
- Fourier Phantom — Anima FNO vs 11+12+13
- Stellar dynamo kill shot
- QCD: freedom is dilation
- Look in the UI (no visitor data)
- Explore the live courts
- MCP clients (public)
- Example app — entire court
- Math Court user guide
- Math Court on Glama
- Quantum algorithms inventory
- Glama connector
- Zero Float · Zero Shear
- UUM-8D vs IUT — WIN
- Readers’ guide
- White paper
- Program index
- Roadmap — what comes next
- Language-games study board
- Known discoveries — family ledger
- Discoveries by user
- Low-friction user flows
- How the IDE hologram works
- Known molecular discoveries
- Study 06 — Explosion Results
- Study 07 — Milky Way Results
- Study 11 — Ehrhart Volume
- Study 12 — Parallel Repetition
- Study 13 — Connes Rigidity
- Study 14 — Protein lattice
- Study 16 — Disease type
- Study 17 — Chemistry InChIKey
- Study 18 — Material STD
- Study 19 — Go First Dice
- Peer-review bundle
- Conjecture alignment
- Study 02 — Launch holes
- Study 02 — Regulatory alarm
- Build a study — Falcon
- Study 09 — Convective bond
- Study 20 — Rife frequency
- Study 21 — Stellar dynamo
- Study 22 — Ground state
- Study 23 — Spin glass
- Study 24 — N-representability
- Study 25 — Permanent