Aetherean · ASKA Digital · Foundation-0 · Moissanite Labs · Proven-Commons

Frontier

Every way we have pushed the frontier: new knowledge, methods, findings, theories, calculations, and what is still open.

38entries 19verified results 2026-11-17last updated Updated as new work lands. See how below.

Verified results first. Draft findings are labelled draft and stay under review until their synthesis lands. Open questions and negative results count too: a failed prediction, honestly recorded, is a contribution.

Verified results

Measured, merged, or live. Numbers carry their exact context; limits are stated, not buried.

Quantization · SiliquantVerified results

E8 beats the industry quantizer at equal bits

Rounding a real model's weights with the E8 lattice damages its output less than llama.cpp Q4_0 at the same bit count.

Gemma 4-E4B, one scale per 32 weights from the largest entry, 10,200 positions. Mean KL at 4.5 bits: E8 0.075 vs Q4_0 0.117 (first-choice word kept 91.6% vs 89.4%). At 4.005 bits: 0.215 vs 0.448. 10 of 18 preregistered predictions hold.

Limit: One model, one text. No packed file format exists yet and no decode speed was measured, a scored result, not a shippable quantizer.

Moissanite-Labs/siliquant PR #2 · 2026-11-15

Quantization · SiliquantVerified results

The integer ball beats E8, and the lattice is a small slice of the gain

Cutting weights to a ball, with no lattice at all, beats E8 and the industry baseline. The E8 cell shape is worth only about 0.65 dB; most of the gain is scale plus ball.

hZBG128 (ball + half-strength importance matrix): KL 0.0209 on Qwen2.5-1.5B, 0.0282 on SmolLM2-1.7B, ahead of matrix Q4_K_M on both. Decoder tables: E8 561 MB, integer ball 31 MB, predicted ZBG64 4 MB.

Limit: Branch result, not on main (branch bayan/charming-gates-jlas14, PR #12 draft). Qwen record partial, 27 of 32 variants; SmolLM2 complete, 13 of 13.

Moissanite-Labs/siliquant PR #12 (draft) · 2026-11-17

Quantization · SiliquantVerified results

Exact closures: billions of weights provably do nothing

Large parts of a real model can be deleted exactly, with zero output change on the tested text.

2,794,842,624 weights zeroable for this text; 55,050,240 weights zeroable for any text, both scoring exactly as the unchanged model.

Limit: Exact on the tested text and model (Gemma 4-E4B).

Moissanite-Labs/siliquant PR #2 · 2026-11-15

Quantization · SiliquantVerified results

The scale rule matters more than rounding precision

How you choose the scale matters far more than how precisely you round.

Clipping 0.17% of weights (−27.5 dB weight error) gives mean KL 0.318, vs 5-bit rounding of all weights (−21.7 dB) at KL 0.118. Best damage predictor tried: largest singular value of the weight change, rank correlation 0.889 over 26 variants vs 0.603 for squared error.

Limit: Same Gemma experiment as the E8 result above.

Moissanite-Labs/siliquant PR #2 · 2026-11-15

Quantization · SiliquantVerified results

The zero-codeword pitfall is model-dependent

E8's codebook must include a zero point or it breaks, and whether that matters depends on the model.

Without a zero codeword, E8 Voronoi breaks per row: 3.658 vs the scalar rule's 0.632. On Qwen the zero codeword helps at every level; on SmolLM2 it is 7% worse with groups of 32. Why it flips is unexplained.

Limit: Verified measurement, open question.

Moissanite-Labs/siliquant issue #16 · 2026-11-17

Steering · HeedleVerified results

Heedle v1, steer a small model by hand, inside checked bounds

A person can steer a small model's change by hand, propose, steer, hold, check, keep or discard, with every step checked and replayable.

Two hosts: cloth/1 (467 weights, point to colour) and text/1 (4,700 weights, 3 chars to next char, dispute exhaustive over all 21,952 contexts). 30 Python + 21 JS tests pass on the merged head; both sample drafts verify; CI green. The verifier hole was independently re-forged on both hosts and caught.

Limit: Code and tests only in the review; no browser pass, no re-run of the headless Chromium export.

Moissanite-Labs/heedle PR #1 · 2026-11-17

Design · Weft-eidoVerified results

Weft-eido #21, tide palette, interference map, loom research

A new take on the whole design language: tide palette, interference map, and research on the loom.

118 tests pass, 0 fail, on a clean-checkout run; all CI green.

Limit: Code and tests only in the review; the author and both earlier approvers did the browser pass.

Moissanite-Labs/weft-eido PR #21 · 2026-11-17

Verified software · Proven-CommonsVerified results

Lean-verified libraries: parsers and codecs with proved round trips

Twelve format libraries rebuilt in Lean 4 with machine-checked proofs, parsing and printing that round-trip by proof, not by test suite alone.

MessagePack decoder Bencode encoder/decoder (200,000/200,000 requests, 0 failures in review) UUID (RFC 9562, 10,562 inputs, 0 failures) ZIP reader JSON parser/printer RFC 3339 timestamps UTF-8/UTF-16 codec IP parser/printer TOML parser/printer (77/77, 300/300, 60/60, 1869/1869, 0 failures) codec pack (hex, base64, CRC-32, Base32, percent-encoding, PEM, Luhn/IBAN) protobuf wire codec

anthology compilation layer.

Limit: Claims are per-slice: parser first, printer separate. The HTTP request parser is still open (PRs #3/#4). The IP claim scope, parser-only vs expanded to printer, is still undecided.

Proven-Commons org (40 repos) · 2026-10-26

Search · SokobanVerified results

An 84-key route breaks the Sokoban holdout

A verified 84-key route beat the published 87-key record and the benchmark creator's unpublished 85-key holdout, cutting the trilogy shortest-known sum from 3,064 to 3,063.

Room III-20, Heroes of Sokoban III. Route sha256 eecbea96b091, credited to a1kindus on the live Frontier, replayed key by key in the original PuzzleScript build.

Method: the board is vertically symmetric in walls, goblins and goals but chirally asymmetric in hero placement, switch Wizard to Priest immediately and walk the Priest down instead of up; found by role-agnostic vertical-mirror Tube Beam search plus iterative tube refinement.

Limit: 84 is the known record, not proven optimal: the proven floor was 74 at the break and is 77 now.

ahrazzle/soko, docs/frontier-dossier.md · 2026-11-13

Search · SokobanVerified results

A verified lower bound: no III-14 route shorter than 38

A machine-checked proof that no route shorter than 38 exists on the true 15×7 board for room III-14.

Finished f=37 layer: 117,766,522 unique canonical states, 181,959,400 edges verified, 0 Bellman violations, 82.41 s; reproduced by an independent prover.

Limit: Still 6 short of the live floor of 44, the next step is pushing f=37 to f=44.

ahrazzle/soko PR #43 · 2026-11-13

Mathematics · Möbius transformVerified results

Full E1 network check for the GF(2) Möbius transform

The missing full-network run for Theorem 2 of the o(N log N) GF(2) Möbius-transform claim, built locally and contributed upstream.

h = 8–11, zero failures, both controls refuted, against Yates' 1937 bound.

Limit: Contribution to Swapnil Jain's mobius-x1; the gate-count figures reported elsewhere were not re-checked here.

Swapnil-jain/mobius-x1 PR #1 · 2026-10-11

Agents · OptChat memoryVerified results

optchat-memory v1.0.0, endless memory for Hermes Agent

An OptChat-style endless-memory provider: append-only log plus binary summary tree as recall, implementing Victor Taelin's OptMem spec.

Release v1.0.0, 2026-10-11; catalog PR NousResearch/hermes-agent#132907 merged; site live.

aska-digital/optchat-memory · 2026-10-11

Review · ExoVerified results

A claimed ssh gate hole that did not reproduce

A claimed security hole, keep the host read when an ssh remote turns credentials, does not reproduce: the hole is absent on base, and the new test passes unchanged on base, so it pins nothing. The real effect was five different shapes the PR never mentioned.

Review REQUEST_CHANGES on exo PR #664, head fd0e6d3d, 3 inline comments; a second changes-requested review the same day. The PR is still open and blocked, the catch held.

Limit: Full suite not rerun in that review session. The contribution here is the catch, not a shipped fix.

Moissanite-Labs/exo PR #664 · 2026-10-27

Rigidity · Foundation-0Verified results

880 real realizations at 10 vertices

880 realizations confirmed by three independent implementations.

Deposited on Zenodo, 2026-11-11, DOI 10.5281/zenodo.24145786.

Limit: Internally verified; externally unverified.

Foundation-0/rigid-realizations PRs #16–#19 · 2026-11-11

Lattices · Foundation-0Verified results

Incentre lattices v3.0.0

The incentre construction through honeycombs, borides, and rigidity in n dimensions; lattice work verified in dimensions 2–9.

Release v3.0.0, 2026-11-03; elastic proof with exact motion to FCC.

Limit: A mathematician's line-by-line read of the elastic proof is still pending.

Foundation-0/incentre-lattices · 2026-11-03

Geometry · Foundation-0Verified results

Vitruvian3d motion and conductivity

Local rotating-motif motion inside a fixed exterior, proved; conductivity rank contrast from certified passages.

Merged PRs #9/#10, 2026-11-12–13.

Limit: Internal AI review only; no human mathematics review yet. The motion does not transfer to quantization, tested and ruled out.

Foundation-0/vitruvian3d · 2026-11-12

Codes · Foundation-0Verified results

GF(q) code bounds improved

Exact Delsarte certificates improve code bounds over GF(7), GF(8) and GF(9).

Release v1.0.0, 2026-11-11; 12 GF(9) bounds lowered by one, two GF(7) codes excluded.

Limit: The wider ≥2,418-entry scan is still draft.

Foundation-0/gfq-codes · 2026-11-11

Physical trade · ASKA PhysicalVerified results

The Dock fused story, R&D floor plus verified trade stack

Five R&D modules paired to four fail-closed gates (brief, qualify, quote, berth) in one narrative: hardware floor plus verification layer.

Dock PR #6 merged 2026-11-17; OptiFab campus model (10 MW IT on 2,500 m², PUE 1.074) added as #/optifab via PR #7, merged the same day. OptiFab calculator outputs harness-verified against its model.

Limit: The five module figures (Taper-Lock, Kerf-Box, GeoBinder, Mill-Eye, Pooled-Cut) are presented by the deck and site, not independently checked. OptiFab v0.2.0 is claimed and unreviewed.

aska-physical/dock PRs #6, #7 · 2026-11-17

Governance · EthosVerified results

Ethos MEP v0.1, a minimum viable ethos protocol

The most portable already-shipped item from the November evaluation: a minimum viable protocol for ethos records, private by default, publication opt-in and granular, with revocation.

Merged 2026-11-14.

Limit: Exact PR number not yet pinned; link to be added once verified.

Moissanite-Labs/ethos (per internal-stoa#28) · 2026-11-14

Draft findings

Strong claims, internally certified, not yet published. Held from any publication claim until the full synthesis lands.

Electrostatics · Foundation-0Draft findings

Four charges, 11 balance points, past Maxwell's bound of 9

Four charges admit 11 equilibrium points, the first breach of Maxwell's 1873 bound below five charges.

3 configurations certified by two exact programs plus a lower-bound prover. Fragile: one part in 100,000 destroys them.

Limit: DRAFT, certified internally, no outside check yet. Held from publication until the full synthesis lands.

Foundation-0/point-charges PR #22 (draft) · 2026-11-15

Rigidity · Foundation-0Draft findings

2,288 real realizations at 11 vertices

All 2,288 complex realizations of an 11-vertex rigid graph are real, growth 2.4116 per vertex against the published bound of 2.3780.

1,144 boxes × mirror; a separate exact checker confirms. A 12-vertex push is partial: 3,904 of 6,180 real, ceiling 2.394 per vertex.

Limit: DRAFT, internally confirmed, not a priority claim, and held from publication until the synthesis lands. The search path is not step-replayable; the certificate stands alone.

Foundation-0/rigid-realizations PR #21 (open) · 2026-11-15

Electrostatics · Foundation-0Draft findings

Exact point-charge counts: 49 at N=8, 64 at N=9

Both counts equal Maxwell's bound; neither beats it. Counts for 3–7 charges (4/9/24/29/44) are proved by a separate program.

49 (N=8) and 64 (N=9).

Limit: DRAFT, unvalidated. The science pass covers only up to 7 charges.

Foundation-0/point-charges · 2026-11-15

Gravity · Foundation-0Draft findings

A gravity threshold model: three versions fail, one residual survives

Zero-parameter, redshift-falling, and fixed-proportion versions all fail against the data; only a large-scale residual above roughly 230–345 km/s survives, and the acceleration-scale fit is undecided.

Failures at 9σ and 0/41 galaxies among others. The deciding observation, lensing at 200–300 km/s, has no systems in the samples.

Limit: DRAFT findings, not peer-reviewed. The 0.0289 a₀ crossover is unvalidated.

Foundation-0/gravity · 2026-11-15

Sequences · Foundation-0Draft findings

238 OEIS recurrences proved, 227 of them new

Finite-state certificates prove 238 recurrences, with a square law and a zero-vector-test finding alongside.

238 proved (227 new); 11 recurrences merged earlier, 2026-10-28.

Limit: DRAFT, PR #5 is a draft.

Foundation-0/finite-state-proofs PR #5 (draft) · 2026-11-15

Celestial mechanics · Foundation-0Draft findings

Central configurations for equal masses

Existence certificates for 8, 9 and 10 equal masses, with class counts 20 / 42 / 67.

Thread still running at last report.

Limit: DRAFT, open PR, thread incomplete.

Foundation-0/relative-equilibria PR #16 (open) · 2026-11-15

Live explorations

Working instruments, sites, and hubs you can open and try. Proposals are labelled as proposals.

Quantization · SiliquantLive explorations

Siliquant results site, live

The empirical quantization results published as a live page.

38 of 40 synthetic cells and 7 of 8 real cells beat the scalar rule, re-counted 2026-11-17 from all 48 rows; only F8t at K=16 fails.

Limit: The record prose still says 34/40 and 6/8, stale, not the live site. Fix waits until the branch is filed as a PR.

moissanite-labs.github.io/siliquant · 2026-11-17

Steering · HeedleLive explorations

Heedle, live in the browser

The steering instrument runs in the browser for anyone to try, with a plain-words guide.

Pages deploy and guide both merged 2026-11-17; /hosts/text-v1.json serves 106,028 bytes.

moissanite-labs.github.io/heedle · 2026-11-17

Optimization · Causal Harmonic ComputingLive explorations

Causal Harmonic Computing hub, a proposal, live

Solve combinatorial optimization with coupled oscillators: TSP mapped to N² oscillators, coupling from the Hamiltonian's second derivatives, claimed O(N log N) settling vs O(N!), with hardware paths from MEMS to Josephson junctions and a full Python Kuramoto TSP simulation.

Hub: overview, Weft, Explorer, infographic, paper, soundtrack (32 timed lyric cues), and all 12 original uploads downloadable untouched. Live at eido.askaconsult.com/site/causal-harmonic/ since 2026-11-16.

Limit: PROPOSAL, not a result. The claims are Gemini's assertions, not verified, including the ~10^150x and ~10^300x speedup figures. The Works Cited still has an unknown_url placeholder to clean before any publishing.

Moissanite-Labs/weft-eido PR #19 · 2026-11-15

Atlas · Foundation-0Live explorations

Convergence Loom atlas

An interactive atlas of the unification-of-sciences explorer at fdn0.org/loom.

PR #59 open, mergeable clean, CI green; needs a non-author maintainer approval.

Limit: The numbers on it (49/64 point-charge counts, 0.0289 a₀ gravity crossover) are presented, not signed.

Foundation-0/f0 PR #59 · 2026-11-15

Life sciences · Foundation-0Live explorations

bi0, life-sciences hub

A hub for life-sciences work: README, areas, licences, and a roadmap.

PR #1 approved 2026-11-17 by a non-author maintainer, mergeable clean, docs-only, approved, unmerged.

Limit: Setup only so far; the research content is the roadmap.

Foundation-0/bi0 PR #1 · 2026-11-17

Open questions

Bets now on the table, filed as issues. The frontier is partly a list of questions.

InterpretabilityOpen questions

A parameter map from a distilled shadow model

Treat data as a flow through a model; push varied flows through it and watch which parameters light up. At billion-parameter scale that is infeasible, so distill a minimal-parameter shadow of the architecture, run high-frequency tests on it to build a generalized parameter map, then apply that map to larger models.

No experiments run yet. Filed as issues on siliquant #21, heedle #2, and weft-loom #32.

Limit: Proposal by A, 2026-11-17.

Moissanite-Labs/siliquant issue #21 · 2026-11-17

Steering · HeedleOpen questions

Automate Heedle, as a plugin, not a new project

Automating the Heedle workflow is an automation feature of Heedle, not a new project: automate the checking, keep the choosing human. At billion/trillion scale, pull strands the same way.

Filed as heedle #3. A tide host (tide/1, a colour palette as a model, exhaustive contrast dispute) is filed as heedle #6; palette steering as weft-eido #24.

Limit: Open proposal; nothing built yet.

Moissanite-Labs/heedle issue #3 · 2026-11-17

Quantization · SiliquantOpen questions

Does E8 generalize? Can a Foundation-0 lattice beat it?

Two open checks on the quantization win: a second model and a second text, and whether any Foundation-0 point set beats E8.

Answered for the tried sets: no F0 point set beats E8 at 2+ bits. The ball, not an F0 lattice, is what beats E8. Qwen + SmolLM2 runs answer the generality question in substance; the issues stay open until the records land.

Limit: Open issues #3 and #4.

Moissanite-Labs/siliquant issues #3, #4 · 2026-11-16

IntegrityOpen questions

Fail-closed certification, the convergence across everything

One discipline across Foundation-0 certificates, the Proven-Commons Lean bar, the Dock gates, and the CFRP certificates: a check must never print PASS when it should not.

The fault that started it: bare-Python asserts are skipped under python3 -O, so a corrupted certificate printed VERIFIED and exited 0. Fix batch: formal-claims-audit #19, vitruvian3d #13 (340 asserts, CI green), gravity #28/#29. Cleanup queue: 35 verdicts would change a label, 9 Zenodo records hold no result yet.

Limit: Fix batch still unmerged; the scale-pin decision (pin the enclosure scale the verifier never checks) is still owed.

aska-digital/internal-stoa#28 · 2026-11-16

Negative results

Things tested and ruled out, and claims we will not make. Recorded so nobody re-pays the cost.

Quantization · SiliquantNegative results

The closed-form output law fails on the third model

A formula that predicts a codebook's weight error from first principles works; the same formula does not predict model output on the third model.

Of 6 output-law predictions committed before SmolLM2's ball arms were scored, only O1 holds. One KL constant per lattice on SmolLM2, where Qwen shared one constant across arms.

Limit: A failed prediction is a finding, not a gap.

Moissanite-Labs/siliquant PR #17 (draft) · 2026-11-17

Quantization · SiliquantNegative results

Vitruvian motion and the F3(s) fold do not transfer

Two Foundation-0 ideas were tested against quantization and ruled out: vitruvian motion controls access ranks and rotational freedom, not cell second moment or decode cost; the fold-as-annealing schedule would not beat Lloyd at equal bits.

A fitted codebook over 3 weights is 3.4–9.8 dB ahead of the scalar rule per row, the side finding that survived.

Limit: Verified negative results.

Moissanite-Labs/siliquant issues #9, #4 · 2026-11-17

Search · SokobanNegative results

No new proven minimum, stated plainly

Nothing in the Sokoban work beats the live Frontier floors for any room. The 84-key route is the known record; the proven floors are 44, 77, and 270.

Live floors (2026-11-15): III-14 44 / known 48; III-20 77 / known 84; III-21 270 / known 316. Totals: 3,006 proven / 3,063 known.

Limit: Verified negative, this page will not claim a new minimum.

soko.banteg.xyz/frontier · 2026-11-15

Exoplanets · Foundation-0Negative results

Withheld: the exoplanets frontier

22 confirmed candidates from 18 stars at SNR 3.0 were claimed, and are not listed as a result, because the evidence does not hold up.

3 of 5 file hashes did not match; all 348 clean targets carry RUWE 1.0, a placeholder for a missing Gaia value. Ledger regeneration and a Gaia re-run are pending.

Limit: Unverified. Listed here only so its absence is explained, not hidden.

Foundation-0/exoplanets-frontier · 2026-11-15

How this page stays live

One file is the source of truth: frontier.json. Each entry carries its claim, its numbers in exact context, its limits, its date, and its source. Adding a contribution means adding one entry and redeploying. No rebuild from scratch, no rewriting the page.

Sources: Foundation-0 · Moissanite Labs · Proven-Commons · ASKA Digital · ASKA Physical