Vibe Theory

A discrete possible universe Ω

The goal

Find the discrete base of the universe. The working answer is a 4d hyperbolic mesh, the honeycomb $\lbrace3,4,3,4\rbrace$, whose every site holds 24 ternary values, updated by one deterministic, reversible, integer rule. Space and its dimension, time's arrow, relativity, spin, light, the quantum, gravity and selves each have to come out of that base, or the base is wrong. This repository tests that one measurable claim at a time, and keeps every result that said no.

The program draws on Maurice Margenstern's work on cellular automata in hyperbolic spaces: the $\lbrace7,3\rbrace$ heptagrid and the $\lbrace5,3,4\rbrace$ dodecagrid, Fibonacci-tree addressing and navigation, and the proofs that these grids host universal computation. The substrate survey and the computation experiments are measured against those results.

The companion pieces

• Short high-level audio overview of things
• Vibe (Book)
• Vibe Theory: A Discrete Hyperbolic Substrate for the Emerging Conscious Universe

The model, term by term

Each term is defined before it is used, with the reason it is this and not something else.

term	what it is	why
tone	one trit: +1, 0 or −1. The whole state is tones	three is the smallest alphabet with an empty value and a sign flip (charge conjugation) (E-FND-0038)
vibe	a slot holding +1 (a love) or −1 (a fear). A slot holding 0 is calm	the two signs are particle and antiparticle. Charge is love minus fear, content is love plus fear
dock	one site of space: 24 slots, one per root of the D4 root system, which are the 24 vertices of a 24-cell. Opposite slots pair into 12 lines	D4 is the one 4d root system with triality (E-FND-0038), and its 24 directions carry the spinor double cover, so a 2π turn gives −1 (E-SPN-0042)
mesh	the space of docks. In the model it is the $\lbrace3,4,3,4\rbrace$ honeycomb: 24-cells whose vertex figure is the flat cubic honeycomb $\lbrace4,3,4\rbrace$. In the experiments it is a flat periodic D4 box of side 8, 16 or 24, a disclosed stand-in	the box has the same 24 directions per dock and is finite and exact. The honeycomb unfolds from one 24-cell in shells of 1, 24, 456, 8,376 and 153,192 docks, a ratio settling at 18.278 (E-GMT-0027)
beat	one tick of time. Every dock applies the rule at once	the rule is synchronous and deterministic, so a run is reproducible bit for bit
take	how anything gets anywhere: each beat a slot takes its neighbor's value along its direction (the stream). Nothing moves	a value is read and written, never carried, so the stream is a permutation and runs backward exactly
husk	the 3d surface on which physics is read: the depth-even half of the 4d box, exactly closed (E-FRC-0168). In the honeycomb it is the flat horosphere at a cusp	the vertex figure $\lbrace4,3,4\rbrace$ is flat 3d, so the mesh has flat 3d surfaces at infinity, and a spectral dimension near 3 is read there (E-GMT-0025)
bulk, depth	the 4d mesh behind the husk. A husk column is the stack of bulk docks behind one husk dock, and its size D is its depth	every husk field is a column sum of bulk trits, so the depth sets how finely the husk counts (E-FRC-0207)
role, link	each vibe carries a point on a 3 by 3 grid (its role, a qutrit). A link between two docks holds a grid move, one of the 648 elements of the qutrit Clifford group Σ(648)	the links are exactly that group (E-QTM-0117). A link stores no matter: it is the relation between the vibes on it
store	a dock line's held love and fear pair, made from an empty line and released back to it	this is how the vacuum makes and unmakes pairs while its charge stays 0
amplitude	the state is a finite sum of classical configurations, each weighted by an exact number in $\mathbb{Z}[\omega][1/2]$, $\omega = e^{2\pi i/3}$	every three in the model (charge in thirds, the knot $1 + \omega + \omega^2 = 0$) lives in this one ring, and it needs no rounding
path	one measurement record of the quantum rule: at every split one branch, chosen by a deterministic key	a path is what one observer would record. A path can scramble while the sum it samples stays linear, so fields are read as averages over paths

The rule, as it stands

Each beat applies the coin, then the meetings, then the collision, then the stream. Every piece is an exact, reversible map in $\mathbb{Z}[\omega][1/2]$, with no float, no rounding and no random number, and runs backward piece by piece.

piece	what it does	why this one	code
the knit	the pieces below, keeping W(F4), the full symmetry of the dock (1,152 elements)	a knit that keeps all of W(F4) exists and keeps CPT, so no direction of the dock is preferred	E-RLT-0061
the collision	a dock's full lines turn as −1 and a lone vibe goes straight through	a lone vibe's wake stays on its own line, the free particle's straight path	E-RLT-0084
the doublet lock	a vibe's take direction along its line is its role's spin doublet	the only way a vibe can move with amplitudes and stay covariant (a theorem), and what makes every charge-one state a spinor	E-RLT-0097, E-SPN-0071
the meeting	a love and a fear on one line pass unchanged. Two like vibes keep their points (weight 1/4) or exchange them (3/4), leaving a knot with CHSH √7	the love and fear case is the identity by the conjugation convention, which is helicity suppression	E-RLT-0099
the contact	two like vibes on a full line keep their slots (the pass, phase +1) instead of bouncing (−1)	of the six phases the ring allows, only the bounce and the pass add no new number, and only the pass binds the electron	E-SPN-0092, E-SPN-0093
the store	pairs made and released with no veto, each stored pair holding two points	the only vacuum measured that holds under the pass	E-RLT-0102, E-RLT-0103
the coin	on a line holding one open vibe (a vibe that carries amplitude), $2C = (1 + \omega) I + (1 - \omega) X$ on its two slots	it gives a lone vibe a mass (gap 2π/3, top speed 1/2 a beat). Its cost is exact lattice momentum, the one symmetry that forbids a mass	E-SPN-0090, E-SPN-0091

The light runs beside the knit on the links: every husk angle, flux and counter is a column sum of bulk trits, with coupling $\kappa = 2/(2D + 1)$ (E-FRC-0207). A bounded, reversible column of $2D + 1$ values is a circle, so the light is compact U(1) exactly (E-FRC-0244, E-FRC-0245).

Where the quantum comes from. Without the meeting (the fear beat) the knit is Clifford and can be simulated classically. The meeting is the one gate outside the Clifford group (E-QTM-0118). Under the lock, motion adds a second source: no locked direction is a stabilizer state, so a moving vibe carries a fixed fear share of 0.164 (E-RLT-0099). A static vibe is classical and a moving one is quantum (chosen 2026-09-26).

The working vacuum (chosen 2026-09-26) is the no-veto store under the pass, with the coin on, and one C domain with no walls. Frozen domain walls were dropped because in cosmology they are a liability, not a fact to reproduce. On 51 of 51 gated cases (E-RLT-0105):

• the vacuum keeps its conservation laws, because no vacuum vibe is ever alone on its line, so the coin acts on it 0 times on every path
• the run reverses exactly, charge conjugation (C) holds, and every husk dock is one causal world
• pairs made equal pairs unmade, 466,944 each, on a cycle of 12 beats (E-GRV-0072)
• a lone vibe's wake stays on its line, at most 16 trits a period
• a lone love spreads over 105 to 124 occupations by beat 16 on 17 of 17 starts, and its positions interfere: against a twin run with its phases dropped each beat, the occupation distance reads 0.69 to 0.71 by beat 6 in the vacuum (0.60 on the husk columns), and exactly 0 with the coin off (E-QTM-0157). The earlier reading that the kept branch does not interfere used a configuration only one history reaches, so it could not see interference

The rule has changed often, and old results stay tied to their rule. Earlier committed rules include pairCollision, lineWeave, the turning weave (turningWeave, adopted 2026-09-02), the fear weave and the combined knit (E-FRC-0159). Most pieces above were chosen on 2026-09-26. Every experiment states the rule it ran on, and many held results ran on an earlier rule and have not been rerun on this one.

Where it stands

The working ledger (kept beside this repository, not in it) lists 191 facts the model must produce, from the Born rule to the specious present, each graded under one set of rules: read on the 3d husk, the rule integer and reversible, a stand-in labeled and never graded as a derivation, a numerical match counted only through a pre-registered null (E-MTH-0010), and the verdict judged over 17 link starts (E-MTH-0028).

status	meaning	facts
held	measured on the rule itself, with a control, over the start family	38
partial	measured, but on a stand-in, one sector, one start, or with a gate failing	67
stand-in	only a textbook calculation or a hand-built object the rule has not produced	45
open	worked on and not reached	24
none	not started	17

By area, as of 2026-09-29:

area	held	partial	stand-in	open	none
quantum foundations	13	11	3	5	1
special relativity	3	7	0	0	0
spin and statistics	4	4	0	0	3
light and charge	6	11	0	1	1
atoms and chemistry	0	3	7	0	1
the strong force and nuclei	0	3	3	2	1
the weak force and the Standard Model	0	3	6	2	1
gravity and spacetime	1	0	11	1	0
cosmology	0	4	5	4	3
heat, statistics and time	3	3	1	0	0
matter in bulk and fluids	1	0	1	2	2
the vacuum and walls	4	0	0	0	0
the classical world	1	0	2	1	0
information and computation	0	2	2	0	0
the numbers nature fixes	1	1	4	2	1
experience, selves and consciousness	1	12	0	2	3
the whole, consistent	0	3	0	2	0

The rows that moved on 2026-09-28 and 29: positions in superposition and fluctuation-dissipation to held, and 16 rows from none or open to partial, among them one light speed, parity violation, the path integral, length contraction, Compton scattering, the horizon and flatness problems, Landauer's principle and error correction. The remaining 17 none rows are mostly ones that wait on bound matter moving across lines (fine structure, spin-orbit, crystals, magnetism, fission, beta decay).

The gravity rows count the depth register below as a stand-in, since it is a part added to the rule rather than one the rule produces.

What holds on the current rule

fact	measured	code
Bell violation	CHSH above 2 on exact algebraic values, never 2 on 17 of 17 starts	E-QTM-0140
Tsirelson's bound	the exact largest CHSH of a knot in closed form, and 2√2 reached by a finite history	E-QTM-0132, E-QTM-0133
contextuality is the fear	a knot is contextual for the model's own readings exactly when it holds a fear, with total deficit $18(e^{\text{mana}} - 1)$	E-QTM-0149, E-QTM-0150
no-cloning, teleportation, GHZ	no beat duplicates an unknown signed state, a love and fear pair met once teleports and swaps exactly on 17 of 17 starts	E-QTM-0151, E-QTM-0152, E-QTM-0153
spin statistics	a full turn is $(-1)^N$ with N the doublet count, so turn sign equals exchange sign	E-SPN-0071
Pauli exclusion	one vibe per slot, in the base and in the dynamics	E-SPN-0045, E-SPN-0047
Coulomb's law	a charge's potential is $1/(24\pi r)$ with no $1/r^3$ term, the quartic isotropic by an integer identity	E-FRC-0241
charge in whole units	counted locally on every beat, pair creation included	E-FRC-0243
two polarizations of light	exactly two per momentum on the husk, the bulk's third killed by the projection	E-FRC-0179
isotropic transport	charge, trace, sound and shear equal in every direction, because the medium is blind to store signs	E-RLT-0094, E-RLT-0096
the second law and the arrow	coarse entropy rises a thousandfold forward and under the exact inverse alike, fine-grained entropy exactly constant, 17 starts	E-FND-0146, E-FND-0147
one causal world	every husk dock descends from the seed by beat 10, 17 of 17 starts	E-RLT-0093
the working vacuum	balanced, bounded, reversible, covariant, 51 of 51 cases	E-RLT-0105
positions in superposition	a lone love's positions interfere, 17 of 17 starts, the coin-off control exactly 0	E-QTM-0157
fluctuation-dissipation	a husk block's energy variance equals −dE/dβ on 17 of 17 starts (mean 1.001), an excess relaxes along the equilibrium autocorrelation, one bath at β ≈ 3.9, the non-Gibbs control off by 159	E-FND-0156

Partial, and the step each needs. The electron binds on one husk line as three loves under the pass (binding 1.079, spin one half share 0.982, travelling at 0.030 a beat, E-SPN-0093), but not yet across lines, and g is unread. Matter reads the light as an exact Peierls phase, likes pushed apart at 0.96 to 1.17 of the static Coulomb prediction, on one line with no back-action (E-FRC-0252). An atom's first lifetime decays at 0.975 to 1.000 of the golden rule, on a stand-in atom (E-FRC-0240).

This week: the vacuum is a bundle of lines

Work of 2026-09-26 and 27, after the working vacuum was chosen. It started from four blocked problems (the electron off its line, a bounded self, gravity's reach, and measurement) that all waited on a vibe leaving its line, and ended with one reading of all four.

• The line law (E-SPN-0098, pass). With no mixer, the tone on every straight mesh line is conserved: 6,144 of 6,144 lines on side 8 over 48 beats. Every piece of the rule acts inside one line, so the knit is 6,144 1d worlds that never touch, in the vacuum's orbit. The mixer G below breaks all 6,144. A mixer of stores breaks 41,699 of 49,152, every break made by the collision's bounce piece once a moved pair creates a dock where it acts across lines.
• Line locality is momentum conservation. A lone vibe's momentum is its root, so it cannot change line without changing momentum. That is Newton's first law, not a defect. The lifted mixer changes a dock's momentum on 407,664 of 415,856 hops (E-SPN-0096), while the collision K (two lone vibes meeting in a dock, its lines turned by the dock's momentum) keeps it by construction and stays quiet.
• Every mixer cascades (E-SPN-0094 to E-SPN-0099). Exactly one covariant, real, reversible way off a line exists in the rule's ring: G, Grover's step on a frame of four orthogonal lines (keep 3/4, hand −1/4 to each other slot). Added to the working vacuum, a lone vibe leaves its line (0.37 off by beat 4) and its wake grows to 39,720 trits a period against 15 without it, and the one-line electron comes apart (return probability 0.454 after one beat, E-SPN-0095). No mixer in the ring fixes the vacuum and still mixes, a theorem about the ring (E-SPN-0096). The lifted mixer cascades faster, 0.618 a beat against 0.155 (E-SPN-0097). Every move rate from 7/64 to 42/64 scrambles the box, growth 1.276 to 1.473 a beat, so the prediction that the ω mixer (21/64) sits at a critical point failed: it reads 1.382, not 1 (E-SPN-0099). The relay is the vacuum: a moved vibe breaks a vacuum pair, which leaves a lone vibe, which the mixer moves again. The rule's own gates on pair making scramble a lone vibe too, where the working vacuum holds its wake at 27 (E-RLT-0106).
• The path-key flaw, and its fix (E-MTH-0029, pass). The registered key read the beat as t times the dock count, which is 0 mod 2^16 on side 16, so on that box every beat made the same choice at each dock, and on sides 8 and 24 the choice repeated every 4 beats. Every exchange-path reading was a special, periodic record. The named full-period key takes 64 values in 64 beats on every line. Rerun on it, the findings hold and are slightly stronger: 16 pairs scramble on side 16 (576,000 trits) where the old key's run was only turning (17,544).
• The fracton rungs (E-SPN-0100, fail on F1 and F5). On the full-period key a lone love stays on its line but runs 8 dock steps, half the box. The immobile rung was the frozen key's. Two lone vibes that meet reach other lines quietly on 434 and 436 of 528 seeds (largest wake 62), and a 12-line cluster stays bounded for 128 beats.
• The reach across lines was the vacuum's (E-SPN-0101, pass). A contact that never turns a vacuum line keeps the vacuum, is quiet at every density (about 18 trits a pair), and reaches 0 lines. So under the pass the 10 lines a pair reaches are vacuum lines it disturbed. Sparse disturbances heal and overlapping ones break the vacuum.
• Density scrambles, and the threshold falls with the box (E-GRV-0085, pass). Two quiet clusters interact with no fall with distance (far over near 0.98 and 0.86, where 1/r would give 0.46), and 16 pairs scramble side 16 while 40 scramble side 24. A threshold that falls with box size has no infinite-volume limit.
• The averaged response has no 1/r (E-GRV-0086, pass). Averaged over 24 distinct full-key records, one sparse pair leaves a core of range half a dock and a tail nearer $1/r^2$ than $1/r$: vibes streaming out along lines, not a static field.
• The entanglement runs along lines (E-GRV-0083, E-GRV-0084, both fail). Held exactly, the vacuum's knot entanglement is an area law only at its first meeting, with η = 8.435, 12.65, 16.87 nats a husk plaquette on sides 8, 12, 16, growing with depth. At equilibrium it is a volume law (1,770 to 5,472 nats for equal-area slabs of width 1 to 4). A lump lowers the entanglement near it with no first law. Jacobson's route to Newton's constant needs the area law, so it fails here.
• Stores cannot join lines, as a class (E-RLT-0107, E-GRV-0087). Storing two lines of a frame as one unit gives a sound rule whose vacuum is the line vacuum relabeled. A store piece that is an involution must return every vibe to the slot it took it from, so no store of any kind moves a vibe onto another line.

One property behind all of it: the vacuum is correlated without limit along each line and not at all across lines. A 3d vacuum has to be correlated the same way in every direction. So no 1/r response, no area law and no quiet reach across lines are one fact. Three ways to fix it are closed: a mixer (it breaks momentum), a cost on releasing a pair through the rule's own vetoes (they scramble), and a new store (an involution returns every vibe to its line). A released pair that costs energy through a new energy ledger is untested.

Gravity

What a gravity must do, each a measured fact about the world:

requirement	why it is required	what it forces on a field
falls as 1/r	Newton's law holds from millimeters to galaxies	the field is massless
only attracts	no two masses repel	the field has even spin, 0 or 2. Odd spin (light) repels like sources
the same on everything	a feather and a hammer fall alike to 1 part in 10^15	the source is blind to charge and kind
changes arrive at c	gravitational waves arrived 1.7 s after the light of GW170817, across 130 million light years	the field is retarded
bends light by 2	1.75 arcseconds at the Sun's limb, twice the falling-particle count	clock and space stretch in equal measure (spin 0 alone gives 0 or 1)
positive energy	the world does not run away	attraction lives in the static part, waves carry positive energy

Closed routes. Everything read off the rule and its light failed: entropic pull, time dilation near a held knot, occupancy-gated takes, a neutral lump in the depth, growth, shared information and shared history (E-GRV-0056 to E-GRV-0073), the light's second-order residue, which is van der Waals at $r^{-6.5}$ (E-FRC-0253), and the vacuum's response and entanglement (this week, above). A second light sourced by content is charge-blind with the right $1/(24\pi r)$ shape, but like sources repel, as spin 1 must (E-GRV-0074). Counting its static energy negative gives Newton to 0.37% but a pull that acts at once (E-GRV-0076 to E-GRV-0078).

The route that works: the husk's depth. A husk column's depth, as a spin-0 wave sourced by content (the radion), attracts as 1/r with no chosen sign, is causal, falls alike and bends light toward mass (E-GRV-0079, E-GRV-0080). Why depth: a deeper column holds more docks, so the same count slows a clock and lengthens a span, which is the arena changing rather than a force on a charge, and that is what makes it the same on everything. What was built since, in order:

step	result	codes
depth as a register	light bends by Newton's 1, not 2, for every coupling, because depth entered only the light's inertia (its clock), not its span	E-GRV-0088, E-GRV-0089
depth found, not stored	one trit a link (step up, level, step down). Steps out minus steps in equals content at every dock, so a lump's lines thin as $1/r^2$ and the summed depth falls as $1/r$ by geometry. No depth register exists. Gauss exact on 43,008 checks, the summed depth equal to the radion to 5.6e-17, k = 0.0418217	E-GRV-0090 pass, E-GRV-0091 fail
depth sets span as well as clock	a link divides by the depth count of the columns it joins ($q = 2D + 1$), so light's index is q, matter's clock slows as $q^{-1/2}$ and space stretches as $q^{1/2}$: factor 1.999 and 1.995 read from the exponents, and 2.258 through a lens against the general-relativity control's 2.338 and the clock-only control's 1.091	E-GRV-0092, E-GRV-0093
α was not constant	read in local units, $\alpha = \sqrt3/(48D)$, coupling to the potential with slope 2.10, where atomic clocks allow about 1e-6. The prediction that local units would cancel it failed, because ħ in the light's units is $D/\pi$	E-FRC-0256 fail
α held flat	fix the gauge field's resolution at one depth $D_0$ everywhere and let only the metric divisors read gravity, as general relativity's minimal coupling does. α flat to 2.8e-8 (slope 7e-9), light still bends by 2.258, and gauge covariance now survives 2,120 window crossings where the unfixed light broke on every beat	E-FRC-0257, E-GRV-0099
universal fall	two span lumps fall alike to 0.9% at one point and 6.1% at the mirror, against a 3% gate. The cause is the mesh: the integer depth staircase reflects slow heavy lumps, and a lump smaller than a dock falls by its own lattice ray, which differs by mass. Real matter is far larger than a dock	E-GRV-0096 fail
a growing bulk	lines and radiation drain into a bulk with 8 times the docks per layer, but the pull becomes a Yukawa (cut off exponentially) of range 1.03 docks, because a growing bulk is a leak to ground	E-GRV-0094, E-GRV-0095 fail
a shrinking bulk	the $\lbrace3,4,3,4\rbrace$ seen from its husk shrinks inward, as in Randall and Sundrum. 1/r survives at 0.549 of the husk alone's strength (the zero mode's share is 16/31 = 0.516), with a positive short-range correction of 0.34, 0.15, 0.083 at r = 4, 8, 12, with no free number but not Randall and Sundrum's shape (log slope 1.16 against 1.87)	E-GRV-0100 partial, E-GRV-0101 fail
the bulk's clock warped	a layer beats once every $2^\ell$ husk beats. Exact, bounded, same static pull (to 5.0e-5), because a static field has no rate. Randall and Sundrum's shape needs the lapse (the clock-rate factor) inside the link weights as well	E-GRV-0102, E-GRV-0103 fail
the lapse in the link weights	layer-$\ell$ links store their step $2^\ell$ times finer, so no new register. Exact (force equal to the linear solve to 7.1e-5). 1/r kept at 0.726 of the husk alone's k against 0.751 predicted: the smooth-layer model runs 3% above the lattice on every stack, more than the correction, so the correction's shape is not readable on this box	E-GRV-0105 fail

The pull travels at light's speed. E-GRV-0101 first read the front at 1.26 c and 1.40 c, which GW170817 would rule out, and E-GRV-0103 found the husk alone reads the same by that method: a near-field bias of the reading. A calibrated witness (E-GRV-0104) sends a rod of 8 content units one dock, so one front leaves behind it, and times the arrival at one third of the static height, the point the husk's dispersion leaves exactly at t = d/c. It reads the husk alone at 1.0023 c, the warped stack at 0.9798 c (never above light), the lapse stack at 1.003 c, and the unwarped stack at 1.126 c, so it does catch a faster bulk where there is one.

What this is, stated plainly. A pull that attracts as 1/r with no chosen sign, travels at the speed of light, is the same on everything larger than a dock, bends light by 2 and keeps α constant, held in bounded registers with the depth found by summing. It is graded L2: the step field is an added part of the rule, and the factor 2 follows from the span change by construction, which the runs show the integer rule realizes.

Settled in the last rounds:

1. a horizon at 2GM. Where the clock register runs out of room, time stops and light is trapped, exactly. Its radius grows as M once the box is taken out, bending stays 1.9985 and α flat (E-GRV-0118, E-GRV-0122). It is bald far out, with one short-range number near it (E-GRV-0116)

2. Hawking's scaling. Light leaving a large horizon is redshifted at 0.97 to 0.98 κ, T·M flat to 1.7%, over a window capped at ½ ln C by the clock register (E-GRV-0123). A forming horizon makes particles, exactly counted, with a power-law spectrum at the register sizes run (E-GRV-0132)

3. an area law. Carrying the infall onto the horizon's own surface registers hides 18.18 nats per cut link at every size and every stage of growth. A horizon forms only while what fell in fits on its surface, which is Bekenstein's bound as dynamics. S/(A/4G) is a flat 1.44 to 1.47 (E-GRV-0131, E-GRV-0135). The torn husk alone gives a volume law (E-GRV-0124)

4. source and field. The confining string is the energy-line register mod 3, with Gauss exact (E-GRV-0127). No local move on a closed register turns those lines into the 1/r depth within a run (E-GRV-0130, E-GRV-0133), so the step field is gravity's own field, as the Coulomb field is light's, sourced by the rule's energy

5. matter that stays put, and falls alike. The rule holds the three-love level with a one-trit drift cost and a consistent fermion sign, and its energy sources the 1/r (E-SPN-0104, E-GRV-0119). Inertia over energy is two factors:

   • the walk goes to 1 as a love gets light against a dock (1.006 at n = 8, E-SPN-0107)
   • the binding goes to 1 as a string-bound love-fear pair's string loosens (1.018 at D 30), with its size growing as the predicted cube root (E-SPN-0112)

   The trio does not, because it is held by contact at a full dock, 2.2 docks across at every tension (E-SPN-0111)

6. spin 2. One depth per line class fixes a full metric, and with linearized general relativity as the coupling it carries exactly two polarizations at c (E-GRV-0125)

7. the 3.8% drift was the wrong energy being read. The headroom light keeps the right one to 2.9e-14 and bends by 2.0069 (E-GRV-0128)

8. a composite at speed. Along the joint limit a love-fear meson's inertia over energy falls to 1.003 at rest (E-SPN-0113). The levels crossing its band are the Klein channel, pair tunnelling through the string, sorted by an exact love-fear exchange symmetry. Read through them, the meson stays one particle to momentum equal to its rest energy, with c within 0.14% of the walk's at n = 16, as a resonance of width 1e-7 of its mass (E-SPN-0115)

9. spin 2's structure from the rule's own redundancy. Sliding the unlabeled docks forces linearized Einstein–Hilbert (E-GRV-0138), and a spacetime slide forces the lapse and shift to be Lagrange multipliers with DeWitt's kinetic term (E-GRV-0139). At cubic order the full slide fixes the wave speed at exactly c and λ at 1, which a foliation-preserving slide leaves free (E-GRV-0141)

Still open:

1. bound matter moving in 3d. On the rule as it stands every vibe is a lineon, and no composite can leave its lines (E-SPN-0116 to E-SPN-0120). Line-only motion is excluded by observation, by about 17 orders of magnitude (E-SPN-0126). So the rule has to change. A candidate change is tested, and it splits the problem in two.

   Motion works under the candidate. It has four parts:

   • a filled love sea as the vacuum;
   • a fermionic mixer over all 24 slots of a dock, which Pauli blocking keeps inert on the sea (E-SPN-0140);
   • the swap coin, which keeps a vibe or crosses it with integer entries;
   • one new prime, 7, so the mass can be tuned (E-SPN-0143).

   Under it, a hole moves the same in every direction and obeys special relativity, with R = tan(m)/m. Every excitation shares one top speed, c* = c/2, and nothing exceeds it at any momentum. This holds in the exact many-body rule, with the vacuum inert and gravity's sign kept (E-SPN-0145, E-GRV-0144). The graviton's slide holds at c* too, so matter and gravity can share one c.

   Binding, first attempts. Every one below leaks or is too heavy, and one obstruction runs through all of them: the swap coin's 22 flat bands, which sit in the continuum for a light member.

   • A phase string leaks through the swap coin's flat bands (E-SPN-0146).
   • A mass string binds exactly and moves isotropically, but it needs heavy members, reads the pair's separation nonlocally, and gives R = 1.57 (E-SPN-0147). No dock-local piece can fix it: that is a band-sum theorem (E-SPN-0148).
   • On a line R → 1, with the lattice mass the only excess (E-SPN-0149). In 3d a cost-only string misses the string's own transverse inertia.
   • A Z3 flux register on the links records each member's path, so a member walks a tree and cannot be light (E-SPN-0150). A Z3 plaquette term cannot give light members and a confining string at once (E-SPN-0151).
   • The group the roots already form, 2T, freezes while the string is still heavy. A register in the 4d bulk cannot confine continuously for any group, so the string has to live on the husk (E-SPN-0152). On the husk, 2I, with the golden ratio in the ring, reaches SU(2)'s scaling region before it freezes: its string tension there is 1.6e-18 of the strong-coupling value (E-SPN-0153).
   • But no gauge string reaches the flat bands. The swap coin undoes itself in every static field of every group, so exactly 11 flat states per color per dock survive any string, a tense string makes the member about 6 times heavier (Kesten's tree floor), and a confining ladder meets the pair at separation 2m/σ = 11.19 (E-SPN-0161).
   • A two-beat schedule frees every flat band's rest phase and lifts 7 of them clear, but three transverse partners of the singlet stay pinned in every covariant schedule of any period, because the singlet's partner is a vector: 314 of 314 schedules that keep R are open (E-SPN-0159).
   • A Coulomb pull binds heavy members only, with R 1.41 to 1.44, and leaks a light one (E-SPN-0155).

   The register: a member that is a spinor. Everything since runs on the candidate rule with one more piece, which is not adopted.

   • An 8-value register per member. A 2-value spinor cannot be exact in the model's ring (the quaternion algebra (−3, −1) is ramified at 3, Q8 needs (−1, −1), ramified at 2). The least exact register is the even Clifford algebra Cl⁺(4) = H ⊕ H, 8 values, on which W(F4) acts by minors. The singlet's partner is then a Clifford partner with no transverse state, and the census closes for a light member for the first time: binding room B* = 2M exactly, the ideal Dirac value, for every mass below π/6 (E-SPN-0160).
   • The price, a theorem: such a partner takes a quarter of the stream, so the one speed is c/4 in bulk units. Read on the husk it is √2/4 husk docks a beat, isotropic, and c/4 is only its bulk label (E-SPN-0167). The graviton shares it: every one of a member's 192 bands is the Dirac band or flat, and no register that closes the census keeps c/2 (E-GRV-0145). Gravity's count field on the many-body register rule is one hole's sector count, counted from the sea. The flats neither source nor feel it, two holes pull together with the sign fixed by the clock, and the same piece pushes members apart, so it is universal only because every excitation of the full sea is a hole (E-GRV-0146). Against the rule's own inertia it fails: the charge is one hole a cycle whatever the hole's speed or mass, the response is the scalar density, and the piece acts as two fields, so a hole falls toward its own band and at most 0.082 as much toward the other (E-GRV-0147). Read at separations 3 to 8 the pull follows the kernel's gradient, power −2.02 against the kernel's −2.09 (E-GRV-0148).
   • The many-body rule with registers is exact. Every register piece is one-body, so its fermionic lift is fixed, the full sea is one branch with a unit amplitude every beat, K never fires, and one hole is exactly the member (E-SPN-0163).
   • A light pair binds and holds exactly. A string built only from the register's sector projectors keeps the flat bands decoupled at every separation. The pair holds 128 beats with a leak below 1e-9, isotropic to 5e-7, against the full 192×192 rule to 5e-17. But its inertia over energy is R = 34.65 (E-SPN-0162), and at weaker strings 13.47 and 1.43: R falls toward the continuum, and a static binder crosses R = 1 only at one tuned coupling (E-SPN-0174).
   • The model's light supplies the missing inertia. Coupled to the member's stream as a Peierls phase with back-action (Gauss holds with the member as charge), the light's Coulomb and transverse kernels are one Maxwell Lagrangian to 6e-8, so the transverse exchange is exactly the Darwin term. It removes 0.879 of the static excess on a bound state, and for a light member takes R to 1.012 against its own tan m/m 1.012. This is linear response, not the pair and the quantum light as one dynamics (E-SPN-0169).
   • Held by the light's own pull, a light register pair binds like hydrogen (E_b within 6% and 14% of the continuum), holds with no leak and is isotropic to 2e-6, but R is 8.5 to 16 at the couplings the engine reaches, heavier the tighter the binding. The argued cause is strong coupling across one link: each center-of-mass step crosses the pull's gradient over a link (0.59 rad a cycle against a band of 0.27) (E-SPN-0173). A symmetry-reduced engine reaching Bohr radii of 8 to 20 is running, to read R where the pull is weak across a link. It found that the level is a multiplet of at least 6 lines, each one heavy on its own, so the high R is not a blend
   • The wall as the husk. A Wilson mass built from the rule's own mixers makes the husk a domain wall, second Chern number −1 per copy (E-SPN-0166, a fail on one pre-registered isotropy tolerance). The member on the wall is light by depth alone, its mass falling 940-fold from depth 4 to 12 with the bulk unchanged, relativistic (R → 1.0004) and isotropic. But the pair census on the wall is open, through depth channels and a Floquet resonance with the far side (E-SPN-0168). The resonance closes by moving the far side's phase, which displaces the wall's Weyl node (E-SPN-0172, a fail on one rigidity tolerance)

2. a temperature: a forming horizon's spectrum stays a power law at every register size run (C 13 to 49), so what shows is the growth's quench (E-GRV-0134). Read late, a settled horizon makes nothing (1e-7 of the burst): it is a static wall with its redshift capped at ½ ln C, so a bounded clock predicts a formation burst and no steady evaporation unless ln C ≫ 4π (E-GRV-0136)

3. Randall and Sundrum's short-range number: finer layering takes the scalar to 3/4 of theirs (E-GRV-0137), and the spin-2 tower reaches 0.971 of it (E-GRV-0142), with the 4/3 derived from a slide along the bulk's depth (E-GRV-0143). Left: the warp, the radion's stabilization and brane bending

The full record is note/project/vibe/roadmap/research/remaining-pieces.md and discrete-gravity.md in the ClueSurf notes.

Where the previous readme read otherwise. It gave gravity as solved: the area-law potential giving Newton and the factor two (E-GRV-0012), and an area-law invariant (E-GRV-0002). Both ran on a hand-written walk and a hand-built metric, graded stand-in. On the working vacuum the entanglement is a volume law at equilibrium (E-GRV-0083). It also read gravity as the clock rate alone, which bends light by 1, not 2 (E-GRV-0088).

Other open problems

problem	where it stands	codes
one light speed	on the working rule a locked vibe's top speed is 1/2 a beat and every stable husk light is capped at $c \le 1/\sqrt6 = 0.408$, a theorem on each side. With the register, matter and the graviton share √2/4 husk docks a beat, which sits inside the stable cap. The quantum loop light meets it exactly at κ = 3/16 and the split ρ = 3 (the trit column never can: even against odd). But the split is fixed by equal filling, and at a balanced split exact equality holds only as a limit of the register's size, missing by 1/(2q²) at register 16q, which needs a register near 10¹⁰ to match the 1e-18 bound. The balance is exactly 3/8 counted on the bulk's registers and 0.4614 on the husk's weighted columns	E-FRC-0250, E-GRV-0145, E-FRC-0260, E-FRC-0261, E-FRC-0262
composite light	a bound love and fear pair is massive by a theorem, with one polarization on the husk, not two. Light stays its own field	E-FRC-0246, E-FRC-0247
the electron in 3d, and g	bound on one husk line only. A general direction is a superposition over lines. g needs its coupling to the light, not a mixer	E-SPN-0093, E-SPN-0089
measurement	a frame-covariant rule depolarizes every basis at one rate, so a pointer basis must come from the environment, never the rule (a 2-design theorem)	E-QTM-0154
the weak force	the locked rule keeps P, C, T, CP and CPT exactly. The register supplies what was missing, a spin apart from the slot: its two halves (J = right multiplication by vol) are kept by the 576 rotations and swapped by the 576 reflections, and a chiral mass breaks P and CP by 0.180 with CPT exact and no isotropy lost (Molien). The frames cannot be the three generations (anisotropic at fourth order). On the flat husk quotient parity survives exactly, so P violation needs an oriented depth, and the Wilson wall's chiral face breaks it by a whole chirality, +2 and −2 at the two faces, the rule's depth-keeping mirror reading −2 and +2, with rotations, spin one half (the untwisted rotation, exact) and CPT kept. The rule keeps an SU(2) on each register half exactly, and the full sea keeps it too, so no vacuum yet breaks it. All on the candidate rule	E-FRC-0248, E-FRC-0258, E-SPN-0167, E-SPN-0168, E-FRC-0267, E-FRC-0268
CP and the asymmetry	three flavors keep every invariant, and the trimaximal mixing, exact in the ring with Jarlskog √3/18, is removable at one body (no one-body piece joins the halves, a theorem). A two-body register exchange makes it physical, and C and CP then break together inside one sea with no piece that tells love from fear, so the love-fear mirror stays exact. Member number is kept by every piece and by any gauge winding (Nielsen–Ninomiya in Floquet form), but flows through the Wilson wall under an exact E·B field, +4 per loop at one face and −4 at the other. So all three Sakharov conditions are in the rule. The flow is an index, blind to CP, so a CP-symmetric field history nets 0: a net asymmetry needs a dynamical light biased by the CP-odd currents, which is not built	E-FRC-0259, E-SPN-0163, E-SPN-0164, E-SPN-0165, E-SPN-0168, E-SPN-0171
the value of α	$\alpha = \sqrt{3/(2\rho)}/(12N)$, set by the split and the depth, not by the speed. A frozen prediction over depths 1 to 10,000 at ρ = 3 and 3/8, and at the husk's measured 0.4614, finds no hit within 1e-3 of 1/137.036 (nearest 138 at depth 11, 0.70% off, chance about 0.32 under the null), so nothing is claimed. The husk's split is narrowed to two readings of one set of fills, 0.4614 (average) or 0.4146 (worst register), and the ladder's Planck-optimal split cannot pick between them at N ≤ 13 (a fail). Its departure is a function of the ratio alone, but its spread stays above what separation needs at every N a dense spectrum reaches, and no sparse method computes it (a fail, the clean negative)	E-FRC-0242, E-FRC-0261, E-FRC-0262, E-FRC-0269, E-FRC-0270, E-MTH-0010
a bounded self	shared history gives one component of all 24,576 vibes or closed components of exactly 8 on one line, nothing in between	E-SLF-0177, E-SLF-0178
one rule for every row	the working vacuum and the electron share one rule. Most older rows ran on earlier rules, and a triage found the rest cannot be rerun usefully: most wait on motion across lines, the light's own quantum sector, CP violation or an energy ledger. The register results run on the candidate rule, which is not adopted	the ledger's section Q

Rows started this round

Ledger rows nobody had started, each run once with a control and graded by the ledger's own rules. Every one is partial unless marked, and the reason is stated.

row	measured	why not held	code
the path integral	the working rule is its sum over take histories term for term, 0 mismatches on 17 of 17 starts, and stationary phase takes over in the long run, E[v²] → 1 − √3/2 = 0.1339746	1d along a line, and the meeting-free control was amended after the first run (disclosed)	E-QTM-0158
delayed choice and the eraser	a later meeting erases a which-arm mark, fringe and anti-fringe summing to the washed-out pattern, with the earlier probabilities exactly unchanged	the gate schedule is arranged, not produced by a mesh configuration	E-QTM-0159
weak values	the rule's own histories carry shares −1/2 and 3/2, and a weak pointer reads 9/4 as its disturbance falls as 9/4·4⁻ⁿ	schedule arranged	E-QTM-0160
mixed states and POVMs	a 4-outcome reading whose effects sum to I, are not projectors and do not commute, Born exact, a pure whole leaving its part mixed (purity 23/32)	schedule arranged, not informationally complete	E-QTM-0161
Wigner's friend	the friend's record undoes with probability 1 and restores P(R) = 1/4, where the collapse account predicts 17/32. Once copied, it cannot be undone	the friend is one vibe's point, not a self	E-QTM-0162
length contraction and simultaneity	one Lorentz factor for speed and energy, a moving packet the contracted rest packet, the tilt of "now", within 3e-4 at 0.5 c* and scaling as m²	candidate rule, lone excitations, thresholds set after a probe	E-RLT-0108
relativistic collisions	four-momentum conserved exactly, and the outgoing shell is the boosted sphere to 7.3e-4 at boost 0.83	candidate rule, kinematics only	E-RLT-0109
Landauer's principle	erasing a trit puts exactly ln 3 into the box, 0.462 nats of it invisible to the husk coarse map	the energy side, kT ln 3, not measured	E-CMP-0018
error correction	a repetition code over mesh lines, since no error crosses lines, 2,754 of 2,754 correctable cases decoded, distance exact	classical, and bought by the line law the candidate rule removes	E-CMP-0019
the horizon and flatness problems	the husk is a horosphere, flat by geometry, and every pair of docks one seed reaches shares a causal past (0 of 8,256 disjoint), against 171 of 300 for a start everywhere at once	temperature agreement not read	E-CSM-0058
inflation or its replacement	the husk does not inflate: a cubic ball, 3.22 e-folds, the bulk's ×18.28 being volume off the husk	the ripple spectrum not addressed	E-CSM-0059
the end state	a fixed box reaches coarse maximum entropy in 4 beats and sits at the predicted fluctuation floor, with exact recurrence	a growing mesh argued, not run	E-FND-0155
fluctuation-dissipation	held: variance equals −dE/dβ on 17 of 17 starts, relaxation along the equilibrium autocorrelation, one bath		E-FND-0156
Compton and Thomson scattering	the shift is λ_C(1 − cos θ) to 5e-6, λ_C set by the member's inertia, the forward amplitude Thomson's with σ/σ_T = 1/R² → 1, the dipole response Dirac's	Klein–Nishina and the absolute cross section (α) not run	E-FRC-0263
the photoelectric threshold	ionization above E_b at first order, following the intensity, and 4,300 times smaller below every bound line	a fail: the gate allowed no loss below E_b and missed hydrogen's own lines, and the pair is a stand-in	E-FRC-0264

Also rerun on the working rule this round: positions in superposition, to held (E-QTM-0157), and the Coulomb binding gate registered with its first run (E-SPN-0155, partial on one pre-registered instrument clause).

Order of work

One chain gates nearly everything, because most rows still at none or partial wait on bound matter that moves across lines.

1. R for a light pair held by the light. The symmetry-reduced engine reads R at Bohr radii 8 to 20, static and with the Darwin term. If it falls to tan m/m there, bound matter moving in 3d has its full demonstration on the candidate rule with the register.
2. The pair and the quantum light as one exact dynamics, to confirm the Darwin term beyond linear response.
3. Adopt or refuse the candidate rule. The register results then become the working rule's, or are set aside.
4. Rerun the rows the line law blocks on that rule: bosons from even-N composites, temperature across lines, anyons, the atom, and the self rows.
5. A net asymmetry: a dynamical light biased by the CP-odd flavor currents at the wall.
6. The light's own quantum sector on the whole husk, and an energy ledger for the Casimir row.
7. The paper, text/0017-the-vibe-theory-model, which waits on step 1.

Results on earlier rules

The previous readme led with results from the turning weave and the rules before it. They stand as results about those rules, and each reads differently now:

• CP violation with CPT exact was measured on the turning weave. The locked rule keeps C, P and CP exactly (E-FRC-0248). It comes back on the candidate rule with the register, C and CP together (E-SPN-0164), but not on the working rule.
• The Cabibbo angle within four percent was never tested against a pre-registered null. The ledger grades it stand-in, and the geometric mixing angles themselves come out degenerate (E-FRC-0070).
• No magnetic monopoles was an identity of the earlier light's link potentials. The light adopted since is compact U(1) (E-FRC-0244), which admits heavy monopoles, so exact absence is no longer a prediction.
• The vacuum clock of period three belonged to the charge rule. The working vacuum's cycle is 12 beats.
• A factorization of the lepton mass hierarchy fit, carried a pre-committed number that would kill it, and was killed by sharper data. It stays recorded.

The scoreboard

Every commonly observed feature of nature the model must account for, one row per feature, generated from the working ledger's observation table (see Development for the command, never edit it by hand). The experiment codes resolve in test/catalog.csv. Its marks are coarser than the 191-fact ledger above: many ✅ rows rest on stand-ins the ledger does not count as held, and the gravity rows rest on the depth register, an added part of the rule.

mark	meaning	count
🟢	emergent from the rule	4
✅	reproduced on the model	50
🔵	structure derived, dynamics open	20
📌	free input, not predicted	1
❌	open	0
⛔	blocked on a base decision	0

observation		note
quantum
quantum amplitude at the base	🔵	The bare knit has no amplitudes: a defect never spreads and two add exactly as sets (E-FND-0080). The quantum part is the signed weight on the role grid, fear: the singlet is 72 loves and 18 fears on 54 units (E-FRC-0120), and the cube-root swap phase is an exact reversible beat whose gates close to su(9) (E-FRC-0127). It runs in the adopted knit as the fear beat, the one gate outside the Clifford group (E-FRC-0159, E-SPN-0063, E-QTM-0118), and gives amplitudes to roles, not positions (E-SPN-0081). The turning weave's vacuum clock carries a discrete phase kinematics, listed below.
		E-FND-0080, E-FRC-0120, E-FRC-0122, E-FRC-0127, E-FND-0107, E-FND-0106, E-FND-0105, E-FND-0102, E-FND-0103, E-FND-0096, E-FND-0091, E-FND-0086, E-FND-0123, E-FND-0124, E-FND-0125, E-FND-0126, E-FND-0127
Born rule	✅	Counting weights at the domain wall give the Born statistics.
		E-QTM-0012, E-QTM-0091, E-FND-0091
entanglement and Bell	✅	Bell violation is reproduced on the model walk.
		E-QTM-0011, E-QTM-0038, E-QTM-0057, E-QTM-0100, E-QTM-0111, E-QTM-0140, E-QTM-0141, E-QTM-0143, E-RLT-0055
measurement and definite outcome	🟢	The domain wall projects and amplifies one outcome, from the rule.
		E-QTM-0085, E-QTM-0090, E-FND-0091, E-FND-0123, E-FND-0124, E-FND-0125
QFT vacuum and reflection positivity	✅	The free clock field and the thermal gas are positive semidefinite on a non-trivial estimator, and the amplifying wake violates positivity exactly as it must.
		E-QTM-0096, E-QTM-0025, E-QTM-0017
path integral	✅	The sum over walk paths reproduces the propagator.
		E-QTM-0018
quantum error correction	✅	The clock code protects the phase and corrects single errors.
		E-QTM-0093, E-QTM-0095, E-QTM-0055
double slit interference	✅	Two-path interference with the cosine cross term holds on the model, in the walk and in the domain clock.
		E-QTM-0018, E-FND-0085, E-FLD-0014
quantum tunneling	✅	The exponential tunneling law and Klein tunneling hold on the walk.
		E-QTM-0056, E-QTM-0073
uncertainty principle	🔵	The position-momentum bound is confirmed as the Fourier identity on the lattice.
		E-QTM-0059
casimir effect	✅	Two boundaries in the model vacuum attract by mode exclusion.
		E-SLF-0018, E-SLF-0017
gauge
gauge group SU(3) SU(2) U(1)	🔵	The group and its embedding fall out of the octonions on the 24-cell, exactly.
		E-FND-0021, E-FRC-0022, E-FRC-0025
Weinberg angle 3 8	🔵	The 3/8 tree-level angle is derived, running it down to the measured value is standard.
		E-FRC-0055, E-FRC-0054, E-FRC-0044
hypercharge and fractional charges	🔵	The quark and lepton charges come out of one representation, exactly.
		E-FRC-0002
anomaly cancellation	🔵	The anomaly sums cancel across one generation, exactly.
		E-FRC-0002
dynamical gauge field	🔵	A carrier exists with a measured unit coupling, the emission vertex is measured (the interacting band radiates matter-to-wire conversion, E-FND-0113), and under the turning weave the interaction structure connects all twelve species (E-FND-0117). The remaining work (the gauge algebra of the carrier, the fast photon sector) is open research. The kick generator is charge-signed (E-FND-0128), the abelian coupling measured. In the adopted three-trit model, links hold grid moves and a flow with Gauss's law exact (E-FRC-0123), and they move by reflection with energy exact and every frame change a symmetry (E-FRC-0128), on variant rules. The adopted knit carries the links as Sigma(648) elements, with every dock's held color exact while it makes pairs (E-RLT-0067).
		E-FND-0100, E-FND-0104, E-FND-0108, E-FND-0113, E-FND-0117, E-FND-0128, E-FRC-0016, E-QTM-0093, E-FRC-0073, E-FRC-0123, E-FRC-0128, E-FRC-0134, E-FRC-0135, E-FRC-0144, E-FRC-0145
photon and Maxwell	✅	The Ward identity and the Maxwell spectrum hold on the substrate gauge sector.
		E-FRC-0072, E-RLT-0030
nonabelian gauge and confinement	✅	Wilson loops show the area law and confinement on the mesh, and the classical color group Sigma(648) confines under a rule with no random number.
		E-FRC-0039, E-FRC-0007, E-FRC-0045, E-FRC-0048, E-FRC-0103, E-FRC-0110, E-FRC-0126, E-FRC-0129, E-FRC-0131
running couplings and unification	✅	The one-loop running and the near-crossing of the couplings are reproduced.
		E-FRC-0044, E-FRC-0009
fine structure constant value	🔵	The base has provably no coupling knob (the bare vertex is unity, measured), so alpha is a coarse quantity: how rarely carrier and matter meet.
		E-FND-0108, E-FRC-0044, E-FRC-0019
QED precision g factor	✅	The g-factor structure holds, the 0.00116 radiative shift is unaccounted.
		E-FRC-0021
strong CP problem	✅	Theta has no continuum to live in (the clock vacuum is three exact points) and the measured T-symmetry of the charge knit picks zero.
		E-FRC-0079, E-FND-0097
electromagnetism
Maxwell equations and light	✅	The Ward identity, the photon mode and the emergent U(1) hold on the substrate.
		E-FRC-0072, E-FRC-0041, E-FRC-0016
Coulomb law	✅	Two charges bind with the inverse-square structure.
		E-FRC-0049
magnetostatics and g factor	✅	Magnetism and the Landau g-factor hold on the substrate.
		E-FRC-0040, E-FRC-0021
Aharonov Bohm flux quantization	✅	The flux period quantizes on the walk around a plaquette.
		E-QTM-0062
Faraday induction	✅	The loop EMF equals minus the flux rate to fourteen decimals at every step, an exactness of the potential formulation.
		E-FRC-0077
ohmic conduction	✅	Impurity scattering gives Drude relaxation, resistivity linear in impurity density.
		E-FLD-0017
refraction and dielectrics	✅	Snell and total internal reflection hold on the clock-rate step to a hundredth of a degree.
		E-FRC-0074
magnetic monopole absence	✅	Every link potential has exactly zero flux out of every cube, so monopoles are forbidden identically, not merely rare.
		E-FRC-0076
matter
spinors and the double cover	🔵	The binary tetrahedral double cover lives in the 24 directions, exactly.
		E-SPN-0029, E-SPN-0042, E-SPN-0031, E-SPN-0044
Dirac equation	✅	The Dirac walk gives the right dispersion and Zitterbewegung on the mesh.
		E-SPN-0009, E-SPN-0030, E-QTM-0094
chiral fermions no doubling	✅	The fermion doubling obstruction is dodged on the substrate walk.
		E-SPN-0043, E-SPN-0004
Pauli exclusion and spin statistics	🔵	Exclusion follows from the spinor sign structure as algebra, and the substrate adds a structural half: a slot holds one vibe of the ternary alphabet, so two same-mode excitations cannot coexist (E-FND-0115). The spin-statistics THEOREM remains Hilbert-space mathematics.
		E-SPN-0014, E-QTM-0064, E-FND-0115, E-SPN-0045
one generation representation content	🔵	One generation of states matches the derived representation, exactly.
		E-FND-0020, E-FRC-0002
three generations count	🔵	Triality gives three, the count is algebra and not yet dynamics.
		E-FRC-0017, E-SPN-0016, E-SPN-0015, E-FRC-0140, E-FRC-0141
distinct generation masses	🔵	The hierarchy scale and ratios are derived and the growth-written condensate is an asymmetric vacuum the model already produces, so the open dynamics is the triality-to-shell tie.
		E-SPN-0039, E-FND-0098, E-FRC-0033, E-FRC-0141
mass hierarchy mechanism	🟢	Warped localization on the mesh produces exponential mass ratios, with controls.
		E-FRC-0031, E-FRC-0032, E-FRC-0033, E-FRC-0053
Koide relation	🔵	The Koide two-thirds appears from the derived angle structure.
		E-FRC-0057, E-FRC-0059, E-FRC-0060
CKM and PMNS mixing	✅	The Cabibbo angle comes out of the shell growth rate within four percent with zero free parameters.
		E-FRC-0075, E-FRC-0070, E-FRC-0020, E-FRC-0036
neutrino masses and seesaw	✅	The seesaw structure is reproduced on the model.
		E-FRC-0037, E-FRC-0035
hydrogen spectrum and atomic structure	✅	The hydrogen level structure comes out on the model Coulomb problem.
		E-SPN-0018
periodic table shell structure	🔵	Exclusion plus the level structure give the shell filling, chemistry itself is unposed.
		E-SPN-0014, E-SPN-0018
nuclear binding and stability	✅	The binding-energy curve and the valley of stability follow the Bethe-Weizsacker form.
		E-SPN-0024
neutrino oscillations	✅	Flavor oscillation with the TM2 pattern is reproduced on the model.
		E-FRC-0036
CP violation	✅	The palindrome traveler candidate has exactly nature's pattern: C and CP violated over the whole swept group with CPT exact.
		E-FND-0102, E-FND-0101, E-FND-0097, E-FRC-0066, E-FRC-0142
superfluidity	✅	The gas shows the superfluid signatures, no wavelength-independent damping among them.
		E-FLD-0007, E-FLD-0015
quantum hall quantization	✅	The plateau staircase is measured: the invariant pins at integers against a continuous knob, steps at the gap closing, and the critical point reads the midpoint.
		E-QTM-0097, E-QTM-0077, E-QTM-0080, E-QTM-0079, E-QTM-0078
superconductivity	✅	Flux expulsion with the exact lattice London depth follows on the model's operator once carriers are dissipationless, which the gas measures.
		E-FRC-0078, E-FLD-0015, E-FLD-0017
higgs
Higgs doublet and custodial symmetry	🔵	The doublet and the custodial symmetry exist as algebra.
		E-FND-0027, E-FRC-0143
electroweak boson masses	✅	The W to Z mass ratio is reproduced from the derived angle.
		E-FRC-0012
electroweak symmetry breaking dynamics	✅	The Z_3 clock is a condensate whose phase is picked by growth history, with an exact commensurability law for the wall network.
		E-FND-0098, E-FND-0086, E-FRC-0143
spacetime
momentum conservation	✅	The adopted knit conserves momentum exactly: the isometric collision keeps charge, count and momentum on every dock state tested (E-RLT-0061), and the lone bounce collision keeps momentum with a lone vibe's wake confined to its own line (E-RLT-0084). The turning weave before it conserves love minus fear and not momentum: the charge-signed momentum along x drifts by 830 over 48 beats (E-FLD-0020), the hop 96 percent of the loss (E-FLD-0021).
		E-RLT-0061, E-RLT-0084, E-FLD-0020, E-FLD-0021
Lorentz invariance	🟢	A propagating mode with the Lorentz cone emerges, with computed control.
		E-RLT-0018, E-RLT-0030
CPT	✅	CPT is exact for the adopted weave (verified at zero on generic states) and for the clock sector it contains.
		E-FND-0102, E-FND-0101, E-FND-0011, E-SPN-0037, E-FND-0097
dimensions 3 plus 1	🔵	The 3 plus 1 split is selected by the algebra, not yet by dynamics.
		E-FND-0038
gravity and Einstein equations	✅	The husk's depth, found from one trit a link and never stored, pulls as 1/r, is causal and bends light by two; the earlier area-law route fails on the working vacuum.
		E-GRV-0090, E-GRV-0092, E-GRV-0099, E-GRV-0100, E-GRV-0083, E-GRV-0012
gravitation
black hole thermodynamics	✅	Bekenstein-Hawking entropy, an analog Hawking flux and the shadow, on models.
		E-GRV-0014, E-GRV-0026, E-GRV-0004, E-GRV-0001
classic GR tests	✅	Light bending by two on the husk light itself once depth sets both its clock and its span, time dilation and the Schwarzschild form on the emergent metric.
		E-GRV-0093, E-GRV-0099, E-GRV-0088, E-GRV-0012, E-GRV-0037, E-GRV-0033
gravitational waves	✅	Propagating quadrupole waves at the right speed on the model.
		E-GRV-0017, E-GRV-0043, E-GRV-0044, E-GRV-0045
frame dragging	✅	A rotating well drags prograde and retrograde rays apart by Fresnel drag, linear and odd in the spin.
		E-GRV-0055
equivalence principle	✅	Bodies of every mass and content fall alike down the field gradient, above the size of one dock; below it a body falls by its own lattice ray.
		E-SLF-0046, E-GRV-0090, E-GRV-0096
dark sector
dark matter phenomena	🔵	The candidate exists and is measured: a persistent thin wall the matter particle crosses without scattering, with the bullet separation mechanism banked.
		E-CSM-0057, E-CSM-0053
Hubble tension	🔵	Expansion is measured linear in age, so H is a local clock of domain age and two honest ladders disagree when they sample regions of different age.
		E-CSM-0027, E-CSM-0020
cosmology
dark energy	✅	Uniform growth reads as a small positive lambda.
		E-CSM-0006, E-CSM-0010
proton stability	✅	The proton is protected by the conserved charges.
		E-FRC-0043
CMB acoustic peaks	✅	The spectrum itself alternates ten to one at decoupling and the null moves into a peak on the coherent schedule, which incoherent noise cannot do.
		E-CSM-0056, E-CSM-0052, E-CSM-0051, E-CSM-0044
BBN abundances	🔵	The freeze-out race is measured and the matter sector is now settled by the adoption, so the nuclear network is open research rather than a blocked decision.
		E-CSM-0055
baryon asymmetry	✅	All three Sakharov conditions are measured on the committed rule: a growth quench charges the matter sector from the exactly C-symmetric empty state, quantized in whole hypersheets, with the total vibe sum exactly conserved (the B minus L analog) and the commensurate quench an exact null. The observed magnitude (ten to the minus ten) is not derived and needs the coarse dilution story.
		E-FND-0116, E-FND-0119, E-FND-0115, E-CSM-0004, E-FND-0101, E-FND-0097
inflation and flatness	✅	Growth gives flatness and a horizon without a tuned inflaton.
		E-CSM-0028, E-CSM-0044, E-CSM-0046, E-CSM-0049
Hubble expansion and redshift	🟢	Growth gives expansion and redshift from the rule with computed control.
		E-CSM-0027, E-CSM-0019, E-CSM-0020
large scale structure formation	✅	Gravitational instability runs on the model's own loop: bodies source the well, fall down it, and a seed clump collapses by a factor near five.
		E-CSM-0054, E-CSM-0002
thermodynamics
second law thermalization	✅	Coarse entropy climbs to near maximum while the microstate stays exactly reversible.
		E-FLD-0016
blackbody planck spectrum	✅	Planck comes out of exact counting on the measured harmonic dispersion, and the classical branch alone gives the catastrophe.
		E-QTM-0098, E-FLD-0018
inputs
absolute masses and couplings	📌	The absolute scales are inputs, as in every framework.
		E-FRC-0067, E-FRC-0009

The reading, and what is falsifiable

Vibe Theory reads reality as one thing, a growing mesh of experience. The image above is its simplest drawable face, the hyperbolic $\lbrace7,3\rbrace$ tiling, meant literally: each tile is a vibe, a unit of experience, and its tone is a felt charge (red is pain, green is peace, blue is pleasure). Touching tiles note (experience) one another. A patch of tiles is a thing or a mind, the whole mesh is the universe, and its growing edge is the present. The $\lbrace7,3\rbrace$ stands in for the $\lbrace3,4,3,4\rbrace$, whose flat 3d cusp is the space we live in. Regular hyperbolic honeycombs run out by 5d, and $\lbrace3,4,3,4\rbrace$ is the one whose docks are a crystallographic root system that carries spinors and whose cusp is flat 3d.

That the base IS experience is the one axiom, and it is unfalsifiable. No reading can tell a felt universe from a structurally identical one. It is held as a frame, never offered as a result.

Everything under the frame is a discrete dynamical system, and that part can be refuted. Each deep claim carries a control where the answer should be no, and the failures are kept. From this week alone:

• the light-bending factor could have read 2 on the depth register and read 1 (E-GRV-0088)
• α was predicted flat in local units and read a slope of 2.10 (E-FRC-0256)
• the ω mixer was predicted critical and read 1.382, not 1 (E-SPN-0099)
• the knots were expected to give an area law and gave a volume law (E-GRV-0083)

So "vibe theory is unfalsifiable" is half right: the axiom is, the physics under it is not, and the experiments test structure, never feeling.

The invariants

What the current rule holds fixed, each measured with a control that could have failed. The structures that recur across other physics theories built from different starting points are worked out in note/triangulating-invariants.md, with the per-theory maps in note/link/.

invariant	what stays fixed	experiment
charge	counted in whole vibes, locally, on every beat, pair creation included	E-FRC-0243
reversibility	run forward then backward, the start returns bit for bit	E-FND-0049
the symmetry	W(F4), all 1,152 elements, with CPT	E-RLT-0061
the discrete symmetries	P, C, T, CP and CPT exact on the locked rule, under all 48 husk parities	E-FRC-0248
momentum	exact on the knit without the coin. The coin gives up exactly this, to make a mass	E-RLT-0084, E-SPN-0090
the line law	the tone on every straight mesh line, 6,144 of 6,144, while no mixer acts	E-SPN-0098
the vacuum's cycle	a 12-beat cycle with pairs made equal to pairs unmade, 466,944 each	E-RLT-0105
the arrow of time	the same cone both ways, so the arrow is the low-entropy slice plus the coarse map	E-FND-0147
the distinguishability metric	Fisher-Rao is the one distance no relabeling of the 24 directions can change	E-FND-0057
the light cone	a fixed top speed, the same in every direction	E-RLT-0014
spacetime dimension	the cusp reads a spectral dimension near 3, so space is 3d plus time	E-GMT-0025
the growth ratio	each shell of the mesh larger than the last by a factor settling at 18.278	E-GMT-0027
gravity's source	on the depth-step register (an added part), steps out minus steps in equals content at every dock, 0 off on 43,008 checks. There is no area law on the working vacuum (E-GRV-0083)	E-GRV-0090

Rows dropped since the last version, because they are no longer true of the current rule: the period-three vacuum clock and its clock-phase interference (the charge rule's), the chaos ceiling on records (it fails since a schedule fix), monopole absence (the light is now compact), and CP violation in the clock knits (the locked rule keeps CP).

The experiments

Everything is finite and deterministic, so every result reproduces exactly. Real numbers appear only as measured outputs, never in the base. Much of the code was written with AI assistance, which changes nothing about trusting it: run it and check. Each question is one experiment in test/experiment/<category>/, a single experiment returning a structured verdict (status, metrics, control, claim), and each grades itself by what it establishes, not by whether it passes:

level	meaning
L3	emergent and novel. One base rule produces the result as a measured consequence, with a control, ideally a quantitative prediction that could be wrong. The target.
L2	known physics. Reproduces a known construction on the substrate (a Dirac quantum walk, lattice gauge theory, a ballistic light cone).
L1	known math. Confirms an established mathematical fact (the 24-cell is the binary tetrahedral group, a 2π rotation gives −1).
L0	circular. The answer is put in by hand, so it proves nothing. Kept as a consistency note, never as evidence.

The full rubric and the rules the runner enforces (an L3 claim must carry a control) are in note/experimental-methodology.

test/catalog.csv is the full index, one row per experiment, generated from the registered experiments and sorted strongest first. It holds 1,556 rows in 18 categories:

total	L3 emergent, novel	L2 known physics	L1 known math	L0 circular	backing a paper claim
1556	53	1172	317	14	604

The largest categories are gauge (275), spin (188), selves (177), quantum (172), foundations (169), gravity (145) and relativity (112). The first full depth audit (2026-08-31) found 40 of 92 L3 experiments with no substrate, rule or coin in their import graph, and regraded them down. Every correction is recorded in the experiment it corrects.

The experiment map is the human way in: every arena one line per experiment, the coverage matrix, a concepts cross-index, reading paths, and a guide for adding your own. The quantum sector map is note/experiment/quantum-coverage.md.

Development

Every experiment lives in test/experiment/<category>/<name>.ts as one experiment. The suite runner (test/run.ts) imports them all through test/experiment/all.ts, the shared library is in code/, and the named batteries (conformance, paper) are in test/suite/. The build fails only on a code crash or a conformance failure, never on a scientific negative.

command	does
pnpm test	typecheck, then every registered experiment plus the conformance battery (about half an hour)
pnpm test:full	the registry run plus the legacy test/test.ts battery
pnpm rerun <code or id> [...]	run single experiments by code (E-GRV-0090) or id, and print every number the verdict holds, the fast loop
pnpm result <list, check, reproduce or audit>	the results database: list it, check it against the registry, reproduce a result's experiments, or audit one into a capsule (writes only on --commit)
pnpm call <file>	run one script under tsx
pnpm call test/catalog.ts	regenerate test/catalog.csv
pnpm check:labels	verify every experiment's substrates label against its import graph
pnpm check:constants	verify no typed constant reaches a verdict
pnpm check:coverage	check the quantum coverage map against the catalog
pnpm check:perturbation	the perturbation (robustness) checks
pnpm lint	eslint over code/ and test/
pnpm format	the formatter
pnpm make	compile the library to host/
pnpm host	build and publish the package
pnpm make:sm-scoreboard --commit	regenerate the scoreboard above from the observation ledger (run from the parent workspace, reports without --commit)

What is inside

• substrate: regular {p,q,...} hyperbolic honeycombs through the Coxeter engine, including the $\lbrace3,4,3,4\rbrace$ dock graph with O(log n) addressing, plus hyperbolic random graphs, regular lattices, Minkowski and curved sprinklings, and classical sequential growth.
• tone: the ternary alphabet and the directional fill carried on each dock.
• rule: the current knit's pieces (isometric-knit, bounce-pair-knit for the bounce and pass contacts, doublet-locked-knit, coined-locked-knit for the coin, occupation-veto-knit for the vetoes and the no-veto store), the husk light (trit-column, photon-shaped), the gravity registers (trit-radion, step-depth, depth-span-light, open-husk), and the earlier rules kept as history and controls (turningWeave, lineWeave and pairCollision in collision), with streaming, growth, and the discrete C, P, T transforms.
• measure: dimension, distance, curvature, manifold-likeness, Lorentz isotropy, streaming BFS shells, navigation, CHSH, locality, Wilson loops, Aharonov-Bohm phase, the knot network, and the named full-period path key (full-key-paths).
• operator: graph Laplacian, Kahler-Dirac and overlap fermions, the gauge-covariant Dirac, the cellular-automaton Hamiltonian, and the gauge index.
• algebra: quaternions and the binary tetrahedral 24-cell, the D4 and F4 root systems, spinor and vector rotation, Clifford and exterior calculus, and the linear-algebra kernels (Lanczos lowest eigenvalues, the kernel-polynomial method, Bethe resolvents).
• dynamics: the Benincasa-Dowker action, uniform-measure and Wang-Landau sampling, parallel tempering, coarse graining, and the Wilson heat bath.
• control: the negative controls that make a positive result mean something (the substrate or rule where the answer must be no).
• draw, render, and viz: renderers and figures for the bulk, the cusp, gliders, gravity, and the nesting tower.
• test/experiment: one experiment per question, in 20 directories (addressing, associative, computation, cosmology, data-structure, fluids, foundations, gauge, general, geometry, gravity, holography, matter, method, quantum, relativity, renormalization, selves, spin, substrate-survey), run by the suite runner in test/.

Documentation

All docs live in note/. The entry points:

• The predictions: the falsifiable calls, the technologies the mechanisms suggest, the frontier claims people ask about assessed inside the model, and the next steps.
• The library guide is how to use the code/ library. It opens with a features-at-a-glance page and an overview of how it fits together, then per-domain API guides and engine deep dives (the Coxeter tessellation engine, the reversible rule, the Kahler-Dirac fermion, the spinor coin, the spectral methods, the causal-set sampler, the unitary evolution, the lattice gauge engine, the coarse-graining and selves engine, and the associative memory engine).
• The math catalog lists every piece of math the library runs, what each module depends on, and which experiments use it.
• Architecture is where code and tests live, and how to add an experiment.
• Experimental methodology is the standard every experiment is held to: the depth rubric, the control requirement, determinism, and the negatives.
• Open problems are the negatives written up in full. The hardest is spacelike Bell correlations: what Bell's theorem proves, why a deterministic theory can still match quantum mechanics (it drops measurement independence, not determinism), and the price vibe pays for that.
• Cross-tessellation experiments is how to write an experiment that runs against every regular hyperbolic tessellation at once.

Reference data and verification

The experiments are only as good as the numbers they are compared to, so those numbers live in one cited place: note/data/reference/.

• What we gathered. Every external value an experiment must match or use as a comparison: the fundamental constants, the full Standard Model particle table, the roughly 26 free Standard Model parameters, the CKM and PMNS mixing matrices, the cosmological parameters, and the geometric and group-theory targets the model derives (the ternary 3, the 24 of the dock, the octonion ceiling 8, F4 order 1152, sin^2(theta_W) = 3/8, the Tsirelson bound, the Born exponent, and so on).
• What it contains. Structured CSV plus a machine-readable reference.json, with a prose readme and a bibliography. Every row carries a source tag and a verified date. The empirical values were fetched from and reconciled against their primary sources on 2026-06-24 (CODATA 2022, PDG 2024, NuFIT 6.0, Planck 2018).
• How we used it. The verification folder runs the comparison-bearing experiments live and diffs each measured number against the reference value, recording a status per experiment in cross-check.csv. It confirmed the matches (the quantum bounds, the 3/8 angle, F4, the warp factor) and caught real problems (two mismapped experiments, one circular result whose number was hardcoded, and one result stronger than the table recorded).

License

MIT. Open for science: use, modify, and build on it freely, with attribution. See LICENSE. The written results and figures are shared under CC-BY-4.0 (attribution).

ClueSurf

Made by ClueSurf, meditating on the universe.