Information Physics Institute : September 2026

Observer Patch Holography specifies a twelve-port carrier and one local repair and memory law, runs observers on it, and asks whether the records they share are our physics. An invitation to the Information Physics Institute.

TL;DR. Simulation theory has an argument and no machine. Observer Patch Holography (OPH) supplies a machine you can run. The hardware is a network of twelve-port icosahedral carriers. The software is one local move: where two neighbours disagree about their shared boundary, repair it and keep a record. On top of that law sit observers that remember, act on their records, keep local clocks and agree with each other exactly where their regions overlap. A library of more than 12,700 machine-checked theorems turns their readouts into familiar structure: public records that do not depend on who repairs first, the Born rule and the Tsirelson bound, exactly three spatial directions, a (1+3) causal order, the Standard Model gauge Lie algebra with one anomaly-free generation, and the Einstein equation from null balance. The bet is that one such process carries all of physics. Particles, an interacting 3+1 continuum and gravity on one responsive geometry are where that bet is open, and that is where we would like the Institute’s help.

Mission

The mission of OPH is a consistent and complete, formal and rigorous theory of reality. The theory has no starting point. It is a closed loop, and it has to meet four conditions.

Consistent. The loop closes. Its fixed point exists, constructed explicitly and machine-checked, and a theory with a model contains no contradiction.
Rigid. Every dimensionless quantity is forced. One measured value sets the units and nothing else.
Complete, in the sense Einstein, Podolsky and Rosen gave the word in 1935. Every element of physical reality has a counterpart in the theory, and every measured quantity is derived.
Correct. Derived values agree with measurement, and new predictions are frozen before they are tested.

OPH does not claim to be the only theory with these properties. The four conditions are the yardstick for everything below, and the last section of this post measures the project against them.

How the parts fit together

Basis. Closure hypothesis · internal-realization policy · operational agreement · adopted state selection and response endogeneity · one executable specification

↓↓

Executable specification

▸ twelve-port carrier and native record laws

▸ finite observers with coherent memory

▸ protected sectors and a fermionic field

▸ common clock and particle-changing drive

▸ energy and reference-state observations

▸ separate refining-screen constructions

finite instances
of the premises
⟶

⟵
theorems as
shortcuts

Finite theorem library

▸ normal forms; quantum record identities

▸ rank-three response; gauge Lie type

▸ anomaly-free matter; fermionic exchange

▸ protected quantum transport sectors

▸ local energy and modular identities

▸ Einstein tensor step from null balance

↓↓

Completion target effective descriptions that observers inside the construction must admit

finite observer algebraconstructed

shared operational meaningfinite instances

field–clock experimentsfinite instances

physical predictionsrules frozen in advance

effective quantum fields in 3+1open

classical and gravitational regimesopen

physical particle identificationopen

↺ Closure loop. Observers recover the specification from their records, which leads back to the basis.

How the parts of OPH fit together (after Figure 1 of the flagship paper). The basis fixes one executable specification and the premises of the finite theorem library. The running specification supplies finite instances of those premises, and a theorem replaces simulation wherever its premises are decided on the specification’s finite objects. Both feed the completion target: the effective descriptions that observers inside the construction must admit. The dashed box is the closure loop: observers recover the specification from their records.

The diagram is the whole method. On the left is a program. On the right is mathematics. The arrows between them are the reason the project moves: every time the program produces a finite object whose properties a theorem needs, the theorem applies, and every time a theorem covers a case, the program does not have to simulate it. Below both sits the target, with its status written into each row. Four rows are realized as finite constructions. Three are open: effective quantum fields in 3+1 dimensions, the classical and gravitational regimes, and the identification of physical particles.

Observers first

Physics usually begins with a world and asks how observers fit into it. OPH begins with observers and asks what world they can establish together. Its observers are finite. Each holds a patch of local state, reads a boundary it shares with its neighbours, keeps durable records, and repairs its state when records disagree. An objective fact is a record that survives every authorized comparison. Space, time, matter and classical objects are effective descriptions of what observers can measure and jointly interpret, and no observer-independent continuum is assumed beneath them.

An observer in this sense is defined by what it can access and what it records. It reads, writes and retains records, and it uses retained records to choose its next operation. Nothing in the definition refers to a background space, a global clock or a list of particles. Those have to come out as readouts that observers can share.

An argument without a machine

In 2003 Nick Bostrom published “Are You Living in a Computer Simulation?” in Philosophical Quarterly.¹ The paper is a probability argument about posthuman civilisations. It contains no circuit diagram. Konrad Zuse had asked the engineering version in 1969 in Rechnender Raum, and Edward Fredkin, Stephen Wolfram and Gerard ’t Hooft each proposed cellular automata underneath physics.² None of those automata returns the Lorentz group, the Born rule or $\mathrm{SU}(3)\times\mathrm{SU}(2)\times\mathrm{U}(1)$ , and none of them contains an observer that can be run and asked what it sees.

OPH takes the engineering question literally and turns it around. The machine is specified completely before anything is computed: carrier, record law, memory, scheduler, preparations and allowed interventions. Observers are programs on that machine. The question is whether what those observers can establish together is the physics we measure. A specified law that fails is rejected, along with its interpretation. That is what makes the hypothesis testable.

The carrier

Each local carrier has twelve boundary ports with icosahedral incidence: thirty seams, twenty faces, five seams at each port, every seam on two oppositely oriented faces. This is finite combinatorics, with no embedding in a pre-existing space. The sixty proper rotations of the icosahedron form the alternating group $A_5$ , and the twelve-dimensional space of port coefficients splits without multiplicity into four irreducible pieces,

$P_{12}\;\cong_{A_5}\;\mathbf 1\oplus\mathbf 3\oplus\mathbf 3'\oplus\mathbf 5 .$

That decomposition does a remarkable amount of work downstream. The $\mathbf 3$ becomes the three directions of space. The $\mathbf 3$ and $\mathbf 3'\oplus\mathbf 5$ become the two non-abelian ideals of the gauge algebra. The trivial $\mathbf 1$ becomes the centre. Carriers are glued along shared seams into a federation whose outward support is a sphere, refined level by level along the geodesic icosahedral tower.

The carrier is the committed hardware of the specification. Within the closure hypothesis it is the part observers recover from their own records, and the federation runs below show that they can.

The native repair law

The executable observer runs on a carrier with twelve integer loads of total two: 78 configurations, of which the 66 binary ones form an invariant shell. Pick a seam $(a,b)$ and flip a fair coin $c\in\{0,1\}$ . The native repair law moves

$d=x_a-x_b,\qquad j=\tfrac12\bigl[d-(2c-1)(|d|\bmod 2)\bigr],\qquad y=x+j\,(e_b-e_a).$

Load is conserved, and the signed current $j$ is what the carrier reports. Four requirements fix this rule: conservation, one-step balance, a tie-break that reads no load, and endpoint symmetry. On the total-two carrier they leave only the coin.

The quantum version is a record law, and seam locality plus the local phase symmetry of repair leave a convex family of them, each a table of five coherence multipliers. Two members are used throughout. Law C records the coin and the current. Law E also records the parity of the seam’s total load. They agree on every classical statistic and differ in which coherences they keep, which makes them distinguishable in a laboratory (see the instrument test below).

Some properties do not depend on which record law nature uses. Every admissible law has the uniform binary state as its unique stationary state and the same spectral gap, $(5-\sqrt5)/60$ . The longest-lived datum of a carrier is a classical position-like readout, because every coherence between distinct configurations decays faster than the slowest load pattern. On the binary shell each native repair is the relative-entropy-minimizing resampling of the theorem library, so the second law and the modular Clausius inequality hold for every single attempt. Repair has a direction because each projection discards the specific disagreement it settled and keeps only the fact that agreement was reached.

Public records, at scale

Two observers who share a boundary can disagree about what it shows. If repair stopped in different places depending on who acted first, the boundary would carry no fact of the matter. The consensus theorem says when it always stops in the same place. Under a transactional local-diamond condition (each commit revalidates what it read, and each conflict component has one canonical merge), Newman’s lemma from 1942 upgrades local confluence to global confluence: every maximal asynchronous repair schedule from the same initial state ends at the same consistent normal form.³ That endpoint is the public record, a fact no single observer owns.

The law runs at scale. On the level-six rung of the refinement tower, 81,920 exact twelve-port carriers with 2,580,480 seams settle under sixteen shuffled schedules, and all sixteen return one quotient hash. A standalone verifier that imports no simulator code rebuilds and checks every hash. The level-ten rung, 20,971,520 carriers and 660,602,880 seams, settles all sixteen schedules in 84 to 100 sweeps with one quotient hash.

The same records let observers recover the specification itself. From event logs alone, recovery returns the twelve ports, the thirty seams, the automorphism and rotation groups of orders 120 and 60, the mean rule, the rank-three Gram matrix and the Lie split $1+3+8$ . Tetrahedral and octahedral control carriers recover their own, different specifications.

The executable observer

This is the centre of the project. OPH runs observers inside the dynamics and reads physics off what they record.

An observer that remembers and acts. The native observer executes two repair events with a finite tape. It records the first current, issues a forecast from that record, chooses the next seam from the record alone, records the result and halts. All 264 binary-source paths are checked exhaustively. With the record wired to the seam selector, the final occupation of port zero is 35/264. Disconnect the selector, or shuffle it with the same action frequencies, and the answer is 1/6. The response is caused by the retained record. A playback of the same frames would fail this test the moment a retained state is changed, which is the difference between a realization and a description.

Two observers, one law. Observer A reads ports (0, 1) and observer B reads ports (0, 5) of one carrier, so their regions overlap at port zero. Each receives its initial bits and every incident local current, and its clock ticks whenever an attempted repair meets its region while remote repairs continue on the same source. A direct execution and an independent recursion agree exactly on all 5,768 output words, and at every positive-weight prefix the two observers’ decoded values at the shared port agree. Agreement on overlaps, the central condition of OPH, is instantiated on a running process.

Local clocks. A local event clock counts the attempted repairs that meet its region. Two such clocks on joined carriers share one continuing history, and over each two-tick cycle of one clock the other ticks twice on average with variance 16. Equal mean rates do not imply equal readings. Attempt counts, local event counts and modular parameters are distinct clocks in the model, and none of them is identified with seconds.

One interacting process. One execution puts quantum electric flux, a charged fermion pair and two neutral scalar sites on a single work register, with an operator Gauss law. Native source words implement their evolution and the pointer instruments, a committed matter bit controls feedback, and the final scalar bit controls the next native repair. A retained native phase witness separates law C from law E inside this process.

Identifying the repair law. Nine source-executed measurements read all nine real coordinates of the record-law family, which identifies the instrument on all 78 states.

From one carrier to a screen. Federations of 20, 80 and 320 carriers preserve complete source records and cross-carrier coherence. On screens of 15,360 and 245,760 ports a logical qubit with an untouched reference passes through native ticks and record-controlled feedback, and each full-screen output lies within $7.8\times10^{-10}$ of the local computation.

You can run the basic observer on a laptop with Python 3.11 and nothing else:

python3 -m model.observer run --trials 10000 --seed 42
python3 -m model.verify

The first runs the record-and-feedback experiment. The second replays every retained result and checks its hashes.

These executions are finite, and their apparatus, schedules and preparations are declared in advance. An autonomous unbounded observer, interacting spatial refinement and a laboratory clock calibration are work in progress.

Quantum mechanics on completed records

OPH places measurement at the moment a private, repairable record becomes a protected public normal form. Once a compare, write and verify slice has completed, its events live in a finite-dimensional complex algebra with a density operator, and the familiar identities hold there. Every valuation on the effects that is nonnegative, normalized and additive on coexisting effects has the form $E\mapsto\operatorname{Tr}(\rho E)$ for a unique density operator $\rho$ , in every finite dimension including two, where Gleason’s theorem fails, and with no continuity axiom. Among instruments with the right effects, invariance on the states certain of the outcome selects the Lüders update. Together,

$\Pr(E)=\operatorname{Tr}(\rho P_E),\qquad \rho\;\longmapsto\;\frac{P_E\,\rho\,P_E}{\operatorname{Tr}(\rho P_E)} .$

Correlations obey $|S_{\mathrm{CHSH}}|\le 2\sqrt2$ , attained exactly on an explicit entangled witness, while every state diagonal in the record basis stays at 2 or below. An isometry that copies two sharp states from a common blank forces their overlap to be zero or one.

The executable model adds a sharp point here. Committed native records form a classical process. Replaying the deposited records of the loophole-free Delft Bell experiment gives 245 trials and $S=2.42$ ,⁴ which rejects a local classical reading of the native records. Correlations beyond 2 therefore have to come from uncommitted coherence, and the model has to supply that coherence explicitly. Nothing collapses. A record gets written, and the network repairs toward consistency with it.

Three directions of space

Linearize uniform repair on one carrier and you get the mean $M=I-(5I-A)/60$ , with $A$ the icosahedral adjacency. The adjacency has eigenvalues $5$ , $\sqrt5$ , $-1$ and $-\sqrt5$ with multiplicities 1, 3, 5 and 3. So $M$ has eigenvalue

$1-\frac{5-\sqrt5}{60}$

on a three-dimensional eigenspace and smaller eigenvalues on everything else that is not constant. Repeated local averaging forgets every pattern except a three-dimensional one, and that surviving pattern is the $\mathbf 3$ carried by the port position vectors. The normalized response converges to the Gram matrix $G=4P_{\mathrm{slow}}$ , of rank three. The integer port records embed densely in one continuous three-dimensional Euclidean carrier, the sixty rotations act on it as isometries, and accumulated records act on it as translations. Position is the carrier’s own readback of its records. The same number $(5-\sqrt5)/60$ is the quantum spectral gap of the repair law.

When carriers join into the refining spherical screen, the continuum form of their repair history has spectral dimension two, the dimension of the screen. Reading the screen as the boundary of a three-dimensional bulk takes a radial observation map that is work in progress (see the next section).

Fields, waves and a smooth (1+3) description

Join the carriers into the geodesic screens of the refinement tower and let the native repair run, with no field, metric or Hamiltonian put in. Along the tower, the observers’ cell readouts converge to one positive operator $A$ . It plays three roles at once: it is the inverse covariance of the static field, the generator of the field’s decay, and the squared frequency of the admitted waves. Its heat kernel has a Gaussian bound, so those waves have a finite speed. Rescaled density readouts of the actual repair history converge to a reflection-positive Gaussian field, and a repair biased toward a profile obeys a fluctuation-dissipation relation set by the same $A$ .

On a declared radial extension of the screen, one action fixes correlations, response and a modular clock with signature $(-,+,+,+)$ . Smoothing the native coefficients then gives a theorem: for every finite menu of observer readouts and every $\varepsilon>0$ , one smooth static (1+3)-dimensional Lorentzian metric carries a free scalar field whose predictions for the whole menu agree within $\varepsilon$ with those built on the native measure. Observers on this construction have the effective description standard physics uses, as a theorem about their own readouts. The radial direction and the units are inputs, and a responsive metric and a spatially refining interacting field are work in progress.

Read more: book ch. 21 “Why Does Time Feel Like It’s Moving?” and ch. 22 · flagship §7.1 · integrated verdict

Causal order from who read whom

Every commit certifies which earlier records it read. The transitive closure of that authenticated read-from relation is a locally finite partial order, a causal set in the sense of Bombelli, Lee, Meyer and Sorkin, with the order coming from provenance.⁵ The longest chain of reads ending at an event gives it a height. Adjoin a real axis for height to the rank-three carrier $V$ and you get

$W=\mathbb R\oplus V,\qquad Q(t,x)=t^2-g(x,x),$

of dimension $1+3=4$ and signature $(1,3)$ . Unit directions in the carrier, the sphere and the future null rays are the same set. A closed convex cone preserved by the icosahedral rotations and by boosts along one axis is the light cone, so the cone is selected without a light postulate.

The carriers produce this order themselves. Each holds its record as port loads, reads its position through the rank-three response and decides its neighbours from that readback. Through 704,969 carriers reading one another over ten rounds, the generated precedence equals the layered order. On the largest interior diamond the Myrheim–Meyer dimension reads 3.55, 3.84 and 3.92 at the three largest levels, the ordering fraction approaches its proved flat value $1/10$ , and the fourth root of a ratio of record counts reads a ratio of proper times.⁶ Turning this finite precursor into a physical manifold is work in progress.

Gauge structure and one generation

Twelve ports give twelve independent ways to push the carrier, and a complete reversible response turns them into a twelve-dimensional compact Lie algebra. Require every carrier rotation to be implemented from inside, by the response group itself. Inner automorphisms fix the centre pointwise, and the only direction the rotations fix is the uniform push, so the centre is at most one-dimensional. A centreless algebra would be $\mathfrak{su}(2)^4$ , and a fixed-space count rules that out. The remaining eleven dimensions are semisimple, and $11=8+3$ is the only compact split:

$\mathfrak g\;\cong\;\mathfrak u(1)\oplus\mathfrak{su}(2)\oplus\mathfrak{su}(3),$

with $\mathfrak{su}(2)$ carrying the $\mathbf 3$ and $\mathfrak{su}(3)$ carrying $\mathbf 3'\oplus\mathbf 5$ . The port count fixes the dimension twelve; the content of the theorem is the exclusion of $\mathfrak{su}(2)^4$ and the forcing of a one-dimensional centre. An explicit matrix realization over $\mathbb Q(\sqrt5)$ is checked exactly on all sixty rotations.

On the supplied matter representation $V=\mathbb C^3_{-1/3}\oplus\mathbb C^2_{1/2}$ the exterior algebra branches as

$\Lambda^2V\oplus\Lambda^4V=Q\oplus u^c\oplus e^c\oplus d^c\oplus L$

with the textbook hypercharges, all perturbative anomalies vanish, and the common kernel is $\mathbb Z_6$ , so the global group is

$S\bigl(U(3)\times U(2)\bigr)\;\cong\;\frac{\mathrm{SU}(3)\times\mathrm{SU}(2)\times\mathrm U(1)}{\mathbb Z_6}.$

An exhaustive scan of all 1,024 sub-menus of the exterior algebra leaves exactly two rank-fifteen chiral anomaly-free masks, exchanged by charge conjugation: one Standard Model generation. The forced adjoint contains no mixed $(3,2,-5/6)$ generator, so the minimal grand-unified exchange channel is absent.

Matter as knots in the network

The executable-model hypothesis has a definite picture of particles. A stable massive particle is a localized topological defect of native overlap transport: a bounded core whose surroundings have settled, with a twist in its boundary transport that no local repair can undo. Repair keeps maintaining the core, and its rest energy is that excess maintenance, read in the same clock and energy dictionary as every other energy. Moving the core moves its charge, and creation, annihilation and decay have to follow from the same law.

Several pieces of this are built and run. On the 66-state binary configuration complex the fundamental group is $\mathbb Z/2$ . A coherent ordering memory turns the odd loop into a protected history sector, and independent internal states at occupied ports realize fermionic exchange with the correct anticommutation sign. Complete source factors join spatial propagation, particle-changing Higgs operations, native repair and one calibrated clock on a single fermionic field algebra, with every quantum outcome and reference record retained. The three-dimensional Weyl/contact action has a positive-energy, rigorously non-Gaussian ground on the infinite lattice, and its continuum limit is a field net with all eight Weyl valleys, continuous translations, positive energy and finite propagation.

The measured particle spectrum is the open frontier. The simplest local defect, a doubly loaded port, is transient: it survives $n$ attempts with probability $(5/6)^n$ . The fixed mass laws tested so far give charged-lepton ratios 1 : 1.19 : 1.50 against the measured 1 : 206.8 : 3477. Locating the defect sector whose maintenance energy is its rest energy is the single largest open task in OPH.

Gravity from null balance

Ted Jacobson showed in 1995 that Einstein’s equation follows from thermodynamics on local horizons.⁷ OPH isolates the step that turns such arguments into a tensor equation. Light rays cannot see a term proportional to the metric, so if curvature and stress balance on every null direction, Einstein’s equation holds up to such a term. Nine explicit null directions reconstruct the quotient exactly, with design determinant 8192/27. The contracted Bianchi identity and stress conservation then make the coefficient constant, and that constant is the cosmological constant:

$R_{ab}-\tfrac12 R\,g_{ab}+\Lambda g_{ab}=8\pi G\,T_{ab}.$

The tensor step is machine-checked from the null-balance premise. The joined construction supplies the matter side: the free Weyl continuum has a conserved stress and geometric modular flow on double cones, and at zero coupling finite regional entropy changes converge to the free-field stress response with a controlled small-ball remainder. Supplying the null balance itself means realizing the area law, entropy stationarity and a responsive metric on one refinement tower together with that matter. That join is work in progress, and the tested static radial geometry fails it for its own scalar excitation.

The numbers: two closure coordinates

The architecture carries two dimensionless coordinates, the screen grain $P$ and the record capacity $N$ . Under closure both are fixed points, and the loop licenses reading them off: locate the constant by observation, declare a map, and certify a fixed point inside the measured basin.

The grain loop

$P=\varphi+\frac{\sqrt\pi}{A_T(P)},$

with $\varphi=(1+\sqrt5)/2$ and $A_T(P)$ the inverse coupling a declared map returns at trial grain $P$ , has one interval-certified fixed point per map. The self-consistent map settles at $P=1.6309682414$ and $\alpha^{-1}=137.035660$ , 2.5 parts per million from the CODATA value 137.035999. The capacity, read as de Sitter horizon entropy in nats,

$N=\frac{3\pi}{\Lambda\,\ell_P^{2}},$

gives $\Lambda$ within 0.4 percent of the Planck ΛCDM value. The same reading gives the dark-energy equation of state

$w(a)=-1+\frac13\,\frac{d\ln N}{d\ln a},$

so fixed capacity means $(w_0,w_a)=(-1,0)$ exactly, with no continuous freedom, and any $w<-1$ would mean the universe is losing memory. Under two named premises the deep galaxy regime has a unique law, $v^4=G M_b a_0$ , which forces flat rotation curves.

Both closure roots count as diagnostics. The maps carry the hadronic transport through a declared quark-screening ansatz, and with the measured hadronic payload the grain root lands at 136.38, so the transport from the cell to the Thomson limit is the lane that decides whether $\alpha$ becomes a prediction.

Tests fixed in advance

Two tests follow directly from the construction, and their decision rules are fixed before any eligible data.

A dispersion band at angular rank six. Any kinetic operator with icosahedral symmetry is exactly isotropic through angular rank five, at every order. The first possible directional term sits at rank six and is a multiple of one invariant. For the whole positive-weight carrier class, write

$\omega^2=k^2+C_4k^4+\bigl(B_0+B_6\,\mathcal I_6(\hat k)\bigr)k^6+O(k^8).$

Then

$C_4<0,\qquad \frac{B_0}{C_4^2}\ge\frac{10}{21},\qquad -\frac{16}{135}\le\frac{B_6}{B_0}\le\frac{16}{75}.$

A resolved intrinsic dispersion outside that band, or any residue at ranks one through five, excludes every member of the class at once. The primitive twelve-port branch predicts the exact point $B_6/C_4^2=32/315$ , $B_0/C_4^2=10/21$ , and that prediction is registered with OpenTimestamps proofs.

An instrument test from the executable observer. Laws C and E agree on every classical statistic, and one coherent interference experiment tells them apart. On a fixed two-state preparation the readout probability is

$p(\theta)=\frac38+\frac12\operatorname{Re}\bigl(g\,e^{-i\theta}\bigr),$

with $g$ the coherence multiplier of the law. Reversing the phase moves C from $5/8$ to $1/8$ , E predicts $3/8$ for both, and the law that records only the current predicts $3/8\pm\sqrt2/4$ . A source-executed implementation runs the complete instruments from twelve-port control words, and the evaluator reports “no data” until independent counts are supplied.

The loop

Everything above enters the loop at the executable specification, because that is where observers stand: holding records and inferring. The loop itself is the one hypothesis of OPH. Let $\mathfrak U$ be a complete inhabited history, $D(\mathfrak U)$ the account of its architecture that observers can recover from records inside it, and $R$ a realization of that account built inside $\mathfrak U$ . Closure says

$R\bigl(D(\mathfrak U)\bigr)\;\cong\;\mathfrak U .$

Observers infer the specification of their world, build hardware within it and run that hardware, and the execution realizes the same history, including its builders. No external computer runs the history, and no extra time exists in which the whole is played. The universe reads itself into existence.

A loop has no first point. The pair $x=(y+2)/2$ , $y=(x+2)/2$ fixes $x=y=2$ with no first value from which the other is manufactured. A calculator may approach it by iteration and someone may find it on a particular day, and neither is the reason the answer is two. Self-reference earns content only when the return condition restricts the possibilities: $x\mapsto x+1$ has no fixed point, $x\mapsto x$ fixes every point, $x\mapsto(x+2)/2$ fixes exactly one. On the declared finite consensus branch the closure map is an affine idempotent whose fixed points are exactly the consensus records. The recovery runs above are the finite form of $D$ : observers reading the carrier’s specification back out of their own event logs.

Where each question stands

Question	What OPH establishes	Work in progress	Where
Objective facts	Public records independent of repair order; settles at 20 million carriers	An implemented scheduler inside the process	book ch. 10 · consensus paper
Measurement problem	Measurement is a record becoming public; Born weights without continuity axiom; Lüders; Tsirelson; no-copying	A source-produced complete effect web	book ch. 19, 20
Observers	A running observer whose actions are caused by its records; overlapping observers exact on 5,768 outputs; local clocks	One autonomous unbounded observer	book ch. 19, 22 · observer paper
Arrow of time	Every native repair obeys the second law; gap $(5-\sqrt5)/60$ for every record law	A laboratory clock calibration	book ch. 16, 23
Three dimensions	Rank-three slow response with Euclidean completion	Physical length scale	book ch. 15
3+1 spacetime	(1+3) causal order from who read whom; dimension near 4; proper time from counts; smooth (1+3) scalar description	Physical manifold; native radial direction	book ch. 7, 17
Gauge group	$\mathfrak u(1)\oplus\mathfrak{su}(2)\oplus\mathfrak{su}(3)$ forced; $\mathbb Z_6$ quotient; one anomaly-free generation from 1,024 menus	Laboratory currents; family number	book ch. 26, 27
Matter	Fermionic exchange from a protected history sector; infinite-volume Weyl ground; eight-valley continuum	Localized defects with the measured spectrum	book ch. 28 · issue 1143
Gravity	Einstein tensor completion from null balance; free-field stress and modular limits	Area law, stationarity and responsive metric on one tower	book ch. 25
Dark sector	$(w_0,w_a)=(-1,0)$ at fixed capacity; deep galaxy law $v^4=G M_b a_0$	Source-only value of $a_0$	book ch. 29
α and Λ	Certified fixed points at 2.5 ppm and 0.4 percent	Hadronic transport to the Thomson limit	book ch. 29
Falsifiability	Frozen rank-six dispersion band; C/E instrument test	Independent data	prediction ladder

Where we need you

We think we are on the right track, and the reason is concrete. One specified local law, with no space, clock, field or particle put in, runs observers whose shared records carry public facts, quantum statistics, three directions, a (1+3) causal order, the Standard Model gauge structure and the Einstein tensor equation, each piece proved or executed with retained receipts. Whether the same process carries the rest of physics is the question we want people who think differently to attack.

The open work has a precise shape. Five realization maps connect the finite construction to the laboratory, and each is a place to contribute:

Position and scale. Identify the rank-three carrier completion with physical position, with a physical length and a refinement limit.
Clock and energy. Calibrate the model’s clocks and energies against laboratory duration and energy across the whole process.
Laboratory current. Identify the port response with a physical gauge current, with a matter action and a continuum limit.
One common tower. Realize the event, modular, stress, entropy and vacuum premises on one refinement family.
One matched action. Realize the field and interaction law with the same native services that run the observers.

The most concrete entry point is the particle search. A particle should be a knot that local repair cannot undo, and its mass should be its upkeep. Run python3 -m model.verify, audit a particle readout under model/, or freeze a defect-to-particle dictionary before comparing it with any measured number; issue 1143 has the rules. If you can build a countermodel to the local diamond, if you know how to do the hadronic transport to the Thomson limit properly, if you have resolved dispersion data or an interference setup for the C/E test, or if you think an assumption smuggles in what it claims to deliver, open an issue or write to bernhard@floatingpragma.ai. Thinking differently is good. That is what the Institute is for.

Everything is public. The Lean library contains more than 12,700 theorems and lemmas with no admitted proofs. Every retained computational receipt is replayed under continuous integration, with exact rational arithmetic and outward-rounded intervals in place of floating point. Each result carries its type (finite theorem, certified computation, source limit theorem, reconstruction implication, diagnostic or frozen prediction), and the common objections page answers the questions people ask first. The repository and the reproduction guide are the place to start.

License

This essay links back to the canonical OPH license:

https://github.com/FloatingPragma/observer-patch-holography/blob/main/LICENSE

References

Nick Bostrom, “Are You Living in a Computer Simulation?”, Philosophical Quarterly 53 (211), 2003. https://doi.org/10.1111/1467-9213.00309↩︎
Konrad Zuse, Rechnender Raum, Vieweg, 1969 (English translation “Calculating Space”, MIT Project MAC, 1970). Stephen Wolfram, A New Kind of Science, Wolfram Media, 2002. Gerard ’t Hooft, The Cellular Automaton Interpretation of Quantum Mechanics, Springer, 2016. https://doi.org/10.1007/978-3-319-41285-6↩︎
M. H. A. Newman, “On theories with a combinatorial definition of ‘equivalence’”, Annals of Mathematics 43, 1942. https://doi.org/10.2307/1968867↩︎
B. Hensen et al., “Loophole-free Bell inequality violation using electron spins separated by 1.3 kilometres”, Nature 526, 2015. https://doi.org/10.1038/nature15759↩︎
Luca Bombelli, Joohan Lee, David Meyer and Rafael D. Sorkin, “Space-time as a causal set”, Physical Review Letters 59, 1987. https://doi.org/10.1103/PhysRevLett.59.521↩︎
J. Myrheim, “Statistical geometry”, CERN preprint TH-2538, 1978. David A. Meyer, The Dimension of Causal Sets, PhD thesis, MIT, 1988.↩︎
Ted Jacobson, “Thermodynamics of Spacetime: The Einstein Equation of State”, Physical Review Letters 75, 1995. https://doi.org/10.1103/PhysRevLett.75.1260↩︎

Information Physics Institute

Translate

Wednesday, September 9, 2026

From Simulation Theory to Concrete Implementation