Skip to content

INDEPENDENT RESEARCH & ENGINEERING / d0re

THE PERSON & THE WORK

Anthony
Michael Collado.

Questions become models.
Models become experiments.
The evidence opens the next question.

Mathematics. Simulation. Model recovery.
Systems that retain their context.

Explore the work ↓

01 / THE PERSON

One independent practice.
A shared question.

Read the first About edition →

His published name is Anthony Michael Collado; the about, paper and repository pages use the shorter Anthony Collado.

Source & context

https://www.d0re.com/film → (footer); https://www.d0re.com/about/2026-09 →

He describes himself as an independent researcher. The preprint byline names him and d0re.

Source & context

https://www.d0re.com/paper.html →

Anthony Collado · independent researcher (d0re)

d0re is a one-person project with a single long-running question about systems that can show how they reached a state.

Source & context

https://www.d0re.com/about/2026-09 →

d0re is Anthony Collado's independent attempt to build and measure that.

d0re stands for Dimension Zero Reverse Engineering. It is a project name, not a company.

Source & context

https://www.d0re.com/about/2026-09 →

The name d0re stands for Dimension Zero Reverse Engineering. It is a project name, not the name of a registered company.

His stated through-line across research and products is visible state: what happened, what a claim rests on, and where the evidence stops.

Source & context

https://www.d0re.com/about/2026-09 →

The through-line is visible state: a computer should help a person see what happened, what a claim rests on, and where the evidence stops.

In his own words, the Theory of Reality connects his mathematics, simulations, AI research and working systems.

Source & context

https://www.d0re.com/film →; https://www.d0re.com/journey/content.json ↗ (chapter 'begin')

The Theory of Reality is the thread through my mathematics, simulations, AI research, and working systems.

The site names him as founder and sole contact: there is no sales team, no sign-up form and no payment. He welcomes technical discussion, research collaboration, early-access conversations, corrections and challenges.

Source & context

https://www.d0re.com/about/2026-09 →; https://www.d0re.com/contact.html →

Your message goes to Anthony Collado, the independent founder behind d0re.

He invites people to challenge the research.

Source & context

https://www.d0re.com/contact.html →

Corrections and challenges are welcome.

He builds d0 OS on his own and uses its desktop every day. It is not a public release.

Source & context

https://www.d0re.com/d0os/2026-09 →

Anthony Collado develops d0 OS independently; it is not yet a public software release.

His public code is at github.com/sendmefather (gauge-break, MIT licence), signed with his name and the d0re expansion.

Source & context

https://github.com/sendmefather/gauge-break#readme ↗

MIT licensed. Anthony Collado, Dimension Zero Reverse Engineering

He wants corrections published alongside the product story.

Source & context

https://www.d0re.com/about/2026-09 →

As the work changes, public claims should change with it. Research and corrections belong beside the product story, not behind it.

02 / THE APPROACH

How the work
can be inspected.

Start with a claim. Find its test. Read the boundary. Follow a real case.

01

Set the bar before the data

Anthony fixes the pass/fail gates, frozen parameters or withheld data before a test runs. A missed gate is reported as a finding, not re-tuned.

Source & context

library records: Hydrogen Matrix Lab 0.8.0 →, Reality revision 1.58 — Frozen hydrogen-isotope holdout →, A fitted expansion rate has to survive the next observer. →, Wave oscillation and higher spatial dimensions: calculation audit →

02

Check against exact ground truth, with independent code and controls

Where the space is finite, the programme enumerates all of it and checks with code that shares nothing with the production path. A deliberate control must fail.

Source & context

https://www.d0re.com/paper.html → §7; library records: Dimensional Chess Experiment C1 →, DRCT Black-Box Transformer Stress Test v6 →

03

Keep a ledger: retract, correct and keep the failures

Wrong arguments are withdrawn in writing, the replaced versions are kept, and failed benchmarks stay in the record. Nothing is quietly deleted.

Source & context

library records: Theory of Reality — revision 1.14 →, Theory of Reality →, RRLS v0.27 source audit and v0.27.1 repair →, A fitted expansion rate has to survive the next observer. →

04

Try to kill your own claim first

Novelty and breakthrough claims get adversarial audits and prior-art searches aimed at destroying them. Separate audit documents test the main manuscript without editing it.

Source & context

library records: Dimensional Reality: Court-Grade Mathematical and Physical Audit →, RRLS v0.28 — Adversarial Proof Gate →, Appendix BS: mathematical and evidence audit →

05

State the scope and refuse the overclaim

Each result gives its limits and names the larger claim it does not make: Millennium problems, full chess, P=NP, physical predictions.

Source & context

https://www.d0re.com/research/state-of-the-theory.html →

06

Publish the method so others need not take his word

Where a result cannot close its own gap, he publishes runnable code, run records and hashes so a reader can check it independently.

Source & context

https://www.d0re.com/paper.html → §8; https://github.com/sendmefather/gauge-break#readme ↗

Method descriptions are linked to their supporting records.

03 / THE WORK

Different forms.
Connected questions.

A research programme and the systems built alongside it. Each project has a question, a declared status and an evidence boundary.

01Under continuing revision. The current manuscript is 2.42 (15 October 2026); its full text is not recovered in the register. No physical prediction is admitted yet.

Theory of Reality

A mathematical framework that keeps a complete state, its dynamics, its interactions and the observer's record apart.

57 engine records mapped; 1,365 tests in the recorded native replay (31 groups); 0 / 57 records admitted for a new measured physical prediction.

Explore the research records →
Source & context

https://www.d0re.com/research/current-state.json ↗; https://www.d0re.com/research/index.html →

02Published 15 October 2026. The 18 underlying source reports are not recovered in the register.

Research edition 02: seven investigations

Seven published investigations: Aharonov–Bohm observation, Hubble transfer, measurement discipline, finite wave modes, competing explanations, Scripture-led questions, and the state of the theory.

The withheld-local-H0 study did not single out the local expansion value. "The programme has not solved the Hubble tension or identified a physical extra spatial dimension."

Explore the project →
Source & context

https://www.d0re.com/research/hubble.html →; https://www.d0re.com/research/index.html →

03Published preprint, not peer reviewed. First published 16 September 2026, revised 28 September 2026. Code is public under the MIT licence.

Quantisation breaks the LayerNorm gauge (GPT-2 preprint)

Recovers every GPT-2 small weight from internal activations by showing that float32 rounding breaks a symmetry that blocks recovery in real arithmetic.

124,439,808 / 124,439,808 parameters recovered; the sha256 matches the public checkpoint and was reproduced on two unrelated machines. The copying objection remains open.

Read the preprint →
Source & context

https://www.d0re.com/paper.html →; github.com/sendmefather/gauge-break

05Exact and numerical experiments, v0.13 to v0.30; v0.30 is the latest. P = NP is stated as not proved.

RRLS: relations, rewrites & phase

Asks when recognising the right relation removes unnecessary search, with every cost of recognition and rewriting counted.

"v0.30 reports 120/120 independent small #SAT checks and 100/100 exact integer Walsh controls."

Explore the research records →
Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.rrls)

06Finite-domain reconstruction series, v1 to v21 (all saved 23 August 2026); v21 is the latest.

DRCT: black-box reconstruction & invented concepts

Asks what an oracle reveals about a hidden model, and how a learner can invent reusable relation classes.

"The v6 lookup clone covers all 780 possible contexts, with zero maximum oracle-logit difference and zero argmax mismatches."

Explore the research records →
Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.drct)

20 more projects & investigations
07Exact finite proofs, C1 to C8 plus R1; C8 is the latest. The game value of full chess is not solved.

Dimensional Chess: C1 through C8

Tests which symmetries and abstractions actually reduce the cost of proofs in exact chess endgame universes.

"C1 solves all 399,112 legal states in its declared KRK universe, reducing them to 50,015 symmetry classes; the maximum forced mate is 32 plies."

Explore the research records →
Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.chess)

08Synthetic system identification, 0.14.0. All observations are synthetic.

Rope Tomography & sequential validation

Asks whether causal delays can be recovered from a boundary-driven wave.

"The Markov-sign release crosses on 39/5,000 declared 90%-persistence null histories and detects 4,629/5,000 97%-persistence alternatives."

Explore the research records →
Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.rope)

09Quantum reference and hypothesis tests, 0.8.0. The proposed rotor was rejected.

Hydrogen Matrix Lab

Tests whether a projected rotor geometry can reproduce a declared quantum reference.

"The proposed rotor fails four of five preregistered gates; a four-component rescue improves held-out Hellinger error by only 0.34% and still fails five of six."

Explore the project →

Research records →

Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.hydrogen)

11Synthetic and published-data comparisons, 1.5.0. The holdout shows no claimed improvement.

Matter Harmony & observer banks

Asks which spectral and observation coordinates are needed to tell physical behaviours apart.

"A two-frequency linear 5D observer cannot reproduce four distinct Balmer features; exposure changes visibility, not frequency support."

Explore the research records →
Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.matter)

12Conditional mathematics and numerical audits. Fixed Grid 0.25.0 is a new reconstruction; some original files are lost.

Fixed Grid String & coupled gauge–matter

Coupled field-and-matter models whose conservation laws stay checkable.

"The 2D plaquette release records 142 passing tests and an independent final-state agreement within 1.76e−13."

Explore the research records →
Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.gauge)

15Interactive transport model with uncalibrated closures.

Water, dye, light & fields

An interactive model separating fluid motion, diffusion and prescribed electromagnetic forces.

"At its displayed scale, molecular diffusion alone travels about 0.09 mm in ten seconds; visible curling mainly comes from moving water."

Explore the project →
Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.fluids)

17Generator tests and historical holdouts. No live trading edge is claimed.

PRNG tomography & market interpreters

Asks which apparent randomness can be recovered inside a declared model family, and which predictions hold on unseen data.

"v16 scores 6,000 historical test hours at 52.367% aggregate direction accuracy, with a reported moving-block 95% interval of 51.0–53.533%."

Explore the research records →
Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.prng-markets)

18Proposal with a conditional theorem (GWSN 0.1). Nothing has been implemented or trained.

Grounded Wave Semantic Network

A proposal to encode explicit semantic structure as waves and transform it under stated constraints.

"GWSN 0.1 reports no implementation, training run, or benchmark."

Explore the research records →
Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.semantic)

19Preliminary trained prototype: design 30 September, benchmark 1 October 2026.

POWM / Octave: eight learning observers

A small NumPy recurrent prototype in which eight temporal observers learn to predict through a shared workspace.

"Held-out Level 2 A+B bit accuracy: 92.70% recurrent versus 84.16% dense; complete forecasts: 9.90% versus 4.23%."

Explore the research records →
Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.octave)

20Live on the site.

Living Engine

Ten interactive browser instruments that connect a defined model state to an observation, each with its scope labelled.

"Ten browser instruments · individual model scopes"

Explore the project →
Source & context

https://www.d0re.com/film →

21The desktop is in daily use on one laptop. Home is in test. Not for sale or download.

d0 OS

A NixOS/GNOME working environment: a daily desktop, an alpha server and a home edition in test.

Secure Boot has been enforcing with d0's own keys since 26 September 2026. The Home installer passed a VM test. "This daily-use build has been tested on one laptop."

Explore the project →
Source & context

https://www.d0re.com/d0os/2026-09 →

22The local tools run daily. The shared service is in test. The source is private.

d0fs: versioned agent memory

Versioned work records, with hooks that bring earlier file and task context into coding-agent sessions.

23 / 23 local and service hook cases matched; 8 sessions at one call per second gave zero errors. "not a claim that a computer or its agent sessions are tamper-proof."

Explore the project →
Source & context

https://www.d0re.com/d0fs.html →

24In daily use; the wider server is alpha.

D0 Remote Desktop

Remote desktop with TLS, certificate pinning and checked computer identity.

About 51 ms from key press to screen at 1920 × 1080 in a 27 September 2026 bench test; 30/30 identity and connection UI checks passed in a VM.

Explore the project →
Source & context

https://www.d0re.com/d0os/2026-09#remote →

25Alpha on one test server. Monitoring and the private CA are in test. No production deployment.

d0 OS Server

Accounts, devices, certificates and health checks managed from one web console.

"The d0re home page counts 46 health-monitor rules from source on 27 September 2026."

Explore the project →
Source & context

https://www.d0re.com/papers/catalogue.json ↗ (projects.server)

26Private. They are mentioned on the d0fs page but have no download or evidence page.

prism and d0hub (named private projects)

prism grades how well each memory is supported. d0hub applies d0fs rules to Git, refusing history rewrites without a receipt.

"These are private projects, not public downloads."

Explore the project →
Source & context

https://www.d0re.com/d0fs.html →

Follow the papers, versions & next questions.Enter the library →

04 / THE RECORD, WITH ITS DENOMINATORS

Figures as of 29 Oct 2026

197

records in the library

167 public, 18 not-recovered, 12 on-request

Source & context

https://www.d0re.com/library/register.json ↗

439,505

words

counted across the 171 of 197 records with recovered text

Source & context

https://www.d0re.com/library/register.json ↗

Records dated 2026-08-23 to 2026-10-15

171 of 197 records carry a date

Source & context

https://www.d0re.com/library/register.json ↗

10

version series

110 records belong to a series

Source & context

https://www.d0re.com/library/register.json ↗

3

versions of the Theory of Reality manuscript in the library

including versions whose text is not published

Source & context

https://www.d0re.com/library/register.json ↗

Status of every record: 68 superseded, 62 reference, 56 current, 7 corrected, 4 unavailable

137 of 197 statuses come from reading the record with an independent second check; 51 follow version order; 9 are declared by the edition or mark an unpublished source

Source & context

https://www.d0re.com/library/register.json ↗

Record types: 64 report, 35 note, 26 manuscript, 18 documentation, 15 revision-note, 14 audit, 7 article, 6 claim-ledger, 5 test-log, 4 dataset, 2 proof-attempt, 1 source-note

197 records

Source & context

https://www.d0re.com/library/register.json ↗

254

entries in the ledger of negative results, corrections, failed attempts, repairs and retractions

each with a verbatim quote from its record

Source & context

https://www.d0re.com/library/register.json ↗

9

reading paths

66 steps, each anchored to records

Source & context

https://www.d0re.com/library/register.json ↗

606

archive files behind the register; 160 hosted catalogue documents, each SHA-256 hashed

23 project briefs in the catalogue (snapshot 2026-10-01)

Source & context

https://www.d0re.com/library/register.json ↗; https://www.d0re.com/papers/catalogue.json ↗ (audit)

57

engine records (40 starting + 17 auxiliary): 27 synthetic, 16 formal, 7 measured reanalysis, 3 mixed, 2 ML, 1 interpretive, 1 proposal

57

Source & context

https://www.d0re.com/research/current-state.json ↗

1,365

tests in 31 groups, 71 distinct pinned archive inputs

recorded native replay from the saved package; not rerun for the website edition; software tests, not physical laws

Source & context

https://www.d0re.com/research/current-state.json ↗; https://www.d0re.com/research/state-of-the-theory.html →

0

records admitted for a new measured physical prediction

of 57 engine records

Source & context

https://www.d0re.com/research/current-state.json ↗; research/index.html

Rerun on 15 October 2026: 38 software tests plus 11 synthetic propagation gates; 4,113 finite-wave model cases with 6 independent checks; C8 verifier with 168,882 move instances

bounded local rechecks only

Source & context

https://www.d0re.com/research/current-state.json ↗ (localRechecks)

22

of 51 claims recorded as rejected (19), retracted (1), negative (1) or false without additional assumptions (1)

51 claims in claim ledger v1.39, 2026-09-20 (an archive attachment, not a register record)

Source & context

https://www.d0re.com/archive/a4ea8f7ee791-claim_ledger.json ↗

124,439,808

of 124,439,808 GPT-2 small parameters recovered bit-exactly; 107,214,677 lattice candidates enumerated for 8 columns, 8 survivors

one public checkpoint, two machines; not peer reviewed

Source & context

https://www.d0re.com/paper.html →

Embedding-gauge lattice attack reached 820,496 of 39,383,808 values (2.08%), reported as a negative result

39,383,808 embedding values

Source & context

https://www.d0re.com/paper.html → §5

Lost original bytes: 17 C8 individual records, plus five older Fixed Grid files and trailing 0.24 content

gap H01, status not_recovered

Source & context

https://www.d0re.com/research/current-state.json ↗ (gaps)

Counts describe the indexed collection. They do not count independent replications or establish the truth of a claim. Inspect the complete register ↗

05 / THE DATED RECORD

See what changed.

  1. Eighteen DRCT manuscripts, v1 to v21, are saved with the v6 transformer stress test; PRNG-tomography and market-state notes are saved the same day.

    Dimensional Relational Closure Theory (DRCT) v1 →DRCT v21 — Morphism Ontology Induction →DRCT Black-Box Transformer Stress Test v6 →Dimensional PRNG Tomography v4 — Formal Core →Dimensional Market State v6 →
  2. Projective Ripple Reality 1.0 formalises observation as a projection of a richer state; Reality Bit Engine supplements v0.11–v0.13 add restricted proofs, a standalone verifier and stated obstructions.

    Projective Ripple Reality →Proof attempts for three Millennium problems →Independent certificate verification →Missing mathematical arguments: reductions and obstructions →
  3. Revision 1.14 retracts the seven-stage refinement theorem after the Rope Tomography Lab 0.2.0 audit: complete deterministic closure is idempotent.

    Theory of Reality — revision 1.14 →Rope Tomography Lab →Universal-Cycle Tomography: a causal boundary model and its limits →
  4. A pasted note names RRLS; RRLS Studio v0.7–v0.26, the v0.13 audit with its SAT extensions, and engine notes v0.14–v0.20 are saved. The P = NP gate stays closed.

    Relational Rotation-Lift Solver: origin note →RRLS Studio v0.7 →RRLS v0.13 — P≟NP Proof Search →RRLS: A Proof Attempt at Universal Relational Compilation →A computable all-time bound for small periodic fluid data →Null-Axis Probes and the Limits of Projection →
  5. RRLS v0.27 is audited and repaired as v0.27.1; v0.28–v0.30 locate the missing operation in polynomial semantic canonicalisation of residual futures.

    RRLS v0.27 source audit and v0.27.1 repair →RRLS v0.28 — Adversarial Generator Compiler →RRLS v0.30 — Dual-Phase Expansion/Collapse Probe →
  6. A court-grade audit finds a coherent mathematical framework with an eight-cycle hypothesis, not proven physics; Wallbreaker Protocol P3 defines the unactivated three-qubit P3-R1 pilot.

    Dimensional Reality: Court-Grade Mathematical and Physical Audit →Wallbreaker Protocol P3 →P3-R1 Hardware Package Audit Report →
  7. Dimensional Chess C1–C8 and run R1: exact KRK and KQK solutions, and independently verified no-forced-mate results from the initial position.

    Dimensional Chess Experiment C1 →Dimensional Chess C8: Movement-Frequency Relations →Dimensional Chess: Initial-Position Mate Frontier R1 →
  8. Rotational complexity of matrices: exact Fourier-rotor reconstruction of every benchmark matrix, with rotor count shown to be a basis-dependent measure.

    Rotational Complexity of Matrices — Theory and Experimental Results →
  9. Hydrogen Matrix Lab 0.1–0.7, Matter Harmony revisions 1.44–1.58, Element Projection Lab 0.6–0.12 and Fixed Grid String Lab 0.1 are saved; the hydrogen rotor fails its frozen holdout.

    Hydrogen Matrix Lab 0.1.0 — checked result →Hydrogen Matrix Lab 0.3.0 — spectroscopy calibration and frozen rotor holdout →Reality revision 1.58 — Frozen hydrogen-isotope holdout →Element Projection Lab 0.6.0: ATLAS publication-level dimension-evidence audit →Fixed Grid String Lab 0.1.0 results →
  10. Manuscript draft 1.70 lists the engines and labs as its computational formalisation and adds an Inversion Theory appendix; audits on 21–22 September reject several of its inferences.

    Theory of Reality →Appendix BS: mathematical and evidence audit →Wave oscillation and higher spatial dimensions: calculation audit →
  11. Musical readings of hydrogen and element spectra are tested and not supported; Element Projection Lab 0.13–0.16, Fixed Grid 0.2–0.7 and the Wave–Observer formulation are saved.

    Hydrogen, musical octaves, and alpha: checked results →Do atomic spectra share a musical correction factor? →Crystal wave observer — Element Projection Lab 0.15.0 →Fixed Grid String Lab 0.2.0 — bounded results →Theory of Reality: a Wave–Observer Formulation →
  12. Manuscript 1.97: appendices through CT on coupled gauge–matter checks and on the identifiability of shadows, optical tails and paired references.

    Theory of Reality →
  13. The spectral studio is validated and gains a Yee-grid Maxwell diffraction mode, presented as a demonstration without convergence or experimental agreement.

    Spectral engine validation →Maxwell diffraction build — v0.3 →
  14. OCTAVE's eight-observer design review and the GWSN theorem; a saved-artifact collection of 666 project files and 88 release ZIPs is assembled.

    POWM / OCTAVE — Eight-observer predictive memory architecture →Grounded Wave Semantic Network Architecture and Theorem →The Theory of Reality — complete saved-artifact collection →
  15. The OCTAVE school benchmark reports a mean per-bit transfer gain, but every new recurrent checkpoint scores 0/8 exact basic-pulse forecasts.

    OCTAVE — School training and benchmark results →
  16. A finite observer and nested-clock lab is executed: cosine readout hides rotation direction, and dephasing changes interference without added energy.

    Finite observer and nested clock lab →
  17. A computation route and the coupled eight-clock Maxwell execution find an exact ordinary Lorentz counterpart, an identification limit; the handoff snapshot packages manuscript 1.97 unchanged.

    Theory of Reality: computation route and a first identifiability result →Eight-clock execution: coupled Maxwell, recovery and identification limit →Theory of Reality mathematical and computational agent handoff →
  18. Manuscript 2.42 and research edition 02: 1,365 software tests pass, and none of the 57 engine records is admitted as a new measured physical prediction.

    Theory of Reality 2.42 →A framework learns to answer for its evidence. →Repetition cannot erase a shared calibration error. →
  19. Five edition-02 investigations keep their failed gates and name a first physical target: a forward Aharonov–Bohm apparatus prediction scored on newly reserved frames.

    Where the field ends, the loop remains. →A fitted expansion rate has to survive the next observer. →A sharper picture can still hide the same ambiguity. →Positive modes. A finite difference. An attractive force. →A question can begin in Scripture. Its prediction still needs a test. →

Dates identify recorded events; they do not automatically identify an invention or launch.

06 / OPEN A CONVERSATION

Bring a question.
Let’s make it inspectable.

For research discussion, technical collaboration or questions about a specific piece of work, contact Anthony directly.

hello@d0re.com ↗Contact & availability →

Originating framework and research direction:
Anthony Michael Collado.
Original source credits remain with each record.