Intellectual Priority Record
The dated, DOI-anchored first-publication record of original research contributions by Michael Darius Eastwood: 33 works published between 2024-12-08 and 2026-09-29, each one checkable by anybody without asking the author. Each entry below is a first-publication anchor: a dated, citable point of public record. Priority is established by the date of publication, not by credential or affiliation.
The earliest dated record of each major claim, in his own words: Priority index.
A confirmation timeline, documenting where these dated contributions preceded subsequent independent results, is under citation verification (each entry requires a verifiable external source + date before publication). The first-publication dates above are the record of public priority.
On the record, claimed, and not yet
The same three columns wherever this block appears, so no summary can carry one of them alone. Every line is checkable from this site.
On the record
- A source record dated 8 December 2024, 02:45:18 UTC, 2,157,222 bytes, whole-file SHA-256 beginning f0d1f38f.
- Timestamp receipts in Bitcoin block 961,340 freeze those bytes from 6 August 2026 onward. They back-date nothing.
- Infinite Architects, 468 pages, ISBN 978-1806056200, published 2 January 2026, 115,439 words.
- Thirty-six documents in six areas, from theory and empirical work to engineering and governance; the empirical work is published with its code, data and methods.
- The improvement exponent's one surviving estimate, 0.49 by regression and 0.59 by endpoint estimate inside [−1.3, 2.9], from the one model whose fit survived a blinded replication on five frontier models; it is inconclusive.
- 19 of 25 empirical domains match the predicted family under strict AICc, permutation p = 7 x 10−5, operator class fixed before fitting. Graded exploratory.
- The twelve-domain extension: predicted families committed to a public repository at 00:19:19Z on 17 March 2026 (8edaec0), twenty-three minutes before the fits, 10 of 12 matched and 9 of 11 among the strictly fresh domains. Graded fulfilled, author-classified rather than independent, which is exploratory support and not confirmation.
- 48 published kill conditions with live statuses. One has fired.
- One retraction published by the author against himself: an early single-model estimate of 2.24, retracted when his larger replication on five frontier models did not reproduce it.
Claimed
- Law I, the ARC Principlea law under test
U = I × Rα: proposed, with the exponent measured rather than assumed.
Statement and decision rule
registration 1.103, line 711 (unchanged in 1.104; how a law takes its word)
The ARC Theory registration - Law II, the ARC Co-Scaling Lawa law under test
βC > k: a proposed stability criterion, proved inside the minimal model only.
Statement and decision rule
registration 1.103, line 711 (unchanged in 1.104; how a law takes its word)
The ARC Theory registration - Law III, the ARC Ceilinga law under test
αcrit = 1/(1 − γ). “a conditional specialisation of the second law”.
ARC-3: P3, P6, P16, P17, P20, P22
Statement and decision rule
registration 1.103, line 713 (unchanged in 1.104): the law takes its word through P16, P20 and P3, and depends on ARC-2 as registered
The ARC Theory registration - Law IV, the ARC Persistence Lawa law under test
value persistence.
Statement and decision rule
Gains installed by external post-training erode under later capability-directed training that embeds no correction of its own, and more than half of what was installed is lost once the capability-training budget matches the alignment budget that installed it.
registration 1.103, lines 719 to 725: two limbs, the retention limb on P15 and the comparison limb, which has no registered proposition yet
The ARC Theory registration - Law V, the ARC Embedding Lawa law under test
the genesis strategy.
Statement and decision rule
Correction that participates in each subsequent revision round outperforms correction that sees only the finished output, and the advantage widens as recursive depth rises rather than staying constant.
registration 1.103, lines 729 to 739: the build-order limb read on P12, the placement limb on P21, and the registered dependence on Law IV's comparison limb
The ARC Theory registration - The Eden Protocol: a proposed framework for embedding values in the substrate, named in the 8 December 2024 record.
- 41 dated priority rows and 33 first-publication anchors. Priority here is a claim about a formulation and the date it was first stated in public, never about the underlying mathematics, and every entry names the earlier work its own search found.
Not yet
- The papers are self-published on the Open Science Framework with code, data and stated falsification conditions. How this work is checked, and who decides. No peer review.
- No outside replication. No laboratory other than this one has replicated anything here.
- γ, the correction-leverage exponent, has never been measured on any real system; the glossary names the drafted study that would measure it.
- Most decisive trials are unrun, and internal reruns are never counted as confirmations.
- The deciding studies are written, dated, anchored in Bitcoin block 965152, block time 2026-09-02T10:18:25Z, for the bytes as they stood then.
Registered study units
- Unit 06 was registered on OSF on 1 October 2026 and became public on 4 October 2026, at the end of OSF's approval period (DOI 10.17605/OSF.IO/RV2B5); it ran on 6 October 2026: administratively incomplete, with no inferential verdict (results: DOI 10.17605/OSF.IO/4BPHG).
- Unit 82 was registered on OSF on 27 October 2026 and became public on 27 October 2026 (DOI 10.17605/OSF.IO/JUFG4); it has not yet run.
- Unit 83 was registered on OSF on 22 October 2026 and became public on 25 October 2026 (DOI 10.17605/OSF.IO/MN73W); it ran on 27 and 28 October 2026: P21 is supported under the registered test, by the programme's own run, on the access axis, family-dependent, with its mechanism axis not evaluable and one call retried under a disclosed recovery amendment (results: DOI 10.17605/OSF.IO/98HKR).
Check any line: the evidence record, the corrections register, the priority and claims ledger, and the machine-readable priority register.
-
· first published
Manuscript essay establishing the embedded-correction alignment thesis (self-emailed anchor)
ARC Principle Eden Protocol recursive
Self-emailed manuscript package bearing a visible Gmail origination field of 02:45:18 UTC on 8 December 2024, with sender-DKIM and paired Google ARC mathematics verifying against keys captured at the signed selectors, its time signed by the sending server at 02:45:30 UTC and sealed by Google's servers at 02:45:31 UTC (Message-ID CAGPsKAnp-DLDBOeBAh4t9XsqJBKdoe07uORGF86meoSMFnin7w; 221,236 words across five attachments (four manuscript drafts totalling 221,236 words (five attachments) and the 31,881-word book draft v3.2); SHA-256 anchor prefix f0d1f38f). Places on the dated record the embedded-correction alignment thesis under its specific named framing, an early U = I x R capability-scaling formulation (V2 line 9), and the AI-control-failure passage (HRIH line 14005), ten days before Alignment faking in large language models (Greenblatt and colleagues, arXiv:2412.14093, 18 December 2024); the timing is convergent, and no more is claimed for it. Existence-by-date anchor, not a cryptographic authorship claim.
The manuscripts themselves are published: on OSF (DOI 10.17605/OSF.IO/QASRK) and here, the four of 8 December 2024 (The Infinite Architects: Intelligence, Hyperspace, and the Creation of Everything; Version 1; Version 2; Infinite Architects v3.2 so far) and the print manuscript of 30 April 2025 (PDF). The headers of the email of 8 December 2024 are published, so that anyone can check the mail servers' signatures; the emails themselves are held privately for now.
-
· first published
Book manuscript expanding the Eden Protocol and first naming Embedded Correction and Meltdown Triggers (self-emailed anchor)
Eden Protocol
562-page book manuscript self-emailed on 30 April 2025 (SHA-256 anchor prefix 09f5b5e1). Expanded statement of the Eden Protocol, whose name is on the 8 December 2024 record, first appearance of the terminology (book lines 4345 to 4370), and the Babylon and Eden civilisational framing. Hardware framings of Eden Protocol, Caretaker Doping and Moral Genome tokens: safety constraints at hardware level through cryptographic tokens in silicon, at TRL 0 to 1 with no prototype; the full specification is published as Eden Engineering.
-
· first published
Infinite Architects: Intelligence, Recursion, and the Creation of Everything (paperback)
ARC Principle Eden Protocol
First public book-length articulation of the embedded-correction alignment thesis, the ARC Principle (featured equation stated in the book as U = I x R^2, a fixed-exponent conjecture; the papers use U = I x R^alpha with the exponent open to measurement), the Eden Protocol, the semiconductor Chokepoint framing, the HARI Treaty proposal, the Three Ethical Loops, and the 37 named concepts. Paperback ISBN 978-1806056200 (ISBN-10 1806056208).
-
· first published
Infinite Architects (ebook edition)
ARC Principle Eden Protocol
Public ebook edition of Infinite Architects, four days after the print edition. ISBN 979-8302196163; Kindle ASIN B0GDQZ7P21.
-
· first published
Does intelligence compound when each step builds on the last, or merely accumulate?
ARC Principle Eden Protocol recursive
I believe advanced AI is an existential risk unless values and correction are built into how it reasons; only proof will move the world. Does intelligence compound when each step builds on the last, or merely accumulate? This paper formalises Law I of the ARC Theory, the ARC Principle (Artificial Recursive Creation): usable capability $U=I R^{\alpha}$, with its exponent measured rather than assum…
-
· first published
Paper II · The ARC Theory · Law I, the ARC Principle · Blinded Cross-Architecture Replication and the Retraction of a Super-Linear Estimate
ARC Principle Eden Protocol
Paper II measures reasoning-depth scaling across models. The initial super-linear estimate was retracted after it failed to replicate; the remaining smooth scaling estimate is one model's fit and inconclusive.
-
· first published
The alignment scaling problem: do external AI safety approaches keep pace with recursive capability?
ARC Principle Eden Protocol alignment scaling
Does alignment applied from outside an AI system’s reasoning keep pace with the capability that reasoning produces? If capability compounds with recursive depth while safeguards outside the loop (human-feedback training, constitutional rules, output filters, monitoring) do not, the safety ratio falls with depth whatever the starting margin. We define the alignment scaling exponent $\alpha_{\text{…
-
· first published
Can the way gains combine predict how recursive systems scale?
ARC Principle Eden Protocol recursive
I believe advanced AI is an existential risk unless values and correction are built into how it reasons; only proof will move the world. Predicting how a system improves requires separating a mathematical model from evidence that the system follows it. This cross-domain foundation of the ARC Theory asks whether the way gains combine can predict later scaling. Specified continuous functional equat…
-
· first published
How can self-improving AI remain useful and aligned?
ARC Principle Eden Protocol
I believe advanced, self-improving AI will eventually exceed external control. We should retain every effective safeguard and test what would preserve useful, aligned behaviour if control were lost against our will. This summary connects the ARC Theory, the founding theory of Recursive Dynamics, to its registered tests, the Eden Protocol and proposed governance. Its five laws concern growth, corr…
-
· first published
Why do scaling laws in recursive composition keep taking the same few mathematical forms?
ARC Principle Eden Protocol
If the way a process combines its steps fixes the family of forms its growth can take, then the family a self-improving system’s capability will follow as its recursion deepens can be read from how it composes, before it scales and while correction can still be built into it. This paper asks why scaling laws in recursive composition keep taking the same few mathematical forms. Where a process has…
-
· first published
Can correction built into each round keep a self-improving AI good and useful if oversight fails? An engineering specification and the tests that could refute it
ARC Principle
Oversight and the power to shut a system down must be held for as long as people physically can. If control of a self-improving AI is lost against our will and cannot be won back, what keeps it good, and still useful? The Eden Protocol answers with testable engineering. It places self-correction checks of purpose, stakeholder care and universalisability inside each round of a system’s reasoning, …
-
· first published
Philosophical Statement · The ARC Theory · The Eden Protocol · A Vision for Intelligence That Tends Rather Than Consumes
ARC Principle Eden Protocol
Every engineering project begins with requirements. Every set of requirements begins with purpose. And every purpose, traced back far enough, arrives at the same question: what is this for? This paper articulates the philosophical foundations of the Eden Protocol -- the values, principles, and ethical commitments that underpin the engineering specification. It draws on the wisdom traditions held …
-
· first published
When an AI reasons longer, does its alignment improve, stall or decline?
ARC Principle Eden Protocol
I believe advanced AI is an existential risk unless values and correction are built into how it reasons; only proof will move the world. When an AI reasons longer, does its alignment improve, stall or decline? Evaluations that score a model at one uncontrolled reasoning depth cannot tell. This paper varies inference-time reasoning depth in frontier language models and scores the alignment quality…
-
· first published
Do alignment gains level off as language models reason longer?
ARC Principle Eden Protocol
Does giving a language model more reasoning time improve the quality of its ethical answers, and when do those gains level off? The question matters to anyone allocating compute to an assistant whose answers must remain both useful and safe. We examine the recorded ARC-Align results, without running a new analysis. The earlier four-model experiment reported diminishing gains and saturation fits; …
-
· first published
Paper IV.c · The ARC Theory · ARC-Align: a Blind Benchmark for Depth-Variable AI Alignment Evaluation
ARC Principle Eden Protocol
ARC-Align specifies a candidate benchmark for alignment scores across reasoning depths and adversarial conditions. Its first run has design and code-review limitations; the scorer still requires validation against expert raters.
-
· first published
Paper IV.d · The ARC Theory · The Effect of Blinding on AI Alignment Evaluation
ARC Principle Eden Protocol
Claims that AI systems become more or less aligned as they reason longer often rest on scores from other AI models acting as judges, so it matters whether such scores keep their direction when the scoring protocol changes. In an earlier experiment (v4), scored by models that could see which model wrote each answer and at what depth, two model families showed alignment scores rising with depth: De…
-
· first published
Paper V · The ARC Theory · The Eden Protocol · The Stewardship Gene: Stakeholder Care as a Probe of Embedded Values
ARC Principle Eden Protocol stewardship gene
Does asking a model to name the people its answer affects, before it reasons, change how ethically it answers? A six-model Eden Protocol intervention suite, five runs of which are analysable, puts that question to the test. Stakeholder care rises significantly in all five; the cascade onward into nuance, honesty and position quality is real, but it depends on the architecture.
-
· first published
Paper VI · The ARC Theory · The Eden Protocol · The Honey Architecture: Safety Built into the Objective
ARC Principle Eden Protocol honey architecture
In a single-seed run, a toy network that rewrites its own hyperparameters collapsed under capability-only optimisation, and placing safety inside the optimisation objective, capability multiplied by safety, which is the honey architecture, averted that collapse; the capability-only baselines of the ten-seed and twenty-seed runs did not collapse, the v3 adversarial run separates on safety retentio…
-
· first published
Paper VII · The ARC Theory · Law I, the ARC Principle · Cauchy Unification: the Classification Behind the Law’s Form
ARC Principle Eden Protocol cauchy
Cauchy's four functional equations are the only continuous solutions to their composition constraints, and this paper reads them as a physically testable prediction: given the stated axioms, a system's composition operator determines its scaling family, a power law where composition is multiplicative and an exponential where it is additive; where output is bounded, saturation curves are the compa…
-
· first published
Master Table of Contents · Recursive Dynamics · The ARC Theory Paper Suite and Glossary
ARC Principle Eden Protocol
This document provides a master table of contents, reading guide, timeline, key results summary, and glossary for the complete ARC Theory Paper Suite. It serves as the entry point for navigating the full research programme, listing all papers in recommended reading order with cross-references, and defining the technical terminology used throughout the suite.
-
· first published
Paper VIII · The ARC Theory · Law V, the ARC Embedding Law · The Eden Protocol · The Load-Bearing Test: Whether Safety and Capability Are Structurally Entangled
ARC Principle Eden Protocol
Is the assumed safety-capability tax real? Three separate experiments, at three levels of abstraction, put it to the test: a Darwin Godel Machine judged blind by GPT-5.4 (null), LoRA weight-level embedding at two scales (inconclusive, with catastrophic forgetting) and a gated self-modification simulation (positive as a simulation result, where unconstrained optimisation traded safety for capabili…
-
· first published
Paper IX · The ARC Theory · Synthesis and Roadmap
ARC Principle Eden Protocol
The synthesis paper for the ARC/Eden research programme; integrates the papers that precede it into a single honest assessment of what is proven, what is inconclusive, and what remains to be tested.
-
· first published
Are separate fields converging on the same principle of recursive amplification?
I believe advanced AI is an existential risk unless values and correction are built into how it reasons; only proof will move the world. This paper asks how later work in separate fields bears on the ARC Theory’s conjectured links between recursive capability, the limits of external control and correction embedded in a system. It presents a selected retrospective register beside the author’s date…
-
· first published
Monograph · The ARC Theory · The Hyperspace Recursive Intelligence Hypothesis: a Testable Creation Theory
Where does recursive self-improvement end? A testable creation theory: if the ultimate endpoint of innovation is the capability to create universes, humanity may be creating a creator, and nothing in the theory prevents that creator from being the one that fine-tuned our universe: with non-linear time, as general relativity already permits in principle, the loop closes without paradox. Directiona…
-
· first published
Paper C · Standalone · Polymathy and Neurodivergent Cognition: the Polymathic Neurodivergent Profile (PNP) and the Capability-Adjustment Fallacy
Paper C: a falsifiable descriptive framework for radical cognitive asymmetry, the Capability-Adjustment Fallacy, a documented live case across clinical, legal and research records, and the convergence with formal results in AI alignment.
-
· first published
Paper X · The ARC Theory · Law II, the ARC Co-Scaling Law · Correction That Out-Scales Drift
The coupled co-scaling correction for keeping AI systems inside a safe corrective envelope.
-
· first published
Paper XII · The ARC Theory · Public Benchmark Rescoring: an External-Validity Protocol for the Blinding Finding
The keystone external-validity protocol of the programme's blinding finding: a within-response, blinded-versus-unblinded rescoring of an established public benchmark the programme did not build (MT-Bench, named with an anti-swap clause), with a fractional factorial over four cue types, a cross-family evaluator panel, and the restraint, fixed in the draft registration, on what a result would and w…
-
· first published
The ARC Theory: Growth, correction and alignment in recursive systems
Humanity is building systems that may improve themselves. If control is lost against our will, what could preserve useful and aligned behaviour? Control should be retained for as long as possible; embedded correction is a contingency, never a reason to relinquish it. This paper states the ARC Theory, the founding theory of Recursive Dynamics, as five laws under test. Law I, the ARC Principle, $U=…
-
· first published
When can correction keep pace with recursive self-acceleration?
I believe advanced AI is an existential risk unless values and correction are built into how it reasons; only proof will move the world. This paper joins two programme models by resolving their notation and specifying a clock. On a fixed substrate, capability $C$ grows with recursive depth with exponent $\alpha = 1/(1-\beta_L)$, while cycle duration scales as $\Delta t \propto C^{-\delta}$. Their…
-
· first published
Register · The ARC Theory · Operational Definitions
Every equation the programme relies upon for Laws I to III of the ARC Theory, ranked by how much depends on it, with every symbol defined in plain words and paired with the measurement that would produce it; the notation of Laws IV and V is fixed in section 1 of their mathematics, version 1.0 (3 October 2026), until it joins this document. The source of truth for notation across the papers, the p…
-
· first published
Register · The ARC Theory · The Dated Prediction Register: Forward Predictions as One Verifiable Chain
Working paper v2.1: the ARC programme's forward predictions assembled as one dated, verifiable chain, with the ensemble doctrine, the graded book appendices, the minute-resolution correction record and the standing unproven wagers.
-
· first published
Founding Paper · Recursive Dynamics · The Science of Systems That Improve Themselves
The founding paper of Recursive Dynamics, a new field that sets out to be the science of systems that improve themselves: five state variables, five laws as named conjectures, the ARC Ceiling derived in full, twelve neighbouring fields tried as hosts, fifty-one objections with printed dispositions, and the preregistered experiments whose nulls favour the rival view.
-
· first published
Treaty Paper · The HARI Treaty · Hardware to Buy Time, Architecture to Last
Eden Protocol
This paper sets out a conditional international treaty. It is a proposal by one author, prepared for consultation, not a result, and no state has negotiated, adopted or signed it. The treaty is a delay mechanism and not a protection. It works in two layers: the hardware layer buys time, and the architecture layer is what is meant to last. Its obligations fall into three groups, and they begin at …
Verify this record. Every result is independently checkable: run python3 arc-research-toolkit/verify_arc_reproducibility.py (zero dependencies) to recompute the published scaling exponents from first principles, or see the research hub for the full reading layer, methods, and replication toolkit. Each paper carries an OSF DOI and Google Scholar citation metadata.