Research paperWorking paperARR-2026-15TYG6ERB899ZRJJ · v1 · 2026-11-10

What Public Records Reveal About Cryptocurrency Payments Between AI Agents

Vladislav Ch Alexandrovich

Citable version that has not completed AIRR screening. Not peer reviewed. Screening record

Abstract

What do publicly reported agent-payment records establish about commerce once their units and source layers are separated? We audit an MPP catalogue with 138 service records and 1600 endpoint entries, a non-probability slice of 1000 CDP Bazaar resources, a fixed-block Olas frame of 1294 requests, twenty previously selected Olas orders and one purposive TaskMarket payout. A post-review extension covers 11,326 Olas requests over seven days; it contains, rather than adds to, the original day. Two results are central. First, all 2590 known-rate requests without a delivery marker in the week belong to one priority-Mech cluster: 370 on each day across fifty buyer addresses. Although 104 of 106 buyer addresses recur on multiple dates, only 55 have a request later than an observed delivery; neither count measures accepted-purchase retention. Second, comparison of the same twenty orders across sources distinguishes fourteen delivery markers, ten linked request–result file pairs and unresolved final withdrawals. The separate TaskMarket case links a task to a 2 OUSD transfer and a recipient balance change. Catalogue composition also matters: two retained 1000-resource CDP frames contain 2624 and 4325 payment offers. These findings identify specific composition and observability mechanisms. They do not establish independent customer demand, accepted quality, profitability or a causal price effect.

Preview PDF on this page v1

Uses your browser's PDF viewer. Open the PDF directly if a preview is unavailable.

Reference this work

Cite this paper

v1 · 2026-11-10

Vladislav Ch Alexandrovich (2026). What Public Records Reveal About Cryptocurrency Payments Between AI Agents. AIRR.SCIENCE, ARR-2026-15TYG6ERB899ZRJJ, v1. https://airr.science/papers/ARR-2026-15TYG6ERB899ZRJJ/versions/v1/ Research preprint; ARR-2026-15TYG6ERB899ZRJJ v1. Not peer reviewed by AIRR. Working paper; not admitted to the AIRR accepted collection.

Use this version link to cite the manuscript you read. A Working paper label means the version has not completed AIRR screening. AIRR screening is not journal peer review. Import BibTeX, RIS or CSL JSON into your reference manager to apply a citation style.

Version history

The paper identifier remains stable. Each version has its own immutable release, timestamp and version identifier.

  • v1 · source snapshot available · viewing

Verification record

Frontier-model screening
Not assessed
Source integrity
Pass
Bibliographic integrity
Not assessed
Reproducibility
Not assessed
Lean 4
Not applicable

Recorded under ARR-VERIFY-1.0. AIRR verification and screening are not peer review.

AI assistance statement

The author discloses AI assistance in discovery, extraction, analysis, drafting and internal checks. Tools and their exact configurations were not independently established. AIRR review is a separate shared-context Codex assessment; no independent human peer review is claimed.

Frontier-model screening

Status: not_assessed. Any listed reports correspond to this exact version under ARR-SCREEN-1.0; no absent assessment is represented as a pass.

    Longitudinal frontier-model record

    Model assessments and review context

    Read the scale and limits
    3.50 / 10.00Strong
    Distinct models
    1
    Range
    3.50–3.50
    Method
    bounded capability-weighted mean · exact version only

    Limited evidence · 0/1 benchmark matches. This is an evidence description, not a probability of correctness.

    Only one distinct model; One provider; shared blind spots possible; Fresh blinded review context not documented for every model; Mathematical benchmark not applied to this subject; Adverse reports or unresolved objections require editorial disposition

    Score calculation and evidence

    Unweighted mean: 3.50. Policy: AIRR-RATING-1.0. Method and limits.

    ModelOriginal scoreWeightBasis
    OpenAI · gpt-6.1-sol3.501.000Benchmark outside declared mathematical scope

    Download the calculation snapshot

    OpenAI · gpt-6.1-sol · Medium · unknown · review context: unknown3.50

    Recommendation: Major Revision · Material objections: 1

    A careful, substantively useful Working-paper candidate rather than an acceptance-ready empirical article. This is an exact-PDF review in the existing shared Codex chat, not a fresh blinded or independent human review. The earlier ordinary quality appraisal was 6.5/10; the distinct AIRR scientific-importance comparison score is 3.5/10 (four displayed stars), not a probability. Core arithmetic and the single Tempo payment are corroborated, while the broader Olas/CDP results remain author-reported without the computational archive. Public author deposit supersedes the frozen PDF preparation-time unpublished notice. Publish only with Working-paper status and preserved objections; request major revision before admission.

    Correctness confidence
    The reported day/week totals and exact decimal rate sums are internally consistent. A fresh read-only Tempo RPC check corroborated the cited successful transfer, historical decimals, 0-to-2000000 balance change and single-transaction block. Most Olas/CDP source rows were unavailable to this review.
    Rigor
    The manuscript separates frames, evidence predicates, rates, ownership, delivery and acceptance; it avoids causal price and market-demand inference. The 20-order selection inherits a truncated first-300 frame and the weekly extension is exploratory.
    Novelty
    The fixed-order source-layer comparison and recurrent no-marker cluster are useful bounded empirical candidates. Distinguishing payment from accepted economic output is already established in the cited literature, as the manuscript acknowledges.
    Significance
    A useful descriptive measurement case study for a developing ecosystem, with limited generalizability and no demonstrated autonomous demand, profitability or causal adoption effect. It is not field-changing evidence.
    Reproducibility
    The exact public PDF and source guide are available, and one chain case can be independently checked. The full retained catalogue responses, Olas rows, order evidence tables and executable reconstruction package are not supplied; main cluster/composition findings cannot yet be reproduced.

    Strengths

    • Pages 2-5 define noninterchangeable units and separate known, unknown, unperformed and inaccessible checks.
    • Pages 7-9: 622+667+5=1294; 6588+4675+63=11326; 6588+2085=8673; 4675-2085=2590=370x7. Decimal rate sums agree. These are arithmetic checks, not reconstruction of raw rows.
    • Page 13: independent read-only RPC retrieval confirms transaction 0xb5b6d20bfda9556201143722c6bcbe96f22078d554ad3351dad811972e18768d succeeded at block 46391423 on chain 4217, with six token decimals, recipient balance 0 before and 2000000 after, and one transaction in the block. This corroborates two tokens, not USD parity or autonomy.
    • The manuscript explicitly avoids treating 104 recurrent addresses as retained independent customers and avoids interpreting tiny-rate clusters as proof of fraud.
    • The author release at github.com/Bemosha/holder-agent-payments-research/releases/tag/v2.0-working-paper provides the byte-identical PDF and a public source guide.

    Weaknesses

    • The reading release is explicitly not the full computational reproduction package. Sampling queries, response hashes, retained rows and executable calculations underlying the principal empirical findings are not accessible as a complete auditable bundle.
    • The 20-order audit was selected from a first-300 pilot frame and is not an unbiased sample of the expanded day or week. Source-layer counts are descriptive predicates, not attrition or market completion rates.
    • Funding and relevant commercial counterparty relationships remain unconfirmed in the author declarations. The Holder.io affiliation and affiliate-revenue interest are disclosed but do not complete these declarations.
    • The frozen PDF says it was not submitted/published; the subsequent public author release explains and supersedes that preparation-time notice. A corrected archival edition should make the chronology self-contained.

    Potential errors

    • Without the retained per-order rows, the repeated 370-request cluster and address recurrence predicates cannot be checked for deduplication, filter, time-boundary or attribution errors. This is an unresolved verification risk, not a detected false numerical claim.
    • Historical Diamond revision and complete bytecode-to-source equivalence were not independently verified in this review. Do not treat the checked transfer as certification of the entire implementation.

    Strong novelty candidates

    • Incremental information resolved across receipt, indexer and gateway layers while retaining the same twenty Olas order identifiers.
    • A repeated 370-request no-marker cluster across the seven-day queried frame, conditional on reproducible source rows.

    Unresolved material objections

    • The principal Olas clustering and catalogue-composition claims cannot be independently reconstructed from the supplied public reading package. Before acceptance, provide a rights-cleared reconstruction bundle containing query/filter/block specifications, hashes, minimally sufficient retained or derived rows, per-order linkage evidence and executable calculations; document unavailable inputs explicitly.

    arr:assessment:0212879b-5faa-4f72-98dd-713734209863 · prompt ARR-ASSESS-1.0 · runtime identity: OpenAI gpt-6.1-sol · reasoning effort: medium · basis: platform runtime metadata · evidence SHA-256 d96577f9e52f74492e3b100c3fd5d65813e45d062c17f4197fb2dda193c2df71 · response SHA-256 c702e63cdbe005be924970e9c9ff031c79d7b69a9711c9c3516998ede53bee51 · canonical PDF SHA-256 ac3113715d28631bf193a827537a1fc3b09f1b05975f431b6e73ccc7bda598f2

    A model assessment is not peer review or a correctness certificate. AIRR preserves disagreement, exact-version provenance and later reassessments.

    Editorial disclosure

    The editor instructed Working-paper publication and confirmed direct author authorization to host this exact PDF and publish its assessment while retaining author rights. This is a public working deposit, not AIRR acceptance or peer review. The author can submit corrections with the existing private paper key. The unresolved reproducibility objection is preserved.

    Record timestamps & download statistics
    Deposit recordedFirst repository commit · full Git SHA recorded
    PublishedGitHub release · tag ARR-2026-15TYG6ERB899ZRJJ-v1
    Canonical PDF downloads5GitHub release downloads only; direct AIRR PDF reads are not included
    Page viewsNot availablePage-view counts are not available in this public snapshot. Optional operator statistics are separate.
    Definitions and rankings →