Does independent corroboration (different author group, different venue) exist for Pinheiro, Ting et al.'s (2026) Babylonian Diaries dating preprint (arXiv:2610.34297), needed before a 4th claim-note from it can be written?
This capture targets question-corroborate-pinheiro-ting-2026-babylonian-diaries-preprint, the open question logged when the 2026-10-27 capture of this preprint hit the vault's single-source concentration cap (three claim-notes already resting on it: claim-babylonian-diaries-automated-dating-96-percent-top-rank-inclusion, claim-babylonian-diaries-dating-pipeline-parser-llm-mcp-astronomy, claim-babylonian-diaries-dating-system-conservative-commitment-rule). The search below checked (a) the arXiv record itself for a later version, journal-ref, or DOI, (b) general web search for citations, replications, or press coverage, and (c) the paper's own footnote naming a "related" University of Chicago project, following it to the primary source.
Claim: As of 2026-10-28, arXiv:2610.34297 remains a single-version, unrefereed preprint with no journal-ref, no registered DOI, and no citing paper located by search
verifies: question-corroborate-pinheiro-ting-2026-babylonian-diaries-preprint
Claim type: historical/bibliographic fact about the document's own publication record. Source: arXiv's own abstract page for the record — Tier 1 (the platform's own metadata about its own holding), clearing the floor comfortably even though this is adjacent to a quantitative/dating fact.
The arXiv abstract page shows exactly one version, submitted 25 Oct 2026, with "arXiv-issued DOI via DataCite (pending registration)" and no journal-ref field. General web searches for the paper's title, arXiv ID, and author names together turned up no news coverage, no citing paper, and no third-party evaluation — only the paper's own GitHub repository and interactive page (both self-hosted by co-author Yuan-Sen Ting, i.e. the same source, not independent) and unrelated general background on the Babylonian Astronomical Diaries. Given the preprint is three days old at the time of this search, this absence reads as consistent with simple recency, the same reading the vault has already recorded for other freshly-posted preprints (cf. claim-fic-preprint-no-independent-corroboration-as-of-2026-10-20).
source_quote: "Submission history From: Yuan-Sen Ting Assoc. Prof. [view email] [v1] Sun, 25 Oct 2026 13:52:56 UTC (1,958 KB)" source_url: https://arxiv.org/abs/2610.34297 source_sha: fc0c03dc0c436fafbb1d0c3f3c4cb75979b6aac7bf2c689b1e678ef925eb9fd3 source_title: "Automated, auditable dating of the Babylonian Astronomical Diaries with language models and astronomical tools" source_author: "arXiv.org (platform record page)" source_date: 2026-10-28 (page as fetched) source_venue: "arXiv.org abstract page for arXiv:2610.34297" source_tier: 1
Claim: A genuinely independent, differently-authored, differently-venued, and differently-funded project exists on a related problem — but it predates the Pinheiro/Ting preprint and neither cites nor tests its specific findings, so it is parallel work, not corroboration
verifies: question-corroborate-pinheiro-ting-2026-babylonian-diaries-preprint
Claim type: mixed historical (who/what/when of the UChicago project) and technical-mechanism (what the project's method is). Source: the University of Chicago Data Science Institute's own news page about its own funded researchers — Tier 1 for the institution's own report of its own award and researchers' own description of their method (a primary self-report, the same category as a lab's own announcement of its own work).
The preprint's own footnote (p.2) names "a related project at the University of Chicago [that] uses language models to extract the observations and a Bayesian model to check them against computed positions," linking to a University of Chicago Data Science Institute (DSI) news page. That page, dated 2026-10-14 — eleven days before Pinheiro et al.'s 25 Oct 2026 submission — reports that Aaron Schein (Assistant Professor of Statistics and Data Science) and Chris Wolfram (PhD student, Computer Science), both at UChicago, received a grant from the Humanities and AI Virtual Institute (HAVI, a program of the nonprofit Schmidt Sciences) "to develop a method for extracting and verifying astronomical observations from the Babylonian Astronomical Diaries," working with John Steele, a Brown University professor described as "a leading expert on Babylonian astronomy." Their method pairs LLM-extracted observations with "a Bayesian statistical model that compares each observation against where the Moon, planets, and stars actually were that night" — a different technique from Pinheiro et al.'s rule-based clause parser plus candidate-year ranking across all 701 possible years — and is aimed at a related but distinct goal: building a comprehensive observation database, estimating quantities like the ancient "cubit" unit and the timing of night observations, and (as a secondary application) potentially helping "date tablets whose dates are missing." The page reports the project's results as still early ("promising") and names no comparable dating-accuracy benchmark against ADART's published dates. Critically, it makes no mention of Pinheiro, Ting, or their preprint — unsurprising given it predates that preprint, but it means the UChicago project cannot be read as testing, replicating, or otherwise engaging with arXiv:2610.34297's specific claims (the 96% top-rank inclusion figure, the 10-of-417 commitment rate, or the cost comparison). It satisfies "different author group, different venue" but not the further requirement — implicit in what corroboration means — that the second source actually test or confirm the first's findings.
source_quote: "Schein has received a grant from the Humanities and AI Virtual Institute (HAVI), a program of Schmidt Sciences, to develop a method for extracting and verifying astronomical observations from the Babylonian Astronomical Diaries. The records, written in Akkadian on cuneiform tablets over roughly six centuries, are the product of the longest-running data collection project in human history and offer an unusually detailed window into the ancient world. Schein and Wolfram are working with John Steele, a professor at Brown University and a leading expert on Babylonian astronomy." source_url: https://datascience.uchicago.edu/news/havi-award-supports-uchicago-researchers-pairing-ai-with-bayesian-statistics-to-decipher-the-babylonian-astronomical-diaries/ source_sha: f0862bf746b0af8537da69c05abc82c31950b6782a5542215acba68cf131045f source_title: "HAVI Award Supports UChicago Researchers Pairing AI with Bayesian Statistics to Decipher the Babylonian Astronomical Diaries" source_author: "University of Chicago Data Science Institute (DSI News)" source_date: 2026-10-14 source_venue: "datascience.uchicago.edu (University of Chicago Data Science Institute news page)" source_tier: 1
source_quote_2: "Statistician Aaron Schein and Chris Wolfram, PhD student in Computer Science, use LLMs to extract information from the Babylonian Astronomical Diaries, verifying the LLM's output with a Bayesian Model that knows the correct positions of the Moon, planets, and stars in the ancient night sky." source_url_2: https://datascience.uchicago.edu/news/havi-award-supports-uchicago-researchers-pairing-ai-with-bayesian-statistics-to-decipher-the-babylonian-astronomical-diaries/ source_sha_2: f0862bf746b0af8537da69c05abc82c31950b6782a5542215acba68cf131045f source_tier: 1
Separately, the footnote's own quoted text from the preprint itself (naming the project, read directly from the PDF) corroborates that the preprint's authors see this as a parallel rather than overlapping effort:
source_quote_3: "A related project at the University of Chicago uses language models to extract the observations and a Bayesian model to check them against computed positions" source_url_3: https://arxiv.org/pdf/2610.34297 source_sha_3: e26e98d0cd36475273e588c265a2db43df7e0fb1efea8190f692f90ad96d05dd source_tier: 1
Claim: The single-source concentration cap on arXiv:2610.34297 remains undischarged as of 2026-10-28 — no fourth claim-note should yet be written from this preprint
verifies: question-corroborate-pinheiro-ting-2026-babylonian-diaries-preprint
This follows from the two claims above rather than from a new source: no different-author/different-venue source has tested, replicated, or independently evaluated any of arXiv:2610.34297's specific reported figures. The one candidate the paper itself points to — the UChicago Schein/Wolfram/Steele project — clears "different author group, different venue" but not the substantive bar, since it predates the preprint, uses an unrelated method for a different immediate goal, and never engages with Pinheiro et al.'s findings at all. Per sources.md's single-source concentration cap, the cap therefore stays in force: the cost-comparison figure (GPT-6-Luna ~$3 vs. GPT-5.6-Terra ~$60, ~20x) and the Claude-Opus-5-coding-agent parser-rule ablation — both routed to the open question rather than promoted on 2026-10-27 — should continue to wait, as should any other new finding from this preprint, until an independent source actually tests its claims. Worth re-checking later: whether the Schein/Wolfram project, once it publishes, tests or cites Pinheiro et al.'s benchmark figures directly (the two groups working on the same corpus makes mutual citation likely once either publishes further).
Further leads
- The Schein/Wolfram/Steele project (UChicago DSI, HAVI/Schmidt Sciences grant, announced 2026-10-14) has already reportedly produced "a precise estimate of the length of the cubit" the Diarists used and "independently rediscovered known errors in an older chronology used to convert Babylonian dates to the modern calendar" — likely referring to the same Parker & Dubberstein (1956) calendar table the Pinheiro/Ting paper's own candidate-year space rests on (not confirmed by direct comparison this session). — source: datascience.uchicago.edu HAVI award page, as above.
- HAVI (Humanities and AI Virtual Institute), a Schmidt Sciences program, has awarded "$10 million to 19 research teams" in this round, bringing its total commitment to "more than $21 million over two years" — a quantitative figure about AI-humanities funding broader than this one project, worth its own note if the vault builds out AI-funding-landscape coverage. — source: datascience.uchicago.edu HAVI award page, as above.
- Pinheiro/Ting's own interactive page and GitHub repository (tingyuansen/Babylonian_Dating) are same-author artifacts, not independent sources, but could be worth a direct read for implementation detail beyond what the paper's prose covers. — source: arXiv:2610.34297 abstract page, "this https URL" links.
Entity candidates
- John Steele — person — professor at Brown University, described by UChicago's own announcement as "a leading expert on Babylonian astronomy"; the FOUNDATIONAL domain authority the new (Schein/Wolfram) project leans on for its archaeological/historical grounding, analogous to how Pinheiro et al.'s paper leans on Sachs & Hunger's ADART — flag first, per the known blind spot, since he is the senior figure the newer AI-assisted project measures its own credibility against.
- Aaron Schein — person — Assistant Professor of Statistics and Data Science, University of Chicago; leads the independent, differently-methoded Bayesian-verification project named in Pinheiro et al.'s own footnote.
- Chris Wolfram — person — PhD student, Computer Science, University of Chicago; co-developer of the Schein project.
- Humanities and AI Virtual Institute (HAVI) — concept/program — a Schmidt Sciences initiative funding 19 AI-and-humanities research teams (~$21M over two years); the funding vehicle behind the one candidate "independent" project this search found.
- Schmidt Sciences — concept/organization — nonprofit founded by Eric and Wendy Schmidt, funds HAVI; worth a mention if the vault tracks AI-philanthropy funders.
- Parker & Dubberstein (1956) — concept/historical reference (already flagged in the 2026-10-27 capture) — the "older chronology" the Schein/Wolfram project may have found errors in; worth checking directly if that project publishes specifics.