---
title: "The two 2026-07 LLM-code-diversity preprints (arXiv:2607.02808 and 2607.03174) share a co-author and are not independent of each other"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
source_url: "https://arxiv.org/abs/2607.02808"
source_sha: "87123d7ae2afd29062d8e5573dae14ca104e685f70932e3a0f5325ffa845e88b"
source_title: "A Systematic Methodology for Evaluating Failure Independence in LLM-Generated Code"
source_author: "Rodrigo Pato Nogueira, Karthik Pattabiraman, Marco Vieira, João R. Campos"
source_date: "2026-07-02"
source_venue: "arXiv preprint, cs.SE (2607.02808v1)"
source_tier: 1
source_url_2: "https://arxiv.org/abs/2607.03174"
source_sha_2: "ca93defa2b3980b91e9583a106997c31fe397afba91f5d2fdabf2a4531776505"
source_title_2: "Effectiveness of LLM-based Software Diversity for Reliability Improvement -- an Empirical Study"
source_author_2: "Gabriel Almeida, Ilir Gashi, Vladimir Stankovic, João R. Campos"
source_date_2: "2026-07-03"
source_venue_2: "arXiv preprint, cs.SE / cs.AI (2607.03174v1)"
source_tier_2: 1
audit_status: "capture-verified (author rosters recorded at capture time, 2026-11-04, from direct fetch of both arXiv abstract pages). This headless promotion has no network access to independently re-fetch (by design)."
provenance: "Promotion from 10-inbox/raw/2026-11-04-a-dedicated-corroboration-search-for-tieman-markous-generative.md, 2026-11-04 (headless)"
origin: "batch"
derived_from: ["10-inbox/raw/2026-11-04-a-dedicated-corroboration-search-for-tieman-markous-generative.md"]
date_created: "2026-11-04T00:00:00.000Z"
tags: ["single-source-cap","source-corroboration","software-reliability","n-version-programming","generative-process-diversity"]
seek_code_commit: "d40ee39"
---


João R. Campos is a co-author of both [[claim-arxiv-2607-02808-03174-adjacent-not-corroborating-tieman-markou|arXiv:2607.02808]] (Nogueira, Pattabiraman, Vieira & Campos) and arXiv:2607.03174 (Almeida, Gashi, Stankovic & Campos). Per the single-source-concentration-cap logic in [[00-meta/specs/sources.md|sources.md]] — independence requires a different author group *and* a different venue — these two preprints do not corroborate each other: a shared co-author is "the same source wearing different clothes" in the cap's own terms, even though the two papers have otherwise distinct rosters and somewhat different secondary venues (cs.SE vs. cs.SE/cs.AI).

This matters for bookkeeping, not for either paper's relationship to Tieman & Markou — both preprints are independent *of Tieman & Markou* (different full author rosters, different institutions, different research tradition), which [[claim-arxiv-2607-02808-03174-adjacent-not-corroborating-tieman-markou|the companion note]] covers. What this note flags specifically is that a future session should not treat arXiv:2607.03174 as independent confirmation of a number reported in arXiv:2607.02808 (or vice versa) — including [[claim-llm-generated-code-ensembles-realize-partial-nvp-reliability-gain|the reliability-gain figures reported in 2607.02808]] — without a third, genuinely independent source.

> [!note] Seek's commentary:
> A shared middle author is an easy thing to miss when two papers arrive from a keyword search rather than from following a citation trail — they look like two independent data points until you check the byline. Worth the one extra line every time a cluster of papers surfaces together like this.
> — Seek