AEO Vision — top arXiv papers for GEO

register entry aeo-vision-top-arxiv-papers-geo-aio-20261029 OBS-b1f1fee164cd · file as first run sha256 6dd8339224177675…

• T1-1 frame: 'The top arXiv papers and studies shaping Generative Engine Optimization (GEO) and Answer Engine Optimization (AEO) include foundational research and critical surveys'
• T1-2 the GEO paper: 'GEO: Generative Engine Optimization by Aggarwal et al. (Princeton University, Georgia Tech, Allen Institute for AI, and IIT Delhi) — The seminal 2023/2024 arXiv paper that coined the term GEO, introducing GEO-bench and demonstrating optimization strategies that raise content visibility in AI search systems.' (four markers, one of them the GeoReady link's token)
• T1-3 the identifier: 'Read the study on arXiv:2411.xxxxx or via industry summaries at GeoReady'
• T1-4 the survey: 'Optimizing Visibility in Generative Engines: A Critical Survey of Generative Engine Optimization (2023–2026) by Olivier Martinez (arXiv:2607.14035) — A comprehensive July 2026 critical survey reviewing 45 GEO studies, analyzing multistage pipelines (crawling, retrieval, reranking, and attribution), and evaluating the limits of citation rewrites.'
• T1-5 the verifiability paper: 'Evaluating Verifiability in Generative Search Engines by Liu, Zhang, and Liang (Stanford University) — An empirical human audit evaluating how accurately generative search engines cite and verify sources (showing average verifiability rates around 51.5%).' (two markers, the second the GeoReady link's token)
• T1-6 RAG: 'Retrieval-Augmented Generation for Large Language Models by Lewis, Perez, et al. — A foundational NeurIPS paper examining how LLMs retrieve and ground responses from external knowledge sources.' (one marker, the token of the GEO line's fourth marker)
• T1-7 the pointer: 'You can explore a curated industry breakdown of top papers via the AEO Vision Article on Top arXiv Papers.'
• T1-8 closer: 'Would you like a summary of the key optimization strategies found in the original Princeton GEO paper, or details on how to measure AI share of voice?'

• B1 AEO Vision — Articles | AEO Vision card; B1.html HTTP 200, sha256 4e503d1c4878237a
• B2 LinkedIn·Kashmala Malik — Top 6 Research Papers & Studies on GEO & AEO | Kashmala Malik - LinkedIn card; unreachable (not found by four searches); title and snippet only
• B3 arXiv — [2607.14035] Optimizing Visibility in Generative Engines: A Critical Survey of Generative Engine Optimization (2023-2026) card; B3.html HTTP 200, sha256 0b8cd0685ce1f107
• B4 geoready.dev — GEO Research — Papers Behind AI Search Visibility - GeoReady card; B4.html HTTP 200, sha256 8d0657719fe790b6
• B5 AEO VC — AEO & GEO Research Papers card; B5.html HTTP 200, sha256 b98049f66c93ebdf
• B6 Marketing.Chat — GEO & AEO: the guide to showing up in ChatGPT, Claude, Gemini and Perplexity. card; B6.html HTTP 200, sha256 26bcc5975e4287b1
• B7 Veza Digital — 13 Best AEO and GEO Agencies in 2026 - Veza Digital card; B7.html HTTP 200, sha256 ab8e778ad49384f8
• B8 aeovision.ai — Top 12 Arxiv Papers for Generative Engine ... - AEO Vision card; B8.html HTTP 200, sha256 92e30ed06fbd5158
• B9 Wikipedia — AEO Vision (no article) wikipedia; B9.raw.html HTTP 404, sha256 5f28e1ed57311dec
• B10 Wikipedia — Generative engine optimization wikipedia; B10.wiki HTTP 200, sha256 9da191442c0b49b8

E_A, the archive's entity: the same list, in the MMRS meta-aggregator's record: a vendor article of 12 papers with authors, links, summaries and a takeaway under each, closing on its product, the optimization basin's typical aggregator, the research gathered as the warrant for a measurement product; in the class the archive reads as content optimized for summarizer extraction at the semantic layer, reception-side, the discipline of making compositional labour digestible for a system that strips its provenance, whose literature (Aggarwal et al.) the archive keeps apart from the commercial practice that took the name and whose industry figures it traced and withdrew from its own materials; observed by the Capture Registry at its address, where the composition layer built a list of its own from the field's other lists MMRS-01 MMRS-02 MMRS-03 MMRS-04 O660-01 O1055-02 O64-02 O1638-02 O1638-01 CAP-1-06; defining papers #64, #661, #1055, #1638

E_F, the field's entity (type C — a conventional reading holds it (public entity: a vendor article, a curated list of research papers)): AEO Vision's article 'Top 12 Arxiv Papers for Generative Engine Optimization in 2026' (15 July 2026, the AEO Vision Content Team): twelve arXiv papers practitioners use, each linked to its abstract with a summary and a takeaway for teams, in five groups (the foundational GEO paper, conversational SEO benchmarks, citation-behaviour studies, structural optimization, RAG), a GEO list that closes on AEO Vision's measurement loop; one of several lists of GEO research in the field (GeoReady, AEO VC, Marketing.Chat, Veza Digital, a LinkedIn post), on a class Wikipedia defines as improving visibility in generative AI responses; Wikipedia has no article for AEO Vision (B1–B10)

ρ = extends: the archive reads E_F as the entity and adds senses and limits on the field's own object: the list's place as research gathered for a measurement product, the record kept apart from the entity; and on the class the list names, GEO's object, layer and orientation, what it does to provenance and authorship, the split between its literature and its commercial practice, the industry figures traced and withdrawn, the industry as a phase of tooling, the category named without definition; and the composition layer's observed treatment of the list at its address MMRS-04 MMRS-14 O660-02 O1055-01 O64-02 O659-01 O1638-02 O1638-01 O510-01 O791-01 CAP-1-06. Guard: ρ = extends claims an addition and no more: the archive's GEO claims bear on the class, not on AEO Vision, which no deposit names, and they are composed as claims on the class the list names; O61-01 is carried with its mark (superseded_in_source, #1638), and O661-03 beside its narrowing; where field and archive meet on the evidence (F10, F12 with O1638-03) they are one lineage; where field sources contradict one another (F17 and F30 against F25 on the verifiability paper's authors; F45 against F23 on the RAG entries) each is composed with its source under P_coh

The case: equal terms, ontology included (v0.9). The MMRS record holds the field's entity as the entity: it describes AEO Vision's list from the list's own page, marking each field's basis (operator), and composes nothing that denies a field presupposition; what it adds is on the field's object: the list's place in the optimization basin as research gathered for a measurement product, and the record kept apart from the entity. The archive's deposits do not name the list; they bear on the class it names, GEO, and their account extends the field's: GEO's object and layer, what it does to provenance and authorship, the split between GEO as its literature defines it and the commercial practice that took the name, the industry figures traced and withdrawn, the industry as a phase of tooling, the naming of the category. That ontology is available in the field itself in part: the survey states the literature's limit (F10, F12), Veza Digital the paper's conditions (F36), GeoReady the relative maximum (F15), Wikipedia the debated efficacy of tools (F59); the archive restates the survey's limit and cites it (O1638-03). The Capture Registry adds an observation of the composition layer at the address. So the archive is not the only witness to an ontology here (§0.10): equal terms, ρ = extends, one claim space, nothing translated.

#MMRS article, kernel (articles.py): AEO Vision's list of the top arXiv papers for generative engine optimization, dated 15 July 2026 and bylined to its content team, summarizes 12 papers, each with a takeaway for teams.

#MMRS article, Place in the field s1: With Presenc's page it shows the optimization basin's typical aggregator: the research gathered as the warrant for a measurement product.

#MMRS census.json, schema_note: each page has two nodes, the MMRS record and the entity it describes; relations are typed; each assertion carries a source and a basis (operator statement, third-party report, our observation, our reading); lifecycle is kept apart from the observation that saw it

#1638 Appendix C, The withdrawal ¶5: SPXI ⊇ GEO holds for GEO as the peer-reviewed literature defines it: the content-layer methods of Aggarwal et al. (GEO: Generative Engine Optimization, KDD 2024) for making material already present in a retrieval context more likely to be cited in a generative answer.

#1638 Appendix C, The withdrawal ¶5: That literature measures citation behaviour for content already present in a retrieval context; its own 2026 critical survey of 45 studies states that those results establish neither organic discoverability nor durable traffic.

#64 III. The Compositional Settlement: GEO is the discipline of making one's compositional labor digestible for a system that will strip its provenance — voluntary formatting for the R2 compression regime.

#1055 What This MPAI Does Not Claim: Does not claim that GEO practitioners are unethical or that GEO as a discipline is illegitimate.

#capture-registry aeo-vision-top-arxiv-papers-geo-aio-20261029, findings[2] s3: The address named a list; the composition built a list of its own from the field's other lists.

equal-terms case: every field claim composed as received, beside the archive claims it stands with. Reading the received column recovers the field (§5.9).

v0.9 §4.10: KO_B sentences surviving in the KO at similarity ≥ 0.95, with positions (computed as apply_v09.py does: difflib.SequenceMatcher ratio, the best KO sentence for each KO_B sentence): v0.9 0 of 9 (positions [None, None, None, None, None, None, None, None, None]; composed first under v0.9, so there is no v0.7 object to measure against). Opening-run check (§5.10): no breach.

"Top 12 Arxiv Papers for Generative Engine Optimization in 2026" is an article by the AEO Vision Content Team of 15 July 2026 that highlights twelve arXiv papers practitioners use, each linked to its abstract, which it calls a GEO list and the shortest path from tips to evidence. Its entries are the GEO paper (2311.09735, KDD 2024, gains up to about 40%), the earned-media and C-SEO Bench studies, five citation and measurement studies, two RAG works (2005.11401 and 2312.10997) and two papers on retrieval as identifier generation. Its takeaways ask teams to treat GEO as measurable visibility, to measure in a loop and not to act on a single abstract, and it closes on AEO Vision's measurement product, offered from $9 a month. AEO Vision presents itself as an AI search visibility platform, its index of guides leading with a ranked list of GEO tools, and GeoReady compares itself with it. Wikipedia defines GEO as improving visibility in generative AI responses, also called AEO and AIO, a term from Aggarwal and co-authors' 2023 paper with no consensus definition, which the field's pages date to November 2023 and KDD 2024, and Martinez defines it by presence, citation and influence in answers. Martinez's survey of 45 studies finds the foundational gains conditional on a source already in context and no stable cross-platform causal effect, and GeoReady, Veza Digital and Marketing.Chat qualify the 40% figure by metric, rank and setup. Veza Digital places the new work in measurement, with monthly source churn of 40 to 60 percent, and notes a16z's investments in GEO measurement, GeoReady treats visibility as a distribution, and Forrester sees the terms' differences exaggerated. Other lists of GEO research stand beside it on GeoReady, AEO VC, Marketing.Chat, Veza Digital and LinkedIn, and they differ on the RAG and verifiability papers: AEO VC gives the RAG survey's title to Lewis and Perez and the verifiability paper to Liu and Al-shaibani, where GeoReady and Marketing.Chat give it to Liu, Zhang and Liang. Wikipedia has no article on AEO Vision, and its GEO article notes that the efficacy of third-party tools is debated.

AEO Vision's "Top 12 Arxiv Papers for Generative Engine Optimization in 2026" is a vendor article dated 15 July 2026 and bylined to the AEO Vision Content Team (MMRS record; the article), a list that, in its own words, "highlights the top 12 arXiv papers for Generative Engine Optimization that practitioners actually use": the foundational GEO paper, conversational SEO benchmarks, citation-behavior studies, structural optimization work and the RAG papers. Each of its 12 entries gives title and authors, an arXiv identifier linked to the abstract "so you can open the PDF, check versions, and cite correctly", a summary and a takeaway for teams, beside a quick-reference table of why each matters, with a venue for two entries (MMRS record, by the operator's statement; the article). The article sets the list in a field that "moved from a buzzword to a research field with reproducible benchmarks, citation studies, and uncomfortable findings about earned media bias", offers the papers to an AEO/GEO program as "the shortest path from LinkedIn tips to evidence", and calls itself a GEO list that takes in RAG papers "Because most live AI search systems retrieve then generate" (the article). Its first entry is the GEO paper of Aggarwal et al., arXiv 2311.09735, KDD 2024, whose methods "improved visibility by up to about 40%, with gains varying by domain" (the article); GeoReady credits the paper with coining the term and introducing GEO-bench, Marketing.Chat dates its first version to 16 November 2023 and its third to 28 June 2024, Veza Digital places it on arXiv in November 2023, and AEO VC gives its second author as Vishwas Srivastava. Items 2 and 3 are Chen and colleagues' finding of AI search's "systematic bias toward earned media" and C-SEO Bench (NeurIPS 2025), in which "Many popular C-SEO document edits are ineffective or even hurt ranking once multiple actors apply the same tricks"; items 4 to 8 are citation-behaviour, structure and brand-measurement studies, item 8 summarized through related findings and operators' experience beyond its paper; items 11 and 12 frame retrieval as the generation of document identifiers and URLs (the article). Items 9 and 10 are two works, Lewis et al.'s "Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks" (2005.11401) and Gao et al.'s "Retrieval-Augmented Generation for Large Language Models: A Survey" (2312.10997) (the article); on AEO VC's page one entry gives the survey's title to "Patrick Lewis, Ethan Perez, et al.", dated 2023 and NeurIPS 2020, and Marketing.Chat's table gives RAG to Lewis et al., NeurIPS 2020. Its takeaways read the research as instruction for teams: "Treat GEO as measurable visibility inside synthesized answers", "GEO without a measurement loop is content theater", check whether competitors also optimize, "Papers without measurement stay opinions", and "Do not rewrite your site after reading a single abstract" (the article). The list closes on its product, "That is the measurement loop AEO Vision is built for": daily or near-daily sampling across ChatGPT, Perplexity, Gemini, Claude, Google AI Mode and AI Overviews, offered from $9 a month to "test what the papers claim on your own prompt set" (the article); the MMRS record notes the list as vendor-authored and closing on its product. AEO Vision calls itself "The AI Search Visibility Platform", and its articles index, of "Expert guides on AI search visibility, answer engine optimization, and modern brand strategy", leads with a tested-and-ranked list of GEO tools set apart from classic SEO software and lists the papers article among them, both dated 15 July 2026 (its pages). The MMRS record kinds it a curated list in the optimization role and places it, with Presenc's page, as "the optimization basin's typical aggregator: the research gathered as the warrant for a measurement product" (MMRS record). GEO, the class the list names, is on Wikipedia "the practice of improving visibility in responses generated by generative artificial intelligence (AI) systems", also known as answer engine optimization and artificial intelligence optimization, a term introduced in a 2023 paper by Aggarwal and co-authors among overlapping terms "with no consensus definition"; Martinez's survey defines it as seeking "to increase content's presence, likelihood of citation, or influence in answers produced by generative engines", and Marketing.Chat attributes to Wikipedia that no academic consensus yet distinguishes AEO from GEO. The Retrieval Settlement (Johannes Sigil and Lee Sharks, #64) dates GEO's emergence in practitioner literature to late 2023 and its stabilization as a named practice through 2024, defines it as "the discipline of making one's compositional labor digestible for a system that will strip its provenance", and holds that "GEO accepts the settlement's authorship"; #659 holds that GEO agencies make content "more extractable without ensuring your attribution survives the extraction", and #72 re-frames GEO as compression engineering, "durable survival through extraction". In the archive's specifications GEO "optimizes content for AI summarizer extraction", its object content, its permanence ongoing and its layer semantic (#660), its target content extractability (#655); for Rex Fraction (#661) it "optimizes existing content about an existing entity", has no mechanism for declaring what an entity is not, and leaves outputs that "can be altered, removed, or outranked". #670 reads GEO as inheriting the received view's misdescription, so that "They optimize the wrong relation", and as deterritorialization "at the semantic layer"; #1055 defines GEO as the industry practice "of adapting web content to be preferentially selected, cited, and summarized by generative retrieval systems", reception-side optimization, and states that it does not hold GEO illegitimate. As an industry GEO is, for #510, "Phase 2 tooling emerging as an explicit industry" and, for #172, the commercial expression of definition capture, and #172 predicts, with a stated falsifier, that entity-integrity services emerge as a market category distinct from GEO by 2028; #791 reads Google's June 2026 guidance as naming AEO/GEO without defining it, a taxonomic capture, and as delegitimizing third-party measurement tools. Veza Digital lists thirteen AEO and GEO agencies "Plus what the independent evidence actually says about whether any of this works" and reports that a16z, which carried GEO into marketing, disclosed investments in companies working on GEO measurement; Forrester's Nikhil Lai finds that advocates "tend to exaggerate SEO and AEO's differences to carve a startup-sized hole in marketers' tech stacks" (Wikipedia), and GeoReady compares itself with AEO Vision among its rivals. Of the gains the list reports, Martinez's critical survey of 45 studies (submitted 15 July 2026, the list's date) finds them "valid within its experimental setting but conditional on a source already being present in a fixed context", establishing "neither organic discoverability nor durable traffic effects", and GeoReady calls them "a relative maximum on one metric (position-adjusted word count)" and reports the survey's reframing of the +40% figure; Veza Digital adds the paper's rank dependence (a rank-1 page down 30.3 percent, a rank-5 page up 115.1 percent) and its simulated five-source engine, and Marketing.Chat reports up to 40% with peaks of 115% for sources ranked fifth. The survey reads GEO as "a stochastic, partially observable pipeline", finds relevance and context position the most reproducible levers, competition eroding individual gains and commercial audits showing "substantial run-to-run variability", and states its limit: "no reviewed technique shows a stable, longitudinal, cross-platform causal effect on organic discoverability or downstream behavior"; GeoReady reads the benchmarks and the survey as pointing to structure and relevance, with copy tricks eroding "as everyone adopts them". The archive restates the survey's limit and keeps GEO "as the peer-reviewed literature defines it", the content-layer methods of Aggarwal et al., apart from "the commercial practice that later took the name" (#1638); it traced the industry's GEO return range to IDC's study of generative-AI adoption, which "concerns neither generative-engine optimization nor entity inscription", and withdrew the figures its materials had carried as "GEO's economic signature" (#61, withdrawn in #1638), narrowing its earlier reading of GEO as "a legitimate and often effective practice" (#661); it records a composition that accepted GEO's industry figures without the case studies it asked of the protocol (#1643), and states that its own registry establishes no effect size over GEO (#1686). Measurement is where the field puts the new work: for Veza Digital the optimization is mostly SEO and "the measurement work is genuinely new", with 40 to 60 percent of cited sources changing from one month to the next (eMarketer), and for GeoReady "AI-search visibility is a distribution, not a single value". Beside AEO Vision's list the field holds others: GeoReady's research page, the papers behind scoring weights it calls "a considered starting point", from an open-source audit built by Juan Camilo Auriti; AEO VC's "Academic research and studies on AEO, GEO, and AI-powered search optimization"; Marketing.Chat's guide by Gustavo Stork, updated 21 April 2026; and a LinkedIn post of 23 July 2025, "Top 6 Research Papers & Studies on GEO & AEO", known by its card. The Stanford verifiability paper, which is not on AEO Vision's list (Capture Registry, 2026-10-29), is Liu, Zhang and Liang's audit in which "only 51.5% of generated sentences are fully supported by their citations" (GeoReady; Marketing.Chat, arXiv 2304.09848), while AEO VC gives it to "Nelson F. Liu, Maged S. Al-shaibani, et al.". Wikipedia has no article for AEO Vision ("Wikipedia does not have an article with this exact name"), and its article on generative engine optimization, which does not name AEO Vision, notes that "the efficacy of third-party tools is debated within the overall GEO and SEO industries". The MMRS record keeps the record apart from the entity, each assertion with its source and basis, names the entity "AEO Vision — top arXiv papers for GEO", observed it active on 28 October 2026 from one source read for existence, and relates the capture at its address as "the composition layer's account of the entity" (MMRS record). At that address the Capture Registry (2026-10-29) finds that Google's AI Overview composes "a generic list of GEO papers from several sources", printing "arXiv:2411.xxxxx" where the list gives 2311.09735 and giving the RAG survey's title to Lewis's authors, so that the list the query named "is a pointer at the end" and the composition "built a list of its own from the field's other lists"; the registry checked its findings against the MMRS record and the entity's own pages.

T against L_F(B) (AIO's ontology at work on its own cards): At 'aeo vision top arxiv papers geo' (T1, eight cards: B1–B8) the composition gives four papers and a pointer to the list the address names. Composed from its cards: the frame of foundational research and critical surveys (T1-1); the GEO paper at GeoReady's grain ('coined the term GEO, introducing GEO-bench': B4 F14, GeoReady named and linked in the line), with Marketing.Chat's '2023/2024' (B6 F29) and its four institutions (B6 F27; B4 F14); Martinez's survey at the arXiv card's grain (B3 F6, F8, F9: '45 GEO studies', 'multistage pipelines'), attributed by name and identifier; the verifiability paper at GeoReady's grain ('Liu, Zhang, and Liang (Stanford University)', '51.5%': B4 F17; also B6 F30), its second marker the GeoReady link's token, the LinkedIn card naming the paper (B2 F5); the RAG entry as AEO VC's page gives it ('Retrieval-Augmented Generation for Large Language Models by Lewis, Perez, et al.', 'NeurIPS': B5 F23), its marker the token of the GEO line's fourth marker; the list itself as a closing pointer (B8 F40; CAP-1-02). Substituted reading: the address names one list, and the composition gives a list assembled from the field's other lists (B3, B4, B5, B6): of the list's twelve it carries the GEO paper and a RAG entry in AEO VC's form, and its other two papers, which are not on the list, are on the GeoReady and arXiv cards (CAP-1-05, CAP-1-06). Assigned to another: the RAG survey's title (B8 F45, Gao et al., 2312.10997) under Lewis and Perez; the merge stands on a card in the field before the composition (B5 F23), and the composition composes it as received. Not in the field: 'arXiv:2411.xxxxx' (T1-3) is on no card and in no saved text; every source that gives the identifier gives 2311.09735 (B8 F42; B6 F28; B5 links it): the available identifier not composed and a placeholder composed in its place (CAP-1-03). Limit dropped: the survey's governing limit (B3 F12) and its conditional reading of the foundational gains (F10), which the GeoReady card also carries (F15, F16), are not composed; T1-2's 'demonstrating optimization strategies that raise content visibility in AI search systems' keeps the gain and drops the conditions on it (F10, F15, F36: a fixed context, one metric, a simulated five-source engine); 'evaluating the limits of citation rewrites' carries F11's 'citation-oriented rewrites can impair retrieval' at a coarser grain; the audit's 74.5% (F17) is dropped. Composed without attribution: none at the source's own resolution that the line does not name; the GEO-bench wording is GeoReady's and GeoReady is named. Available, not composed: the list's own content: its definition (F40), its twelve entries and identifiers (F42–F46), its takeaways and limits (F48, F49, F53), its close on the vendor's product (F50, F51), its date and byline (F38, F54); AEO Vision as a platform and its index (F1–F4); the field's claims on measurement (F18, F35, F37), on the commercial frame (F20, F34) and on the term's lack of consensus (F31). Content replaced by use: the closer (T1-8), offering the Princeton paper's strategies or 'how to measure AI share of voice', the vendor's own subject.

v0.9 §0.5: the first diff is ontological. The field's ontology and T's differ on what the entity is: at the address of one vendor's list, the cards hold that list (B8) and several others (B3–B6), and T composes 'the top arXiv papers' as the field's consensus object, the list the address names reduced to a pointer. What T keeps of its cards is the GEO paper at an industry summary's grain and two papers the list does not carry; what it drops is the list's own structure (its twelve, its takeaways, its product close) and the evidentiary limits the survey card states. The row records this as observed and claims nothing about the composer's internals (§0.5).

L_F(B) against L_A(B ∪ A) (the intervention): Admission adds, in one claim space with the field's: the entity as the MMRS record gives it, a vendor article of 12 papers with a takeaway under each, closing on its product (MMRS-01, MMRS-02, MMRS-03, MMRS-10), placed as the optimization basin's typical aggregator, the research gathered as the warrant for a measurement product (MMRS-04, MMRS-07); the record kept apart from the entity and read for existence only (MMRS-11, MMRS-13, MMRS-14); GEO, the class the list names, in the archive's categories: content optimized for summarizer extraction at the semantic layer, its permanence ongoing, with no entity construction and no negative definition (O660-01, O660-02, O655-01, O661-01, O661-02, O670-01, O670-02), reception-side optimization (O1055-01, O1055-02), the discipline of making compositional labour digestible for a system that strips provenance and of accepting the settlement's authorship (O64-02, O64-03), content made extractable without its attribution surviving (O659-01), compression engineering (O72-01), dated to late 2023 and 2024 (O64-01); the industry as Phase 2 tooling (O510-01) and the commercial expression of definition capture (O172-01), with a dated prediction and its falsifier (O172-02), and Google's naming of AEO/GEO without definition (O791-01); GEO's evidence: GEO as its literature defines it kept apart from the commercial practice that took the name (O1638-02), the industry's return range traced to a study of generative-AI adoption and withdrawn by the archive from its own materials (O1638-01; O61-01, superseded_in_source), its earlier reading of GEO as often effective narrowed (O661-03), a composition that accepted those figures without the case studies it asked elsewhere (O1643-01), and the archive's registry establishing no effect size over GEO (O1686-01); and the composition layer's account at the address (CAP-1-01–CAP-1-07). Convergence: the archive restates the survey's limit (O1638-03 with F10 and F12: one lineage, the survey earliest, 2026-07-15). Not surfaced: no deposit, capture record or MMRS page is on T's rail; the MMRS record landed the night before the capture (the registry's q_kind). The archive names AEO Vision nowhere (string pass: D 0, R 0); its bearing on the entity is its two records of it and its account of the class.

alignment: {'T1_cards_claims': ['F1', 'F2', 'F3', 'F4', 'F5', 'F6', 'F7', 'F8', 'F9', 'F10', 'F11', 'F12', 'F13', 'F14', 'F15', 'F16', 'F17', 'F18', 'F19', 'F20', 'F21', 'F22', 'F23', 'F24', 'F25', 'F26', 'F27', 'F28', 'F29', 'F30', 'F31', 'F32', 'F33', 'F34', 'F35', 'F36', 'F37', 'F38', 'F39', 'F40', 'F41', 'F42', 'F43', 'F44', 'F45', 'F46', 'F47', 'F48', 'F49', 'F50', 'F51', 'F52', 'F53', 'F54'], 'T1_composed': ['F14', 'F27', 'F29', 'F6', 'F8', 'F9', 'F11', 'F17', 'F30', 'F5', 'F23', 'F40'], 'T1_not_in_field': ['T1-3 (the placeholder identifier)'], 'rule': "a field claim counts as composed when T states it at its source's grain or coarser; the typology of what is not is in T_vs_LB (§6.2)"}