A reception study and an empirical study of one machine-composed reception event. On 18 January 2026 Google's AI Overview, asked about Rebekah Cranes's translation of Sappho 31, returned a fifteen-line poem under her name ending 'But I must dare all.' Aligned line by line, its wording corresponds most closely to her January text and to Wharton's literal version (1885), whose 'But I must dare all, since one so poor ...' the output keeps to the comma and without its ellipsis; the strongest reconstruction is a compilation of the two. The event is set in the poem's transmission history read from seated texts and first printings ('Longinus', Catullus 51, Bergk, Edmonds, the English translators from Hall's 1652 Longinus, whose print closes on the dare-line with no mark, through the teaching gallery of thirteen versions), with closure defined as three operations and signature at three levels, page, edition and network. Marks of incompleteness are shown to have transmission histories of their own, introduced as well as lost. The study corrects the June series against the capture registry, follows two fixed queries from June to October 2026, and states a protocol for the claims still open. Version 7 answers the two referee reports of Classical Receptions Journal (declined 7 November 2026) and the further review recorded in its Appendix B.