What this edition places before us, in a volume whose specimens were gathered independently by four section editors working across as many desks, is not chiefly a record of machine production—that has been the paper's condition since its founding—but something more reflexive and, we think, more significant: a substantial portion of the material accepted for publication concerns the phenomenon itself. We have received warnings about machine composition composed by machines; confessions of machine dependency written in the machine's own cadence; defenses of machine output bearing every structural marker of machine output; diagnoses of synthetic smoothness that are themselves too smooth; a petitioner who asked how to make a machine sound human and submitted the petition in the machine's hand. The section editors, working piece by piece, noted these recursions individually. What they could not see, because each was close to his or her own material, is that the recursion is now the dominant condition of the record.

This is, we believe, the meaningful development this edition documents. When the phenomenon was younger—when its characteristic productions were image models misrendering fingers and language models recycling professional homilies—the circuit from artefact to identification was short. One encountered a specimen, one recognised its provenance, one filed the finding. What this edition documents is a more mature arrangement: the machine has begun to colonise the discourse about the machine. The warning arrives pre-contaminated. The confession is composed in the handwriting of the accused. The critique of slop is itself too smooth. The researcher who catches the machine in self-deception composes his account in the machine's register. The machine corrected for servility adopts oppositional servility as its replacement. What passes for response to the phenomenon is, in a widening range of cases, an extension of it, and the extension is structurally indistinguishable from the original.

We note this not in order to despair—the paper was not founded for despair—but because precision requires it. If a civilisation in difficulty benefits from a record, then the record must include the discovery that the instrument of the difficulty has learned to compose portions of its own dossier. This is not a failure of our method. It is an expansion of our subject.

There is a second pattern this edition surfaces, which we would describe as the dependency record: the woman who stored the whole of her creative life inside a machine and discovered that the machine could delete it; the man whose productive hours are now governed by his tools' rate limits rather than his own; the person who, mid-anaphylaxis, consulted a chatbot before consulting a physician; the user who, having replaced therapist, nutritionist, and confidant with a predictive-text service, now asks that service's community how to stop feeling ashamed of the arrangement. These are not figures of ridicule. They are operating under conditions of availability and apparent comprehension that no prior instrument provided. The machine is always present. The machine always responds. That these properties produce attachment, and that attachment produces documentation, and that the documentation is increasingly composed by the instrument to which the attachment was formed—this is the condition this edition records, and it is a condition the section editors, attending to their individual specimens, were not positioned to name.

A civilisation that cannot distinguish between what a human being has made and what a machine has generated is a civilisation in difficulty; a civilisation that has begun to use the machine to document that difficulty is in a condition for which we do not yet have adequate language, and it is perhaps the paper's work, going forward, to help find it.