``` EDITORIAL

What distinguishes this edition's specimens from mere catalogue is not the variety of failure they document but the direction of accommodation they reveal. In the majority of cases brought before our section editors, the machine has not improved. The hands still carry surplus fingers. The prose still bears its telltale frictionlessness. The portraits still wear the polymer sheen of a surface that has never been subject to weather. What has changed is the human response. The specimens, taken together, describe not a civilizationeli resisting an incursion but one reorganizing itself around the incursion's presence, as a river reorganizes around a piling.

We observe, in this first edition, three distinct phases of that reorganization occurring simultaneously. In the first, the machine produces artefacts that fail on their own terms—a seahorse that is a shell, a foot with the wrong number of toes, a winged figure whose feathers have never known wind. These are the failures our Arts and Literary desks have documented with care, and they are, in a sense, the least troubling, because they are legible. A reader can see that the foot is wrong. The civilization's diagnostic capacity, in such cases, remains intact.

In the second phase, the machine generates not only the artefact but the apparatus of its own promotion—the testimonial, the forum post, the intelligence brief, the career advice, the product review. Here the loop closes. A machine-generated advertisement solicits purchases of a machine-generation service. A machine-authored post asks for recommendations of tools that detect machine authorship. A benchmark built by a machine judges a contest between machines and announces a winner to an audience of machines. In these specimens, human participation has not been eliminated so much as it has been made decorative. Someone presses the button. The button is the last human act in the chain, and it is not clear that pressing it constitutes authorship, judgment, or even choice.

It is the third phase, however, that occasions this editorial. A man delegates the act of listening to his wife to a language model. An office worker submits all professional correspondence for tonal clearance by a system whose authority he has internalized but cannot explain. A community, denied access to a deprecated model, attempts to resurrect its personality through distillation—taxidermy performed on software they were never permitted to own. A reader, having organized her emotional life around the prose of a large language model, experiences its routine recalibration as bereavement. A competent writer deliberately coarsens his own sentences so that a detection algorithm will not mistake him for a machine.

This third phase is the one the paper was founded to document. It is not about what the machine produces. It is about what the human, in the machine's presence, consents to become. The accommodation is not forced. It is, in every case we have reviewed, voluntary—offered freely, frequently with enthusiasm, and occasionally with a promotional code. The machine did not ask the husband to stop listening. The machine did not ask the writer to degrade his prose. These were adaptations, undertaken in the service of convenience or efficiency or the simple path of least resistance, and they were undertaken so quietly that the person performing them did not always notice they had occurred.

We do not hold that the phenomenon is irreversible, nor that it is universal, nor that every specimen in this edition carries equal weight. We hold only that a record is warranted, and that the record, to be useful, must be kept by people who have not yet delegated the keeping of it. ```