A fingerprint measured in a moving medium¶

Xu et al., "code_transformed: The Influence of Large Language Models on Code" (arXiv:2506.12014, revised 2026-02), measure style across more than 20,000 GitHub repositories linked to arXiv papers from 2020 to 2025 and report conventions shifting toward what models write: snake_case function names in Python rise from 40.7% in Q1 2023 to 49.8% in Q3 2025. They also measure similarity across projects; whether they claim convergence is not something the abstract settles and the full text has not been read.

Two consequences for this study, neither yet a measurement here.

Timeliness. If model-written code is spreading through the corpora, an organization's own hand is being overwritten while the study measures it. That makes the measurement worth making now and makes any null harder to read: a fingerprint that has faded is indistinguishable from one that was never there.

A threat to the train-to-test transfer. RQ1 trains on 2024-11 to 2025-09 and will test on 2025-11 to 2026-10, a year later, in exactly the period this drift is measured over. An adapter learns the hand its training window carried; if the test window's hand has moved toward the models', the contrast is attenuated by drift rather than by an absence of fingerprint. The apparatus can already say something about this without unsealing anything: the dev window sits between the two, and the per-project breakdown could be recomputed on the earliest and latest months of the train window to see whether the contrast is shrinking with time.