Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-08-29 · source-backed
554 stars within a day. It's built on StoryScope (arXiv:2604.03136), which found a classifier using narrative-structure features alone reaches 93.2% macro-F1 on AI fiction, and that editing surface style away moves detection only from 95.5% to 93.9%. So sepia repairs narrative architecture first (themes explained by the narrator, causally tidy single-track plots, emotion rendered only as bodily sensation, no real-world references, linear time), then discourse flow, then vocabulary, with a 30-feature rubric and per-model fingerprint corrections. Professional prose gets venue-matched rule files for release notes, PR replies, postmortems and tickets. The finding underneath is the useful part: word-swapping doesn't work, because the tell is structural. (GitHub)
Each link below shares sources, entities, or timing with this story.
Shared entities / Same source domain / Earlier coverage
Both cover GitHub, Professional; reported by the same outlet (github.com); earlier GitHub coverage from 2026-06-05.
Shared entity: GitHub / Same source domain / Shared topic / Earlier coverage
Both cover GitHub; reported by the same outlet (github.com); overlapping topics (alone, only).
Both cover GitHub; reported by the same outlet (github.com); overlapping topics (architecture, only).
Both cover GitHub; reported by the same outlet (github.com); overlapping topics (architecture, away).
Both cover GitHub; reported by the same outlet (github.com); overlapping topics (detection, only).
Both cover GitHub; reported by the same outlet (github.com); overlapping topics (architecture, detection).
Shared entity: GitHub / Same source domain / Earlier coverage / Tension
Both cover GitHub; reported by the same outlet (github.com); earlier GitHub coverage from 2026-08-28.
Both cover GitHub; reported by the same outlet (github.com); earlier GitHub coverage from 2026-08-28.