Fetching from the wire…
01
02
03
04
05
06
07
08
Research2026-08-21 · source-backed
Whole repository states in a QD archive, candidates as Git commits in isolated worktrees, judged by a project-supplied evaluator. In a matched Zstandard experiment, QD landed 0.135% below sequential champion editing (95% BCa interval -0.556% to +0.161%). arXiv Archive retention and stepping-stone sampling demonstrably occurred; the controlled experiment just didn't establish any endpoint benefit. Publishing that is more useful than publishing the version where they only ran the winning arm.
Each link below shares sources, entities, or timing with this story.
Shared entity: Whole / Same source domain / Earlier coverage
Both cover Whole; reported by the same outlet (arxiv.org); earlier Whole coverage from 2026-08-05.
Same source domain / Shared topic / Tension
Reported by the same outlet (arxiv.org); overlapping topics (better, endpoint); pushes against this story (against).
Shared entity: Publishing / Earlier coverage / Tension
Both cover Publishing; earlier Publishing coverage from 2026-07-25; pushes against this story (versus).
Shared entity: Whole / Earlier coverage / Tension
Both cover Whole; earlier Whole coverage from 2026-07-16; pushes against this story (but).
Same source domain / Shared topic
Reported by the same outlet (arxiv.org); overlapping topics (candidat, evaluator).
Shared entity: Publishing / Earlier coverage
Both cover Publishing; earlier Publishing coverage from 2026-08-20.
Shared entity: Zstandard / Earlier coverage
Both cover Zstandard; earlier Zstandard coverage from 2026-08-10.
Shared entity: Publishing / Earlier coverage
Both cover Publishing; earlier Publishing coverage from 2026-07-28.