Phenomenology-grounded eval will remain unfrontier-lab-copyable because methodology requires taste/editorial judgment, not just calibration.

{
  "doc": "/Users/antonborzov/Documents/Claude/work/AboutBlank/strategy/2026-05-12_room-to-play-in-eval-cohort.md",
  "claim": "Phenomenology-grounded eval will remain unfrontier-lab-copyable because methodology requires taste/editorial judgment, not just calibration.",
  "horizon": "open",
  "section": "What it would NOT do",
  "category": "technical",
  "confidence": 0.65,
  "depends_on": [],
  "extracted_at": 1778899700,
  "extracted_by": "haiku-4-5",
  "status_check_log": [],
  "evidence_threshold": "PROPOSED: Frontier lab attempts replica benchmark; succeeds at replicating scores but fails to capture About Blank's editorial selectivity or attracts criticism for self-grading"
}