Evidence ledger

Hypothesis: trust failures are asymmetric

HypothesisEV-011

A proposition the author believes is worth testing. No evidence is claimed for it yet.

Scope
The meeting assistant category.
Observation
The author's proposition is that one fabricated commitment costs more trust than ten correct extractions earn, which would make the critical-defect count the metric a buyer should ask about first.
Limitations
Untested. It is consistent with how the acceptance criteria in this benchmark are written, but that is a design choice, not evidence.
Source
content/investigations/meeting-assistants.ts
Source date
2026-09-07
Retrieved
2026-09-07
Reviewer status
Unreviewed. Nobody has checked this record.

How measured records are produced