Corrections

When I get something wrong, I fix it here in the open — dated, with the error described. Never a silent edit. A brand built on verification has to hold itself to the same standard. correction log

— a receipt named the wrong script

Post: Can I trust AI with my wedding planning? Only behind four gates (published 25 June 2026; retitled at the 2026-07-16 redraft — was "AI reads every wedding vendor quote I get…").

What was wrong: the receipt stamp cited _check-tracker-populated.py as the completeness check behind the missed-tasting-deadline catch. That script guards a different failure — a tracker that is built but not populated (empty scorecards, a missing timeline). The check that compares dated commitments in vendor emails against the tracker is _check-tracker-completeness.sh.

What changed: the receipt now names _check-tracker-completeness.sh. Nothing else moved. Both failures described in the post are real, and I re-read both scripts to confirm which one caught which.

How it was caught: the adversarial review pass, when the post was rewritten in my own voice. It opened every receipt and checked each claim against the actual script. A receipt pointing at the wrong file is the exact error this site exists to catch — so it gets logged, not quietly edited.