Blind Eval

Use before landing anything where taste or output quality is the question and a test cannot decide it. Judges a change on its merits with authorship hidden, then keeps or reverts — a tie reverts, only proven uplift lands. Trigger words: blind eval, karpathy, keep or revert, quality gate, taste call, blind judge, A/B judge, prove uplift.

Tcuzzo Updated 2 repo stars
The Reflex Seam. On the left, a deterministic runtime owns state, files, rules and tests. On the right, a model judgment kernel owns inference, policy, reversib

File contents

Tcuzzo/backs-aios-skills commit c6f9e08444

Frequently asked questions

npx get-skill add tcuzzo/blind-eval