Our AI coding workflow:
Fable 5 leads (plans, briefs, decides) โ never codes.
Opus 5 (1M) implements every change from a full brief.
Medium effort, not high โ high reasoning makes Opus 5 stranger.
Fable reviews every diff. Fixes loop back to Opus, never patched by Fable.
@buildwithrajath Starting to do dumb things when you are about to hit 70% usage. This is base on my experience only.. But I don't know what the hell is happening with the model.