Ex-Anthropic engineer's prompt pairs Jev with Opus 5.5 as an autonomous business assistant
A former Anthropic engineer shares a single-prompt setup where Jev decides what needs doing and Opus 5.5 does the work, with inbox triage costing fractions of a cent and adjustable thresholds.
ex Anthropic engineer just leaked single prompt to turn Jev + Opus 5.5 into fully autonomous business assistant Jev acts as the brain, deciding what needs to be done, while Opus 5.5 handles the actual work. the loop running your day: INBOX → JEV TRIAGES → OPUS DRAFTS → JEV CHECKS → send or surface to you what that buys you: triage costs fractions of a cent, so you can run it on everything, all day. the expensive model only fires on the ~20% that earns it. and your judgment lives in thresholds, not vibes - want it more cautious about auto-sending to clients? move 0.7 to 0.9. done. no prompt rewrite, no praying. small model for system 1. big model for system 2. your business assistant should have been built this way from day one. posting everything I learn wiring these up. send this to whoever on your team is still triaging their own inbox. what's the first decision you'd hand off?
This post quotes another post: x.com/me_barnyx/status/2101976764779999381