A creator decision engine built on Jev for $0.63 a month
The author built a creator decision engine on TypeSafe's Jev model as a 300ms probability layer that scores hooks, routes DMs, filters comments and gates drafts, returning only choices and confidence scores rather than text, at 15m input tokens a month.
You simply IGNORE the real power of the Jev. I built a decision engine for creators that costs $0.63 a month to run. Not an LLM, a 300ms probability layer on top of TypeSafe's Jev model, the one that came out of stealth last week. It never writes a word, it only answers three things: pick one, rate this, yes or no. Every answer comes back with a confidence number attached, and the number is the whole product. > Feed it ten hooks, it returns ten confidence scores and flags the one that sounds like AI wrote it > Feed it an inbox, it routes every DM: paid work, fan, pitch, spam > Feed it comments, it runs a threshold ladder, hide at 0.7, ban queued at 0.9 > Feed it a draft, a regret gate scores it before it goes out 15m input tokens a month, output is free because there is no output, one long Claude session costs more than this thing runs for a month. Zero code, four JSON questions, full breakdown & setup below
This post quotes another post: x.com/1663262095868084228/status/210214035969030…