Jev-powered X content desk that drafts but never auto-posts
The author uses Jev to make a Choice (skip, short draft, full draft, escalate), a fit Score and a safety yes/no for each X post, with drafts going to review and a human always pressing Post.
It is becoming f*cking dangerous: people are wiring agents straight into their X accounts. the article below has the one rule that keeps yours alive a content desk gate, before your writer burns a single token: → the post is under your view floor? skip → the author isn't on your watchlist? skip → strong fit? the LLM drafts a short or full reply into a review folder → then you hit Post. always what Jev answers on every post: → Choice: skip, draft short, draft full, or escalate to a human → Score: how well it fits your account → yes/no: safe to engage without a human? the default is no why the last one matters: X already removed 42,000 accounts for AI auto-replies. a model that decides what's worth answering is fine. a model that answers in your name is how you lose the account the video is that desk: 4 posts in, 2 skipped for free, 2 drafts in review, the Post button locked paste this into Claude Code: "Build a review-only content desk. For each post: skip if under [X] views or off my watchlist. Rate the fit, then draft a short or full reply only for strong fits, into a review folder. Never publish. Every draft waits for me." the full playbook is in the article below ↓
This post quotes another post: x.com/thegreatest_sv/status/2102802378378621355