Articles & threads
Claire Vo opens the hood on her week in Jev: nine cents to classify two years of ChatPRD pull requests, her Claude and Codex sessions sorted in minutes, and a real-time voice app where the quote API, not the decision model, was the slowest part.
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
The through-line is a question she now asks before building anything: does this step need prose, or just a decision? Decisions go to Jev at fractions of a cent; the interesting clusters go to GPT-6 Astra. One pattern from the episode sticks: her AI usage fell from nearly all engineering in January to under 40% by September as agents and media work took over. ChatPRD hosts written walkthroughs of every workflow she shows.
You can sort large archives of past work, such as team project logs or code updates, into clear topics. Before starting, gather the written records you want to organize. A developer can then set up Jev to compare items side by side and decide whether they share the same subject.
Jev only makes quick choices between fixed options and cannot write summaries or sentences. After it groups related records together, your developer can send each cluster to a standard text-writing model to name the topic.