When I had 'Jev' check 20 expense reports, it took 0.5
Ran 20 expense reports through Jev at 0.5 seconds each for under one yen, getting all duplicate, personal-use, and category judgments right when questions were split.
The original post, translated
When I had 'Jev' check 20 expense reports, it took 0.5 seconds per report, and for 20 reports it cost less than 1 yen. It got all answers correct for duplicate submissions, personal use, and account category judgments. However, when I casually asked 'Can I approve this?', the accuracy was 60%, but when I split the questions, it went up to 95%. 'Jev' is an AI that doesn't write sentences and only answers questions with 'yes/no' or multiple choice.
Show the original post in its source language ▾
『Jev』に経費申請20件をチェックさせたら、1件0.5秒、20件で1円未満で終わった。 二重申請・私用・科目の判定は全問正解。ただ「承認していい？」と雑に聞くと正解は6割で、質問を分けたら95%になった。 『Jev』は文章を書かず、質問に「はい／いいえ」か選択肢で答えるだけのAI。
The post above is a machine translation from ja; the untranslated text is in the fold-out.
Engagement when collected
Numbers are a snapshot taken from X when the case was added to the library (schema v1, collected 2026-09-19); they will not match today.
Where this case fits
Filed under general Jev builds. In the pattern Jev is built for, the model answers a bounded question per step — and ordinary code acts on the answer, because the answer is already a value rather than a paragraph. Browse the full case library for the work it sits next to.
Related Jev cases
Jev was free until 9/24, so I tried calling it from Claude
Dev tools
Yeah, it really does feel like an "extension of if," just
Keep browsing: all 1173 Jev cases · more from @AI · builders · what Jev is
Last updated: 2026-09-22 · sources & corrections · every card links to its author's original post