We scored our own AI harness against a published framework and got a 1
中文摘要
作者使用公开框架对其AI编码工具进行评分，结果仅得1分，强调了实测数据与普遍建议之间的差距。
English Summary
The authors scored their AI coding harness against a published framework and received a poor score of 1, contrasting empirical data with general advice.
Original Excerpt
Most writing about AI coding agents is advice. This is a score, and it is not a good one. Continue reading on Medium »