技术负责人应该产生最多的AI代码痕迹
技术负责人应该用最多的AI代码生成量来保持手感。 AI编程工具还没有标准用法，技术决策需要一手经验。实测结论：上下文窗口控制在50%以内，超过20%触发警告；模型过于自主会擅自提交代码，需要明确的指令护栏；人类代码审查仍然必要，AI会自行脑补导致过度设计；代码库要"可grep"才能让AI高效理解。低AI消耗的技术领导者值得怀疑——当编程实践本身还在被设计时，delegate实验等于delegate自己的判断力。 
RSS: https://schipper.ai/feed.xml
I'm currently a Staff AI Engineer at Snowflake working on agents for GTM teams, and the author of nah. This site collects my public work on agents and developer tooling.
提交点赞
评论
点赞
收藏Parallel coding agents with tmux and Markdown specs
I’ve been running parallel coding agents with a lightweight setup for a few months now with tmux, Markdown files, bash aliases, and six slash commands. These are vanilla agents - no subagent profiles ...