AI编码
infoq.com ·
@Steef-Jan Wiggers
Anthropic 称无需完整心智模型,实践者正在检验 Claude Code 评审方式
Anthropic 公开了 Claude 承担约 90% 代码后 Claude Code 团队的代码评审做法,并称无人保有代码库的完整心智模型。文中记录了一个 PR 过夜经历 50 多轮评审、评审者大部分精力花在既有代码上,以及团队承认尚不具备的某项控制。InfoQ 指出实践者正在检验这一说法。
AI推荐理由 InfoQ 转述 Anthropic 关于 Claude Code 团队代码评审方式的公开说明,包含约 90% 代码由 Claude 编写、单个 PR 过夜经历 50 多轮评审、评审精力集中于既有代码、团队自认尚缺某项控制等具体细节,对 AI 编程与工程流程有参考价值;但属二手转述,需回看原文核实。
72
正文

Anthropic has published how the Claude Code team reviews code now that Claude writes roughly 90% of it, including a claim that nobody keeps a complete mental model of the codebase. The post documents a pull request that ran more than fifty review rounds overnight, a reviewer spending most of its effort on pre-existing code, and a control the team says it does not yet have.
*By Steef-Jan Wiggers*