<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[simayi2018的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/simayi2018</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; simayi2018]]></copyright><item><title><![CDATA[AI 时代的研发团队该怎么定考核目标？]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/163540400</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/163540400</guid><author>simayi2018</author><pubDate>Thu, 06 Aug 2026 17:30:12 +0800</pubDate><description><![CDATA[人人都在用 AI 写代码，可绝大多数团队还停在"工程师主导、AI 辅助"的浅水区；能说清"AI 带来了什么"的人，少之又少。]]></description><category></category></item><item><title><![CDATA[砸钱给全员配了AI编程工具，怎么交出漂亮的成绩单？]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/163540329</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/163540329</guid><author>simayi2018</author><pubDate>Thu, 06 Aug 2026 17:29:12 +0800</pubDate><description><![CDATA[灯一亮，摸鱼的会心虚，落后的会露怯，真正的高手则站到了 C 位。也只有把这三件事全部说清、做实，企业的 AI 研发转型才算真正落地，不再困在]]></description><category></category></item><item><title><![CDATA[如何评估团队 AI 编码实践的成熟度？VibeInsight 用数据代替感觉]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/162964143</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/162964143</guid><author>simayi2018</author><pubDate>Fri, 17 Jul 2026 12:15:30 +0800</pubDate><description><![CDATA[什么是 AI 编码实践的成熟度？为什么很多团队明明天天在用 AI，却说不清自己到底用得怎么样？这是研发管理者在推广 AI 编程半年后，最容易卡住的一个问题。思码逸的体会是：评估成熟度不能靠感觉，要靠数据。]]></description><category></category></item><item><title><![CDATA[AI 原生软件工程时代，研发管理平台该如何进化]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/162962307</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/162962307</guid><author>simayi2018</author><pubDate>Fri, 17 Jul 2026 11:22:01 +0800</pubDate><description><![CDATA[半年后的复盘会上，你能指着一条上扬的 Prompt 曲线、一组越来越高的高价值使用占比，说清团队怎样被 AI 改变、下一笔钱往哪投。这，就是 AI 原生软件工程时代，研发管理平台该有的进化方向。]]></description><category></category></item><item><title><![CDATA[Claude Code API Key 怎么配置？新手从本地配置到多设备同步完整教程]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/162703812</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/162703812</guid><author>simayi2018</author><pubDate>Wed, 08 Jul 2026 18:41:17 +0800</pubDate><description><![CDATA[Claude Code API Key 配置，短期看是填几个字段，长期看是维护一份可靠的模型服务配置。只用一台电脑、临时试用时，手写环境变量可以；多设备、多模型、多 Agent 场景下，把配置集中管理并绑定到 Claude Code，会更容易复查，也更不容易漏改。]]></description><category></category></item><item><title><![CDATA[思码逸研发效能分析工具：全面提升团队生产力]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/162700548</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/162700548</guid><author>simayi2018</author><pubDate>Wed, 08 Jul 2026 16:15:07 +0800</pubDate><description><![CDATA[知识密集型的研发工作天然难以量化，管理者往往只能依靠工时、代码行数等粗糙指标，无法看清团队真实的生产力水平。]]></description><category></category></item><item><title><![CDATA[AI 时代怎么做程序员的绩效考核？]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/162699827</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/162699827</guid><author>simayi2018</author><pubDate>Wed, 08 Jul 2026 15:52:26 +0800</pubDate><description><![CDATA[程序员的工作本质上是知识密集型的创造性劳动。有效的考核数据需要满足三个条件：客观性、可量化性和正向引导性。]]></description><category></category></item><item><title><![CDATA[Spec 该怎么写，才能让 AI 可靠执行？GEARS 格式详解]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/162688241</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/162688241</guid><author>simayi2018</author><pubDate>Wed, 08 Jul 2026 10:42:25 +0800</pubDate><description><![CDATA[GEARS：为 AI 原生软件工程设计的规约（spec）语法]]></description><category></category></item><item><title><![CDATA[多设备党必备：轻松配置我的 AI Agent 全家桶]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/162225656</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/162225656</guid><author>simayi2018</author><pubDate>Tue, 23 Jun 2026 10:29:50 +0800</pubDate><description><![CDATA[4 台设备、3 个 AI Agent、8 个 LLM 配置触点，手动管理就是噩梦。KeySync 让你一次配置，全设备同步。]]></description><category></category></item><item><title><![CDATA[User Prompt：AI 编程提效的真正标尺]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/162225583</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/162225583</guid><author>simayi2018</author><pubDate>Tue, 23 Jun 2026 10:28:00 +0800</pubDate><description><![CDATA[我们选择度量什么，本质上反映了我们在乎什么。如果在乎的是成本，那就盯 token——但请从模型端直接获取真实数据，避免"看似 token 实际并非 token"的中间态误导。如果在乎的是 AI 到底有没有让团队变得更强——user prompt，可能是目前最直接、最诚实、最能推动行动的答案。]]></description><category></category></item><item><title><![CDATA[AI 提效指标体系分三层，大部分人只看到第一层]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/161792820</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/161792820</guid><author>simayi2018</author><pubDate>Mon, 08 Jun 2026 11:08:12 +0800</pubDate><description><![CDATA[AI 没有推翻原有的研发模式，而是在其基础上提供了新的改进路径。为此，我们的度量体系也应该在保持原有框架的同时，为新的工作方式补充合适的指标。]]></description><category></category></item><item><title><![CDATA[证明 AI 提效，“相关代码占比“没那么重要]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/161516748</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/161516748</guid><author>simayi2018</author><pubDate>Fri, 29 May 2026 14:13:42 +0800</pubDate><description><![CDATA[当百分之八九十的代码都是 AI 写的，再去看是80%还是90%，意义确实不大了。]]></description><category></category></item><item><title><![CDATA[当多 Agent 连续开发数小时：迎接“人人都是技术管理者”时代]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/161394361</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/161394361</guid><author>simayi2018</author><pubDate>Mon, 25 May 2026 15:29:16 +0800</pubDate><description><![CDATA[Harness Engineering 的核心，就是面向可靠性去构建整个 agent 系统。

]]></description><category></category></item><item><title><![CDATA[代码度量分析实践：发现潜在问题，优化开发流程]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/161198901</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/161198901</guid><author>simayi2018</author><pubDate>Mon, 18 May 2026 17:30:36 +0800</pubDate><description><![CDATA[本文将围绕代码度量分析的实践，探讨如何借助工具和方法，有效提升研发团队的生产力与代码质量。]]></description><category></category></item><item><title><![CDATA[深度访谈：穿越技术周期的头部研发效能度量服务商]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/161198460</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/161198460</guid><author>simayi2018</author><pubDate>Mon, 18 May 2026 17:12:03 +0800</pubDate><description><![CDATA[本篇文章将深入探讨头部研发效能度量服务商思码逸，了解其如何凭借领先的度量分析平台 DevInsight 及其核心指标代码当量，帮助企业穿越技术周期，实现研发效能的显著提升。]]></description><category></category></item><item><title><![CDATA[未来已来：AI 驱动的研发效能平台新趋势]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/160987927</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/160987927</guid><author>simayi2018</author><pubDate>Mon, 11 May 2026 16:14:22 +0800</pubDate><description><![CDATA[研发不再是简单的“人机协作”，而是“人机共生”，AI 将成为我们提升研发效能、释放创新潜力的强大盟友。]]></description><category></category></item><item><title><![CDATA[选择适合你的研发效能平台：避坑指南与最佳实践]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/160987914</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/160987914</guid><author>simayi2018</author><pubDate>Mon, 11 May 2026 16:13:37 +0800</pubDate><description><![CDATA[如何清晰地看到这台“发动机”的运行状态？如何确保它在高速运转的同时，依然保持高效、健康？]]></description><category></category></item><item><title><![CDATA[AI 时代的研发效能度量体系：量化投入，优化产出]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/160987890</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/160987890</guid><author>simayi2018</author><pubDate>Mon, 11 May 2026 16:12:35 +0800</pubDate><description><![CDATA[未来所有的研发都将是数据驱动的。通过精细化的度量和智能化的分析，研发团队将能够更好地理解自身，持续优化，最终实现卓越的研发效能。]]></description><category></category></item><item><title><![CDATA[告别唯代码论：多元化程序员考核体系构建]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/160589346</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/160589346</guid><author>simayi2018</author><pubDate>Tue, 28 Apr 2026 12:29:05 +0800</pubDate><description><![CDATA[唯代码论，简而言之，就是将代码量（如代码行数 LoC、提交次数等）作为衡量程序员工作量和绩效的唯一或主要指标。]]></description><category></category></item><item><title><![CDATA[代码当量已死？AI 原生开发给出了相反答案]]></title><link>https://blog.youkuaiyun.com/simayi2018/article/details/160589218</link><guid>https://blog.youkuaiyun.com/simayi2018/article/details/160589218</guid><author>simayi2018</author><pubDate>Tue, 28 Apr 2026 12:23:11 +0800</pubDate><description><![CDATA[AI 一天写一万行代码之后，我们反而更该关心这件事。]]></description><category></category></item></channel></rss>