整理 | 褚杏娟 Clawdbot（现名：Moltbot）火了到国内，社交平台上到处都是部署教学、使用教学和使用展示。国 内的腾讯云、阿里云等也相继宣布上线 Clawdbot 云端极简部署及全套云服务，钉钉也在 Github 上 开源了 Moltbot 接入方式。 项目背后的创始人 Peter Steinberger 也红极一时，他的构建方式成为很多人的学习对象。Peter 之 前就是一位非常出色的开发者，打造了一个被用在超过十亿台设备上的 PDF 框架。后来他经历了严 重的职业倦怠，卖掉股份，整整三年从科技圈消失。今年，他回来了，而他现在的构建方式、正在做 的事情，已经和传统软件开发完全不同。 Peter 近期在"The Pragmatic Engineer"节目中，用近两个小时的时间分享了他的开发经历。他解释 了，为什么他现在发布的代码，大部分自己都不再逐行阅读，而这其实并没什么大不了；他具体是如 何打造了 ClawdBot 这个看起来就像 Siri 未来版本的个人助手的；他如何利用"闭环原则"，高效进行 AI 编程；为什么代码评审已经过时，PR 应该改名叫 Prompt Request 等，他还分享了很 ...

"One of the biggest challenges is making sure you have the correct feedback loop even when people only want to tell you what you want to hear." — Elon Musk https://t.co/VSmF58CBkp ...

Core Product Development Principle - Shipping something to someone always wins, emphasizing rapid iteration and feedback loops over big launches [1][34] - The key is enabling rapid iterative loops to get feedback from real users and maximize shots at the goal [1] - In the age of AI, this translates to building a "skateboard" first, then evolving it to a "car," ensuring a continuously viable product [2][4] - A continuously viable solution is significantly more valuable because it provides feedback along the way, avoiding building in a vacuum [5][6] Feedback Loop Implementation - Establish a feedback loop with real users who can see something, provide feedback, and allow for iterative improvements, ideally within a day [7] - Being able to ship every day is crucial for a fast feedback loop, requiring specific focus on the target customers [9] - Work with real people (not just personas) to understand their problems and build empathy [10][11] - Write the PI (Product Information) FAQ or launch blog post early to sanity check and communicate the product effectively [12] Navigating Constraints and AI Integration - Design the best product first, before considering constraints like legal, compliance, and financial aspects [15] - AI accelerates all aspects of product building, but the fundamental process of talking to users and getting feedback remains the same [26] - Product management becomes more critical as the cost of writing code approaches zero, emphasizing customer knowledge and rapid feedback [28][29]