A small story: I gave my note folders recognizable icons so I can tell them apart at a glance. But the real lesson isn't the feature — it's why it was me, not the AI, who first noticed the friction.

00:09  My note workflow: one folder per note
00:52  Folder names use the note ID: unique and lean
01:28  The pain: a wall of number IDs, hunt folder by folder
02:05  The spark: give the folder an icon you recognize at a glance
02:37  Icon design: a visual plus the title, instant ID
03:05  Building it with AI: speak the idea, settle the plan, it executes
03:36  Collaboration snags: no Finder access, font too small
04:09  The real insight: the change is trivial, spotting the problem isn't
04:42  The loop: from spotting to solving, the human starts it
05:07  A reminder for everyone: spotting problems is yours to train

这期从一个真实的小功能讲起：我给一堆数字 ID 命名的笔记目录，换上了能一眼认出的图标。但真正值得聊的不是这个功能本身，而是——为什么是我，而不是 AI，先看到了那个「找起来好麻烦」的问题。

00:08  我的笔记工作流：每篇一个目录
00:45  目录名用笔记 ID：唯一又精简
01:17  痛点：一串数字 ID，找产物得逐个翻
01:51  灵光一现：给目录换个一眼认出的图标
02:20  图标设计：视觉图加标题，一眼认出
02:44  和 AI 一起落地：口述想法、定方案、它执行
03:10  协作小坑：够不到 Finder，字体太小我让它调大
03:41  关键认知：功能改动不重要，发现问题才重要
04:09  闭环：发现问题到共同解决，人才是起点
04:35  给所有人的提醒：发现问题得自己练

I deleted every script, JSON config and validator from my AI image pipeline — and the results got dramatically better. This episode is about the deeper philosophy of working with AI: restraint.

00:07  Wrapping AI in scripts and validators made its images worse

01:23  Automation strangled the model's artistic ability

02:27  Deleting the rules unlocked the AI

03:37  Claude Code cut 80% of its system prompts

04:31  The engineer's habit of constraining everything

05:37  Work with AI as a partner, not a commander

06:46  No system is 100% controllable

08:01  Too many constraints create complex systems and surprises

09:08  Two rules: don't constrain what you don't master, keep systems simple

我把给 AI 出图的那套脚本、JSON、校验器全删了，AI 反而画得更好。这期聊聊跟 AI 合作的底层哲学——克制：真正的高手，懂得在不懂的地方放手。

00:08  给 AI 套脚本和校验器，出图反而越来越呆板
01:20  自动化把大模型的艺术创作能力抹杀掉了
02:18  删掉规则后 AI 自由发挥，效果反而惊艳
03:16  Claude Code 适配时删掉了 80% 的系统提示词
04:07  软件工程师的「代码约束一切」通病
05:07  跟 AI 是合作不是指挥
06:02  系统不可能 100% 可控
06:56  约束太多会变成复杂系统，产生涌现行为
07:47  两条正解：别在不擅长处加约束，系统做简单

Everyone is talking about Loop Engineering and Graph Engineering, but I'm still using the simplest way to work with AI. In this episode, I explain why — and why I believe there are no shortcuts for ordinary people.

00:08 I use AI agents every day, but I've never touched Loop Engineering

00:57 Why not? My use case demands Human in the Loop

01:46 Why not chase it? Big names have use cases, free tokens, and practice

02:34 Learning isn't doing — the senior's code doesn't write itself for you

03:20 Practice hasn't reached critical mass — not enough reps yet

03:52 First principles — start simple, switch only when it breaks

04:39 Master the present — practice creates experience others can learn from

05:25 No shortcuts — keep practicing from scratch is the only solid road

当全网都在讲 Loop Engineering、Graph Engineering 时，我还在用最笨的方法跟 AI 协作。这期聊聊为什么——以及为什么我相信，普通人没有捷径可走。

00:08 我天天用 AI Agent，却没碰过最火的 Loop Engineering

00:38 为什么用不上？我的场景必须 Human in the Loop

01:18 那为什么不追？大佬有场景、有免费 token，更有一线实践

01:55 学到不等于做到：高手写得再好，代码也得自己一行行练

02:30 实践还没到质变：频次、数量不够，谈不上升华

02:57 第一性原理：从最简单的方法开始，走不通再换

03:32 把当下研究透，才能不一直跟在别人后头

04:09 没有捷径：keep practicing from scratch 是唯一的康庄大道

本期从一场关于机械厂的梦说起——作者干过钳工、画过图纸，突然发现写代码和造机器是同一个道理。AI 来了，编程像数控机床一样被自动化，但真正值钱的，是留在人脑里的那张「图纸」。

00:08  梦里回到机械厂：我干过钳工，现在写软件
00:41  第一层类比：写代码像装配机器，设计像画图纸
01:17  岗位映射：工程师等于设计人员，编程等于零件制造加装配
02:02  数控机床类比：钳工的手艺活，被机器自动化了
02:28  机器做的也有瑕疵：零件要二次加工，AI 代码也要人盯
02:51  核心：理解图纸，才是自动化时代人的价值
03:16  警惕 AI 万能论：没有洞见，就形不成正反馈

本期从一场关于机械厂的梦说起——作者干过钳工、画过图纸，突然发现写代码和造机器是同一个道理。AI 来了，编程像数控机床一样被自动化，但真正值钱的，是留在人脑里的那张「图纸」。

00:08  梦里回到机械厂：我干过钳工，现在写软件

00:41  第一层类比：写代码像装配机器，设计像画图纸

01:17  岗位映射：工程师等于设计人员，编程等于零件制造加装配

02:02  数控机床类比：钳工的手艺活，被机器自动化了

02:28  机器做的也有瑕疵：零件要二次加工，AI 代码也要人盯

02:51  核心：理解图纸，才是自动化时代人的价值

03:16  警惕 AI 万能论：没有洞见，就形不成正反馈

我把两个内容 skill 折腾了一遍，发现一个有点反直觉的事实：AI 再聪明，也只会「局部优化」——它从不会主动把一个 skill 的功能挪到另一个 skill 去。这期聊聊，为什么系统级的熵减，目前还得靠人。

00:08  1. AI 再聪明，也不会自己把活分出去

01:02  2. 模块化的真实取舍

01:42  3. 系统性缺陷：只会局部优化

02:27  4. 为什么：大模型的上限

03:08  5. 落到 coding agent

03:56  6. 熵减为什么难度量

04:38  7. 未来的指望：度量熵

05:17  8. 连前沿团队都在对抗熵增

I built two content skills and watched one bloat. The counterintuitive takeaway: no matter how smart the model, it only does local optimization — it will never move a function to where it belongs. This episode is about why entropy reduction in systems still needs a human in the loop.


00:07  1. Smart as it is, it won't reassign its own work
01:12  2. The real trade-off of modularization
01:59  3. The systemic flaw: local optimization only
02:48  4. Why: the ceiling of today's models
03:33  5. Down to coding agents
04:26  6. Why entropy is hard to measure
05:21  7. The future hope: measuring entropy
06:02  8. Even frontier teams fight entropy