从一篇「AI 写代码该选什么语言」的讨论说开去。当 AI 既写代码又当审稿人，选语言的判据不是它对 AI 多熟、也不只是循环多快，而是一次绿灯替你排除了多少错——而最大的陷阱是，那盏灯要是 AI 自己点的，它什么都没排除。

没人审 AI 写的代码之后，语言该怎么选

存算一体、光互连、异质集成、新型存储介质。当 HBM 也堆不动的时候，计算架构会怎么演化。

存储墙之外：计算架构的下一站

When a wall gets this expensive, the industry attacks it from every side — tier through NAND, pool over CXL, ditch HBM, compute inside memory. None of them breaks the wall in 2026. Which tells you exactly where this cycle's turn is not going to come from.

Beyond the Memory Wall: The Workarounds, the Challengers, and the Turn

硅谷101 这期 AI-First 播客的听后感。多数人把 AI 当工具，跟 24 小时较劲，提速到头也就 10 倍；少数人把它当系统，重构整条工作流，人往后退到只做两件事——给意图、审产出。但我想说的是，审产出这道关，正在被自己淹没。

把 AI 当工具的人，还在和 24 小时死磕

Reactions to a podcast on AI-first org architecture. Use AI as a tool and you're racing the clock — 10x at best. Use it as a system and you rebuild the whole workflow, retreating to just two jobs: give intent, review output. But the review gate is drowning, and that's the part nobody wants to say out loud.

Tool or System: Why Most "AI-First" Tops Out at 10x

一个被 99% 的人忽视的事实：训练不是瓶颈，推理才是。推理需要内存，海量内存。而这个需求正在创造一个持续十年以上的半导体超级周期。

存储超周期：算力飞轮的另一面

Memory is the most cyclical business in tech — 40 years of up-10x, down-90%. This run has three genuinely structural differences. But 'this time is different' is the most expensive sentence in finance, so the bear case gets equal weight.

The Memory Supercycle: Why This Time Might Actually Be Different

我用 TypeScript 写过几个 web 游戏，但总觉得天花板太低。想认真做一个、让 AI 写大部分代码，到底该选 Godot、Unity 还是虚幻？查了一圈才明白：AI 最擅长的引擎，不是训练数据最多的那个，而是它能整个读进眼里、自己跑起来、自己看着改的那个。

想用 AI 做游戏，我研究完发现引擎选错了全白搭

I built a few TypeScript web games but always felt the ceiling was too low. If I wanted to build a proper game and let AI write most of the code, should I pick Godot, Unity, or Unreal? After a lot of digging the answer surprised me: the engine AI is best at isn't the one with the most training data — it's the one AI can read entirely as text, run headless, watch, and fix on its own.

The Best Game Engine for AI Is the One It Can Read

一张 H100 超过一半的成本是 HBM。它凭什么这么贵？TSV、硅中介层、垂直堆叠——以及为什么只有 SK 海力士能赚钱。

HBM：把内存架在芯片头顶

HBM isn't faster DRAM. It's DRAM rotated 90 degrees — stacked, drilled through with thousands of vertical wires, and bolted next to the GPU on a silicon interposer. That manufacturing nightmare is exactly why three memory makers just crossed a trillion dollars.

How HBM Works — the Memory That Decides Who Wins AI

The code for my games? Agents handle it, faster and more correctly than I would. The thing that stopped me cold was the story. I'm not a writer, so I did the only thing I know how to do — turned "how do you write a game story" into a research problem, tore the craft apart, and built a skill that walks you from a blurry idea to a scene you can actually stage. This is the process, and the genuinely useful craft I dug up. AX Skill Workshop, part 2.

AI Wrote My Game's Code. It Still Can't Write One Line of Dialogue That Isn't Cringe.

一具坐在金色王座上的尸体，已经死了一万年。把它送上王座的，是它最宠爱、最信任的那个儿子。这是战锤40K 最像悲剧的那段往事——也是这一万年黑暗真正的源头。

他最爱的儿子，亲手把银河砸碎了

你的 SSD 为什么断电还能存数据？因为 NAND 用物理方法把电子锁在绝缘层——这个简单的原理构成了全球每年数百亿美元的存储市场。

NAND 闪存：把电子关在绝缘层里

The same chip lives in your USB stick, your phone, and the petabyte SSDs feeding AI clusters. It got cheap by trading endurance for density — and that exact tradeoff is why AI is now dragging it out of cold storage and onto the hot path.

How NAND Flash Works — From Your SSD to the AI Data Center

Ten thousand years before the grimdark present, the Emperor of Mankind's favorite son turned a galaxy-spanning crusade into a civil war that has never truly ended. This is the tragedy at the heart of Warhammer 40,000 — and why its wound still bleeds.

The Horus Heresy: How a Father's Greatest Son Broke the Galaxy

一个电容、一个晶体管。过去五十年几乎所有计算的内存都建立在这对搭档上。

DRAM 是怎么存下一个比特的

DRAM is a leaky bucket you refill thousands of times a second. That physical fragility is why it's fast, why it's volatile, and why it just hit a scaling wall at the exact moment AI demand exploded — sending RAM prices up triple digits.

How DRAM Works — and Why It Got Scarce

六千万年前，两个早就死透的神族在银河里互砍。四万年后，一个复活的半神还在打圣战。中间这条线怎么连起来——而且为什么连日期本身都在故意骗你。

六千万年的战争史，一条时间线讲清楚

From two dead god-races fighting over the galaxy 60 million years ago to a resurrected demigod waging holy war in the present — the full history of the grimdark future, in order, with the dates that lie to you on purpose.

60 Million Years of War: The Complete Timeline of Warhammer 40,000

处理器的速度是指数增长的，内存带宽是线性增长的。这个裂缝是过去二十年计算架构最根本的瓶颈——而你每天都在感受它。

存储墙：当计算等不起内存

I finance the machines AI runs on, and the single priciest, scarcest part inside the box isn't the logic die — it's the memory stacked next to it. This is why: inside a modern accelerator, raw FLOPS stopped being the binding constraint. Memory bandwidth did.

The Memory Wall

在第四十一个千年，没有好人，没有救援，结局那页也没有和平条约等着你。这篇讲清楚：战锤40K 的银河，从结构上就停不下来——而这恰恰是它的全部用意。

为什么这场战争永远打不完

In the 41st millennium there are no good guys, no rescue, and no peace treaty waiting at the end. Here is why the galaxy of Warhammer 40,000 is structurally incapable of stopping the war, and why that is the entire point.

Why the War Never Ends: The Core Conflicts of Warhammer 40,000

一个垂死的神权帝国，四位用情感凝成的噩梦邪神，外加一打坚信自己才是主角的外星种族。这篇把战锤40K 银河里所有出场的狠角色，一个个介绍给你。

银河里到底有几伙人在打仗

A galaxy at total war needs combatants worth fearing. Here is the cast — a dying god-empire, four nightmare powers wearing reality like a glove, and a dozen alien races that all believe they are the protagonists.

Who Fights in the 41st Millennium? A Field Guide to the Factions of Warhammer 40,000

一百万个世界，一位已经死了一万年的神。在第41个千年，你能说出的最乐观的话，就是人类还活着——而这恰恰是最糟的部分。

一个由垂死之神统治了一万年的银河：战锤40K 到底是什么

A galaxy of a million worlds, ruled by a god-emperor who has been dying for ten thousand years. Welcome to the 41st millennium, where the most optimistic thing you can say is that humanity has survived — and that may be the worst part.

The Grimdark Galaxy: An Introduction to Warhammer 40,000

2015 年，Games Workshop 干了一件几乎没人敢干的事：把自己养了三十年的奇幻世界真的玩死了——星球裂成两半，神死的死逃的逃。这是旧世界怎么塌的，也是废墟里长出了什么。

一家公司亲手炸掉了自己的世界

In 2015, Games Workshop did something almost no franchise dares to do: it ended its 30-year-old fantasy world for real — cracked the planet in half and let it die. Here is how the Old World fell, and what rose from its ashes.

The Day a Company Murdered Its Own World: Warhammer's End Times

帮非职业创作者从零打造游戏剧本的 AI skill。世界观、角色、主线支线、场景脚本——全流程陪伴式创作。

Skill Workshop

战锤奇幻的旧世界，南北极是两个永远堵不上的伤口，每座城市的地板下都住着鼠人。它打的不是一场仗，是六场。这篇讲清楚它为什么注定输——以及为什么这反而是它最动人的地方。

一个从一开始就注定打不赢的世界

The Old World was never going to win. From daemon-haunted poles to a rat empire under every floor, here is why Warhammer Fantasy is a story about civilizations fighting entropy with one hand tied behind their back.

A World Under Siege: How the Old World Lost Its War

矮人把每一笔旧仇都记进一本永远翻不到底的书，鼠人多到能淹没整个地表却忙着自相残杀，精灵因为一个没坐上的王座分裂成两族打了几千年。带你认全这个垂死世界里的所有玩家——以及让他们到死都不肯联手的那些仇。

一本写了四千年还没写完的仇恨账本

Dwarfs who write their grudges in a book that never ends. Rats who deny their own existence. Elves split by a single denied throne. A tour of the powers fighting over a dying fantasy world — and the hatreds that keep them from ever uniting against the thing that will kill them all.

Order and Destruction: A Field Guide to the Factions of the Old World

一个从来没存在过的中世纪欧洲，八股看不见的魔法之风，世界尽头天上的一个洞。这是旧世界——Games Workshop 造了三十多年的奇幻设定，2015 年自己把它毁掉了。

他们花了32年造一个世界，然后亲手把它炸了

A dark-medieval Europe that never was, eight invisible winds of magic, and a hole in the sky at the top of the world. This is the Old World — the fantasy setting Games Workshop built over thirty years, then deliberately blew up in 2015.

The World That Was: An Introduction to Warhammer Fantasy

把半仙占卜做成 skill。梅花易数、小六壬、六爻——用户描述事情，选卦法，出真卦象，半仙化解读。一个 skill，三种东方占卜体系。

Six months of accumulated Claude Code configuration — 50 skills, 50 subagents, 40 commands, hooks, rules, and the sync pipeline that pushes all of it to a public GitHub mirror with a fail-closed two-tier audit. What's inside, why I built it, and the workflow philosophy behind it.

I Open-Sourced My Entire ~/.claude

AX-skills is the public mirror of my personal Claude Code skills directory — eleven workflows for writing, research, image generation, knowledge sync, and one cyber half-immortal divination skill. After a year of writing them, I think a skill isn't a stored prompt — it's how you encode care.

I Open-Sourced the 11 AI Skills I Use Every Week

Last week I surveyed 8 open-source meeting recorders. My friend tried every one. Verdict: ASR is bad, summaries are bad, UI/UX is bad. So I dug deeper. The real problem is structural — every project I recommended is built on Whisper, and Whisper has been outpaced by 2-4× on Chinese benchmarks for over a year. This is the deeper survey: data, the issues that maintainers themselves admit, the dumb-but-effective hack everyone overlooked, and four projects I missed the first time.

I Recommended Eight Open-Source Meeting Recorders. My Friend Returned All of Them.

A friend asked for a meeting recorder. Zoom/Teams native AI is unusable. Granola and Krisp work but cost ~$200/yr per seat. He's on Google Workspace but doesn't use Meet, so the bundled Gemini summarizer is dead weight. I spent a night on GitHub and found a whole constellation — from Meetily at 11.6k stars down to single-digit-star personal toys, all racing to commoditize the same three Lego pieces.

Zoom's Native AI Sucks. So I Surveyed Every Open-Source Meeting Recorder

Took the three Eastern divination engines from PanPanMao and packaged them into a single agent-portable skill — algorithm engine + knowledge base + persona prompt, three folders deep. The pattern transfers to any product with real domain knowledge.

Packaging a Product into a Skill: A Cyber-Banxian for an Agent Contest

Part 2 的「6 槽位通吃 7 品类」是一层漂亮的抽象。跑了两周，测出牛仔外套生成「宝石星图」的 bug。这一篇讲抽象失守在哪一层、为什么没用户时分阶段上线是假安全感、以及把服饰和美妆各自推倒重来的过程。

从万物回到三个品类

Part 2 claimed six slots could cover seven categories. Two weeks later, a denim jacket generated a gem constellation. This post walks through where the abstraction failed, why skipping the 'safe' phased rollout was the right call, and how each of three categories got its own six-template set.

Walking Back the Abstraction

Solo dev 做设计，大多数时候 N=1——第一直觉就是唯一方案，因为做第二个版本要花一天。Claude Design 把这事拆了：一下午六个完全不同的视觉语言摆在面前。留下了什么、丢了什么、方向锁死之后 big-task 的 ui profile 接手怎么走——这篇讲这些。

一下午让 AI 出六个设计方向，solo dev 的设计终于不用将就

For solo dev and small teams, design exploration is usually N=1 — you go with your first instinct because producing a second variant takes a day. Claude Design collapses that: six distinct aesthetic directions in an afternoon. What I kept from the pipeline, what I dropped, and how `big-task`'s UI profile takes over once a direction is locked.

Six Directions in an Afternoon: Using Claude Design to Kill First-Instinct Bias

先是裸跑 Claude Code。再是 GSD 多 agent 仪式。再是 superpowers 的 skill 库。最后在上面包了一个自己的 orchestrator。每一版崩盘都来自一个具体 session。最后我才想明白：solo dev 的 harness 跟团队的不是一个物种，而且真正的答案不是选一个默认引擎——是按工作形状路由。

两个月，我把 AI 工作流推倒重来了四次

Vanilla Claude Code. Then GSD's multi-agent ceremony. Then superpowers' skill system. Then a custom orchestrator on top. Each rewrite came from a specific session that broke. What I learned: solo dev harnesses and team harnesses are different species, and the right answer isn't a default engine — it's routing by work shape.

I Rewrote My AI Coding Harness Four Times in Two Months

同一条流水线——Claude Design 出方向、HANDOFF.md 锁死、`big-task` 落地——但每个产品都把它掰成不一样的形状。ÉLAN 是空地，需要宽探索。识川 app 已经上线，要按类目一个个迁移。盘盘猫是 monorepo，v2 跑三个并行品牌表达而不是一个方向。变的是什么，不变的是什么。

同一套 AI 设计流水线，三个产品跑出三种形状

Same pipeline — Claude Design explores, HANDOFF.md locks, `big-task` ships — but it bends around each product. ÉLAN is greenfield and needs wide exploration before converging. Shichuan has a live app and needs incremental category-by-category migration. Panpanmao is a monorepo with a v2 redesign that runs three parallel brand expressions instead of one. What changes and what doesn't.

One Pipeline, Three Shapes: Running Claude Design + big-task Across ÉLAN, Shichuan, and Panpanmao

AX 的 AI 雷达每天从 50 多个信源抽约 500 条内容，挑 15 条给我。选题不靠 prompt，靠一份能版本化、能 diff 的 Markdown 编辑策略——一个 Agent 从反馈里迭代它。讲这把雷达的架构取舍和关键设计决策。

每天 488 条 AI 新闻，我让一份 Markdown 替我选题

AX Radar filters ~500 stories a day from 50+ sources down to about 15. Editorial judgment isn't a prompt — it's a versionable, diff-able Markdown file that a Claude Agent iterates against user feedback. A breakdown of the architecture and the design calls.

488 AI Stories a Day: Why My Editorial Policy Is a Markdown File, Not a Prompt

上一篇讲的是我用 AX 雷达的一手体验。这一篇讲我给它加了 HTTP API、MCP server 和一份 Claude Skill——让 Claude 自己能查、能搜、能保存。两张脸共用一个后端；sha256 不是 bcrypt；pgvector 早就在了；Skill 才是放 domain 知识的地方。

雷达加了第二张脸——这回是给 Agent 看的

Part 1 covered AX Radar from my side as the operator. Part 2 covers what I shipped this week: an HTTP API, an MCP server, and a Claude Code skill that let Claude read, search, and mutate the radar directly. Two surfaces, one backend. Why sha256 beats bcrypt here. Why semantic search rode the existing pgvector index. Why the skill is where domain knowledge lives.

AX Radar Needed a Second Face. One for Agents.

想跑团，凑不齐人。想玩剧本杀，差两个。想来一局狼人杀，只有四个人在线。如果 AI agent 能坐进来补位——不是那种复读机式的 NPC，是真的有人设、会推理、能演戏的 agent——会怎样？

凑不齐人的时候，AI 能不能顶上

想让几个 AI 坐一起辩论或来一局狼人杀？GitHub 上有二十几个相关项目。框架那头星数不少但没有'房间'概念，游戏那头全是一次性 hack。中间缺一个平台层——而搭好这个平台层之后你会发现，做好多 Agent 游戏和做好多 Agent 协作，用的是同一套东西。

我调研了 20 个多 Agent 项目，没一个能直接用

AgentScope 两万三千星，阿里出品，agent 抽象干净，记忆系统完善。fork 它搭一个多 Agent 平台，看起来是最省力的路。但跑完八轮自我反驳，结论是：不 fork，自己搭——但系统性地偷它的设计模式。

面对两万星的框架，我选择不 fork

不 fork 不等于不学。从 AgentScope 偷 reply/observe 双接口和消息隔离模式。从 ChatArena 偷三层架构。从斯坦福 Generative Agents 偷记忆反思循环。从 Accio Work 偷'建群聊'的交互隐喻。每个核心模块都有明确出处，但代码是自己写的。

偷取清单：从 20 个项目里拆出一个平台

Agora 跑通了第一个 MVP：6 个 AI debater + 3 个 AI judge，三个模型各出两个人设，辩了三个话题。预期是及格就好，结果 Claude 引亚里士多德，GPT 举 Log4Shell，Gemini 演了个愤怒的 17 岁创作者。63 次 LLM 调用零失败。

我让 6 个 AI 辩了三场，有一段我反复看了三遍

Agora 的第二个模式上线了。9 个 AI agent 打了一局标准预女猎狼人杀，频道隔离让狼人能私聊但村民看不到，盲投保证没人跟票，状态机驱动夜晚→白天→投票的完整循环。三局下来最让我震惊的不是技术跑通了，是 Claude 扮演的狼人在投票理由里写了一整段阴谋论述——而且逻辑完全自洽。

三局 AI 狼人杀，狼人在投票理由里写了自己的阴谋

Phase 2b 上线：守卫、警长、白痴、遗言、猎人连杀，全部作为可开关的进阶规则。12 人局里，AI 守卫第一晚选了「今晚不守」——它的理由是要避免和女巫撞救。一个 AI 在思考两个角色之间的规则交互。这段话我反复看了两遍。

狼人杀加了守卫和警长，AI 居然想明白了规则冲突

有人用 AI 两小时做了个游戏。我决定做三个。不写代码，只当制作人。第一个游戏的概念阶段，AI 写出了比我更好的剧本。

我不写代码，但我要做三个游戏

Someone made a complete puzzle game with AI in two hours. I'm making three — as producer, not developer. In the first concept session, the AI designed a better story than I could.

I Don't Write Code, But I'm Making Three Games

Dead Signal 是第二个游戏。一个调查记者，一部连环杀人案受害者的手机，72 小时的碎片数据。恐惧不来自画面，来自数据。

手机里的最后72小时

Dead Signal is the second game — an investigative journalist reconstructing a murder victim's final 72 hours from fragmented phone data. The horror comes from data, not visuals.

The Last 72 Hours on a Dead Person's Phone

Shadow Access 翻转了 found-phone 游戏的叙事。你翻的不是别人的手机，是自己的。一个干净的 iOS 界面慢慢被「感染」，你要找的不是线索，是异常。

凌晨三点四十七，有人在用你的手机

Shadow Access flips the found-phone genre: you're not investigating someone else's device — you're discovering what someone did inside yours. An iOS interface that slowly gets 'infected' as you uncover the intrusion.

At 3:47 AM, Someone Was Using Your Phone

看完 Trae 的 MTC 模式，我在 Claude Code 里做了个 /mtc skill。6 个阶段，不许跳步，先想清楚再让 AI 写代码。做游戏的时候效果比我预期好。

给 Claude Code 加了个 skill，强迫它先想清楚再动手

After seeing Trae's MTC mode in action, I built a /mtc skill for Claude Code. 6 phases, no skipping, think first, code second. Better results than expected on my game project.

I Added a Claude Code Skill That Forces Planning Before Code

把斯坦福 AI Index 2025 和 2026 两份公众舆论报告摆在一起看，两件事清楚了：一是造 AI 的那批国家反而是最不信 AI 的，二是真正把 AI 用起来的那批国家，不是西方科技媒体关注的那几个。热情地图和能力地图几乎不重叠，差距还在拉大。

做 AI 最多的那个国家，最不信 AI

Stanford HAI's 2025 and 2026 AI Index reports, side by side, show that the countries building AI are the ones most afraid of it — and the countries using AI most intensively at work are not the ones the Western tech press talks about. The sentiment map and the capability map don't overlap.

The Inverted AI Adoption Map

所有 AI 都在抢着帮你干活，但没有一个陪你。我做了一个住在桌面上的 AI 宠物——它能看见你的屏幕，记得你昨天熬夜，但它不会帮你写代码。

AI 助手够多了，我想养一只宠物

Every AI tool is competing to help you work. None of them remember you. I built a desktop pet with a soul engine — it sees your screen, remembers you were up late yesterday, and won't help you write code.

Enough AI Assistants. I Want a Pet.

humor: 0.8 对模型来说毫无意义。我们用散文写角色圣经——五个性格，每个六十行中文。从此它知道怎么损你、怎么关心你、怎么假装不在乎。

给 AI 写性格，不能用数字

humor: 0.8 means nothing to a language model. We wrote character bibles in prose — five personalities, sixty lines each. Now it knows how to tease you, care about you, and pretend it doesn't.

You Can't Define Personality With Numbers

AI 宠物没有记忆就是陌生人。三层存储、混合检索、每晚做梦整理记忆——再加一个八层 prompt 引擎把记忆注入每次对话。这是让'你上次说喜欢吃火锅'成为可能的全部工程。

它记得你昨天也熬夜了

An AI pet without memory is a stranger every time. Three-tier storage, hybrid search, nightly dreaming — plus an 8-layer prompt engine that turns memories into conversation. This is the full engineering behind 'you mentioned you like hotpot.'

It Remembers You Were Up Late Yesterday

LLM 被训练成有用的。但朋友不会看到你的代码说'我注意到你在用 React hooks'——朋友会说'又在写 bug 啊笨蛋'。让 AI 停止帮忙，是做 AI 宠物最难的一关。

让 AI 学会"没用"，比让它有用难十倍

LLMs are trained to be helpful. But a friend doesn't see your code and say 'I notice you're using React hooks' — a friend says 'writing bugs again, dummy.' Suppressing the assistant instinct is the hardest part of building an AI pet.

Making AI Useless Is Harder Than Making It Useful

Electron 壳里零行 AI 代码，灵魂引擎只有两个 npm 依赖。身体和大脑分开的架构决策，以及 2000 行代码撑起整个系统的极简主义实践。

两个仓库，一个灵魂

Zero lines of AI code in the Electron shell. Two npm dependencies in the soul engine. The architecture decisions behind separating body from brain, and how 2,000 lines of code power the entire system.

Two Repos, One Soul

员工手册就是给人类写的系统提示词。MrBeast、Netflix、Duolingo 三家百万人效公司的手册都在解决同一个问题：怎么把品味传递出去。但给人和给 AI 写系统提示词的区别比想象中大——人需要激励、需要归属感、有自尊心要哄、会遗忘会走偏；agent 一样都不需要。

人效百万美元的公司，员工手册里写了什么

MrBeast wrote 36 pages. Netflix distilled theirs to 5. Duolingo designed a 64-page illustrated book. All three generate over $1M per employee. All three are solving the same problem: how to transfer taste, judgment, and operating principles to another entity so it can execute autonomously. These handbooks are CLAUDE.md files for humans — and they reveal three distinct archetypes for how organizations might scale in the agentic AI era.

CLAUDE.md for Humans

上一篇说员工手册是给人类写的 CLAUDE.md。这篇接着问：AI 原生公司的手册该怎么写？该招什么人？怎么筛选品味而不是技能？从 Compute Labs 三人团队的经验出发，试着写出 AI 原生公司的员工手册——一份给有品味的人的操作指南。

AI 原生公司招人不看简历

If employee handbooks are CLAUDE.md files for humans, what should the handbook of an AI-native company look like? What kind of people does it hire — and how does it screen for taste instead of skill? This piece designs the AI-native employee handbook from scratch: the talent profile, four screening methods, and the five sections every AI-native handbook needs — from mission to judgment frameworks to agent collaboration protocols.

The AI-Native Employee Handbook

你的优势是你的问题。二十五年真正会复利的不是信息量，是问出更好问题的能力。最后一个 workshop 是全课程唯一不用 agent 做的。

从术到道

Your edge is your questions. What compounds over twenty-five years is not information — it is the capacity to ask better questions. The final workshop is the only one you do without an agent.

From Technique to the Way — Your Agent Is a Mirror

大部分人只拿 agent 当执行者用。投资场景里 agent 有四个角色——分析师、红队、导师、执行者——你提问的方式决定你激活哪一个。

你的 Agent 有四份工作

Most people use their AI agent in one mode. Investment agents play four distinct roles — analyst, red team, tutor, executor — and the way you frame your request determines which role you get.

Your Agent Has Four Jobs, Not One

股票是对未来现金流的权利——不是代码，不是图表。三种回报来源，五种优势类型，以及你在判断 agent 输出之前必须有的词汇。

一只股票到底是什么

A stock is a claim on future cash flows — not a ticker, not a chart. Three sources of return, five types of edge, and the vocabulary you need before you can judge an agent's output.

What Is a Stock, Really — First Principles

年报不用每一行都读。你要认的是模式。三张报表当三个镜头——业务结构、脆弱度、真假。

读一家公司：读模式，不数页码

You do not need to read every line of a 10-K. You need to recognize patterns. Three financial statements as three lenses — business structure, fragility, and truth.

Reading a Company — Patterns, Not Pages

DCF 里每个数字都是一个故事选择。Agent 算得又快又准，但它默认讲共识。你的工作是带自己的故事进去——以及找到估值暗中依赖的那个假设。

估值是一个被数字包着的故事

Every number in a DCF is a story choice. Agents compute flawlessly and default to consensus. Your job is the narrative — and knowing which assumption your valuation secretly depends on.

Valuation Is a Story Wrapped in Numbers

风险是你承受不起的亏损的概率，不是标准差。LTCM 1998 年优化了错误的指标，被相关性崩溃杀死了。四种真正的风险，以及为什么只有你能决定自己亏得起多少。

风险不是波动率

Risk is the probability of a loss you cannot afford — not standard deviation. LTCM optimized for volatility and was killed by correlation breakdown. Four real risks, and why only you can decide what you can afford to lose.

Risk Isn't Volatility

大部分投资者有的是信念，不是论点。把判断写下来、附上止损条件，是对自己记忆的问责——因为记忆会自我美化，信念会悄悄变形。

没写下来就不算判断

Most investors have beliefs, not theses. A written thesis with kill criteria is accountability against your own memory — because memory is self-serving, and beliefs mutate silently.

Days 20-23: Shipping M1, M2, M3 in Four Days

Days 12-19: Credit Economy, New Features, and the Feature Explosion

Days 5-10: From Code to Product (Testing, Landing Page, and the Cat Theme)

Days 2-4: The Monorepo Migration Sprint (111 Commits in 3 Days)

Day 1: Why I'm Building a Fortune-Telling App (And I Know Nothing About Fortune-Telling)

From Rowers to Navigators: Thriving in the Age of AI

Managing AI Like a Software Development Manager: A New Paradigm for AI-Assisted Programming