I’m going to show you, step by step, how to turn Fable 5 into a machine that knows your business inside out… and ships outputs that look nothing like what everyone else is getting 我将逐步向你展示，如何把 Fable 5 变成一台对你的业务了如指掌的机器……并产出与他人截然不同的成果。

the tool is a second brain built in Obsidian, and a small group of people is already running one… exactly the same model, a world of difference in every output 这个工具是在 Obsidian 中构建的第二大脑，已经有一小部分人在运行它……完全相同的模型，每个输出却天差地别

the smartest model on the market ships average work all day for one reason: it knows nothing about you 市场上最聪明的模型整天都在输出平庸的作品，原因只有一个：它对你一无所知

no context on your business, your audience, your past decisions… so it guesses, and guesses read generic 不了解你的业务、你的受众、你过去的决策……所以它只能猜，而猜出来的东西读起来就很空泛

plug it into your own knowledge base and the same model becomes a different machine 将其接入你自己的知识库后，同一个模型就变成了不同的机器

the code follows your architecture, the content sounds like you, the writing stands on research you own… and you see it from day one 代码遵循你的架构，内容听起来像你，写作基于你所掌握的研究……而从第一天起你就能看到这一切。

and this holds for ANY workflow you run: coding, marketing, content, sales, research 这对你运行的任何工作流程都适用：编程、市场营销、内容创作、销售、研究。

running an AI agent without a second brain is wasting your time… and the gap only widens, because every file the brain gains makes every future run smarter, forever 在没有第二大脑的情况下运行AI代理是在浪费时间……而且差距只会越来越大，因为大脑每获得一个文件，都会让未来的每一次运行变得更聪明，且效果持久。

i run this exact machine every day… it sits behind every article, every guide, and every product i ship 我每天运行这套机制……它支撑着我发布的每一篇文章、每一份指南和每一个产品。

this is the full system… what a second brain really is, the folder structure the agent navigates on its own, how to fill it with goals, keep it alive with loops, run a real research machine on top of it, read it without burning money, and wire it into everything you build 这就是完整的系统……第二大脑的真正含义、智能体自主导航的文件夹结构、如何用目标填充、通过循环保持其活力、在其上运行真正的研究机器、无需大量消耗即可读取、并将其与你构建的一切相连接

if you want to learn how to get the most out of Fable 5 and how to make money with it, that’s what the real time AI ops community is for: weeklyaiops.com如果你想学习如何最大化利用 Fable 5 并以此盈利，那就加入实时 AI 运维社区：weeklyaiops.com

same model, different league 同款模型，不同层次

here are the numbers behind that claim 以下是这一论断背后的数据

in accounting, a model working without the client’s history lands around 70% accuracy 在会计领域，不使用客户历史数据的模型准确率约为70%

give it the client’s transaction history and it starts at 85% and climbs past 90% 若提供客户的交易历史，模型准确率则从85%起步，并攀升至90%以上

nothing about the model changed, the knowledge did 模型本身未作任何改变，变化的是知识

writing works the same way 写作也是如此

a mid-tier model with a well-built voice profile produces more recognizable output than Fable 5 with no profile at all 一个配置了良好语音资料的中端模型，其输出比完全没有配置资料的Fable 5更具辨识度

the files carry more of the result than the model tier does 文件资料对结果的影响比模型等级更大

and the model itself rewards this harder than anything before it 而模型本身对此的奖励力度比以往任何模型都强

Anthropic’s own testing had Fable play a full deck-building game with file-based memory, and it improved three times more than the previous flagship Anthropic的测试让Fable通过文件记忆完整玩了一款卡牌构筑游戏，其表现比上一代旗舰机型提升了三倍以上

one game, tested by the vendor, replicated by nobody yet… but the move the number points at costs you a folder of markdown, so you take it either way 一款游戏，经过供应商测试，尚未有人复现……但数字指向的那步棋会让你损失一个Markdown文件夹，所以无论如何你都得走这一步

before we build, i have to tell you something: the model doesn’t magically find everything in your notes 开始构建之前，我得告诉你一件事：这个模型并不会神奇地找到你笔记里的所有内容

what it does is act on knowledge that lives outside the conversation, and cite where every piece came from 它的作用是调动对话之外存在的知识，并注明每一条信息的来源

the memory is yours, on your disk, in plain text you can open and read 记忆属于你，存放在你的硬盘上，以纯文本形式呈现，你可以随时打开阅读

give it a few weeks and the agent starts citing decisions you forgot you made 几周后，智能体开始引用那些你已经忘记自己曾做过的决定

the first question is where this memory should live 首要问题是这些记忆应该存储在哪里

the answer costs nothing and you might already have it installed 答案无需任何成本，而且你可能已经装好了它

what Obsidian is, in one minute 一分钟带你读懂Obsidian是什么

Obsidian is a free app that sits on top of a folder of markdown files on your computer Obsidian 是一款免费应用，基于你电脑上的 Markdown 文件文件夹运行

no database, no cloud lock-in… your notes are plain text files you own, and the app is just a beautiful window into them 无需数据库，不受云服务限制……你的笔记是纯文本文件，归你所有，而这款应用只是一个浏览它们的精美窗口

you only need two of its features: 你只需要使用它的两个功能：

> [[wikilinks]]: type double brackets around any note’s name and the two notes are now connected > [[维基链接]]：在任意笔记名称两侧输入双括号，两篇笔记即刻建立关联

> the graph view: Obsidian draws every note as a dot and every link as a line, so you see your knowledge as a web > 图谱视图：Obsidian 将每个笔记绘制为节点，每条链接绘制为连线，让你以网络的形式直观呈现知识体系

and it fits agents perfectly: because the vault is just a folder, Fable works on it directly through Claude Code, the terminal app the model runs in on your machine 与智能体完美契合：由于知识库本质上是一个文件夹，Fable 可通过 Claude Code（模型在终端运行的应用）直接对其进行操作

no plugin, no connector, no special setup… the agent reads and writes markdown files, Obsidian shows you what changed 无需插件、无需连接器、无需特殊配置……智能体直接读写 Markdown 文件，Obsidian 则为你展示实时变化

you use the app, the agent uses the folder, both are looking at the same brain 你使用应用界面，智能体操作文件夹，两者共享同一个知识中枢

and before you brace for a big project: the starter version of everything in this article takes about an hour, and once the reading rules are set it runs on pennies… the money part gets its own section 在你准备迎接一个大项目之前：这篇文章里所有内容的入门版大约只需一小时，一旦阅读规则设定好，运行成本极低……资金部分会单独讨论

what separates a brain from a pile of files is the structure… and structure is where almost everyone gets it wrong 大脑与一堆文件的区别在于结构……而结构恰恰是大多数人出错的地方

the structure: four pieces, nothing else 结构：四个部分，没有别的

the thinking comes from karpathy’s llm-wiki idea: treat your knowledge base like a codebase 这个思路源于卡帕西的”大语言模型维基”理念：将你的知识库视为代码库来对待

Obsidian is the editor, the model is the programmer, the wiki is the code Obsidian 是编辑器，模型是程序员，维基是代码

after digging through the setups people run in public, the repos, the viral templates, the failure threads, four pieces kept showing up: 翻阅了人们公开分享的配置方案、代码仓库、爆款模板和失败经历帖后，发现有四个要素反复出现：

> raw - everything you capture goes here untouched: articles, transcripts, call notes, competitor pages… read-only history, the agent never rewrites it > 原始素材 - 你捕获的所有内容原封不动存放于此：文章、转录稿、通话记录、竞品页面…只读历史记录，代理程序绝不修改

> entities - one page per concrete thing: a client, a competitor, a tool, a person > 实体 - 每个具体事物对应独立页面：客户、竞品、工具、人物

> concepts - one page per idea: a strategy, a pattern, a lesson > 概念——每页一个想法：一个策略、一个模式、一个教训

> INDEX.md - the front door: every page listed with a one-line description, so the agent knows what exists without opening everything > INDEX.md——前门：列出每一页及其一句话描述，这样智能体无需打开所有内容就能知道存在什么

the agent’s job is compiling: it reads new material in raw/ and updates the entity and concept pages, linking them as it goes 智能体的任务是编译：它读取 raw/ 目录中的新素材，并更新实体和概念页面，同时建立链接

and the writing rules are simple enough to fit in four lines: 而写作规则简单到只需四行就能概括：

> one lesson per file, with a one-line summary at the top > 每份文件一课，顶部附有一行总结

> update the existing page instead of creating a duplicate > 更新现有页面而非创建重复内容

> delete notes that turn out to be wrong > 删除经证实有误的笔记

> keep raw sources and compiled pages separate, always > 始终将原始来源与整理后的页面分开保存

why raw/ stays untouched: when the same agent reads and rewrites the same notes over and over, details blur and errors compound 为什么raw/文件夹保持原样不变：当同一个代理反复阅读和重写相同的笔记时，细节会模糊，错误会叠加

the raw folder is your ground truth, and the wiki keeps getting smarter on top of it 原始文件夹是你的真实依据，而知识库则在它的基础上不断变得更智能

and the pages are only half the value… the links between them are where the real edge hides 而且页面本身只占价值的一半…页面之间的链接才是真正的独特优势所在

knowledge graphs: why this gets better as it grows 知识图谱：为什么它会随着规模增长而变得更好

every [[link]] the agent writes between two pages is an edge in a graph 智能体在页面之间创建的每一个[[链接]]，都是一张图中的一条边

this is the part that separates a vault from a pile of notes: a search-based knowledge base gets noisier as it grows, because more files means more junk in every search 这正是区分资料库与笔记堆的关键：基于搜索的知识库会随着规模扩大而变得杂乱，因为文件越多，每次搜索中的无效信息就越多

a linked wiki gets stronger as it grows, because every new page connects into the web and makes the surrounding pages more useful 基于链接的维基系统会随着发展而愈发强大——每个新页面都会融入网络，让周边页面变得更有价值

when the agent needs to answer something, it doesn’t scan everything… it walks the links 当智能体需要回答问题时，它不会扫描所有内容……而是沿着链接逐步探索

from the client’s page to the campaign concept to the competitor page, following connections the way you’d follow your own memory 从客户页面到营销概念再到竞争对手页面，像追随自己的记忆一样追踪这些关联脉络

karpathy’s own vault sits around 100 articles and 400,000 words, all compiled by the model, all connected 卡帕西个人知识库大约包含100篇文章和40万字，全部由模型整理完成，相互之间建立了联系

open the graph view after two weeks of this and you’ll see your business as a living map… that picture alone changes how you think about what you know 坚持两周后打开图谱视图，你会看到自己的业务变成一幅活地图…仅凭这幅图像就能改变你对自己所知内容的认知方式

so how do you fill it without spending a month copy-pasting? 那么该如何填充它，而无需花一个月时间复制粘贴？

populate it with goals 填充目标

the first move is a backfill, and Fable’s goal system was built for exactly this job 第一步是反向填充，而Fable的目标系统正是为此而生

/goal in Claude Code lets you write one finish line, and the model keeps working on its own while a second, smaller model reads along as the judge and confirms when the line is crossed 在Claude Code中使用/goal指令，您只需设定一条完成标准，模型便会自主持续工作，同时由第二个小型模型担任评审，同步阅读并确认是否达到标准

the trick is the judge only sees what’s in the conversation, so the goal must demand proof it can read: 关键在于评审只能看到对话中的内容，因此目标必须要求提供它能够读取到的证据：

feed raw/ with what you already own before you run it: old chat transcripts, bookmarked threads, your notes app export, client folders, past research 用你已经拥有的东西填充原始数据/后再运行：旧的聊天记录、已收藏的帖子、笔记应用的导出文件、客户文件夹、过往研究

then walk away, and come back to a compiled brain 然后离开，回来时你将拥有一个整合完毕的”大脑”

two rules keep the backfill honest: 两条规则确保回溯的真实性：

> every change ships as a diff, the exact before-and-after lines, never as a claim… if the agent says it updated a page, the diff proves it > 每一次修改都以差异形式输出——精确记录修改前后的行文，绝不凭空断言……如果智能体声称更新了某个页面，差异记录就是铁证

> a page without a source link back to raw/ gets flagged, not trusted > 没有指向原始内容源链接的页面会被标记，不被信任

the backfill hands you a compiled brain… keeping it alive is a different job, and it’s the one everybody skips 回溯会为你提供一个编译好的大脑……让它保持活力是另一项任务，而这项任务每个人都跳过

keep it alive with loops 通过循环让它保持活力

a second brain that only grows when you remember to feed it is a dead brain in three weeks 一个只有在你记得喂养时才会生长的第二大脑，三周内就会死去

so the maintenance runs on schedules, not on memory: 维护是按照日程进行的，而不是靠记忆：

> after every session: a hook, a small script that fires on its own when a session ends, mines what just happened… decisions made, mistakes caught, patterns confirmed, written into the vault as dated notes… work you already did becomes memory without you filing anything > 每次会话结束后：一个钩子，一个在会话结束时自动触发的小脚本，会挖掘刚刚发生的事情……记录下决策、捕捉到的错误、被验证的模式，将这些以带日期的笔记形式写入资料库……你已完成的工作无需整理，就会自动转化为记忆

> every night: a compile pass on a cheap model reads the day’s new raw material and updates the wiki pages… routine work, routine tier > 每晚：由廉价模型执行的编译程序会读取当天的新素材，并更新维基页面……这是日常任务，属于常规层级

> every week: a lint pass hunts contradictions, duplicate pages, and dead links… this is the loop that keeps the graph clean, and it exists because unmaintained wikis rot > 每周：一次整理程序会检查矛盾内容、重复页面和无效链接……这是保持系统整洁的循环，之所以存在，是因为无人维护的维基会逐渐腐化

> every week: one synthesis pass on the big model reads across the whole vault and writes what changed this week, what’s drifting, what deserves attention > 每周一次：对大型模型进行一次综合梳理，扫描整个知识库，记录本周的变化、趋势以及值得关注的内容

that last one is the only pass where the premium model earns its seat 只有最后一次梳理才值得高端模型大显身手

everything else runs on the cheap tier, because updating notes is routine work and routing routine work to Fable is how people burn money for nothing 其余工作都用经济型模型完成，因为更新笔记属于常规任务，而将常规任务外包给Fable只会白白烧钱

maintenance keeps the vault clean… but where does the material that makes it valuable come from? 维护工作能让知识库保持整洁……但真正让其具有价值的素材从何而来？

the research workflow that feeds it 为其提供素材的研究工作流

this is where the vault stops being storage and becomes an edge, and it’s also the step where garbage usually gets in 正是从这里开始，知识库不再是存储，而成为了一种优势；而这一步，也通常是垃圾信息混入的地方

the default AI research is one prompt to a chatbot, and the answer dies in the scrollback 默认的AI研究方式就是向聊天机器人发送一个提示，然后答案就消失在历史记录中

worse, it’s built on stale knowledge… in AI, advice from six months ago is often actively wrong, and the real practitioner layer, what people are running right now, what breaks, what works, lives on socials, not in official docs 更糟的是，它建立在过时的知识之上……在AI领域，六个月前的建议往往已经是错的。真正实践者的层面——人们此刻正在运行什么、什么会出问题、什么能奏效——这些信息存在于社交媒体，而不是官方文档中

so the research machine works like this: 研究所的运作方式如下：

> one question goes in, and it gets split into 3-5 sub-questions > 一个问题输入后，被拆分为3-5个子问题

> parallel agents fan out, each searching a different surface: socials for the practitioner layer, the web for documentation and pricing, scrapers pulling the full text of everything worth reading > 并行代理向外辐射，各自搜索不同层面：社交媒体寻找实践者层，网络检索文档与定价信息，爬虫抓取所有值得阅读内容的完整文本

> every finding becomes a receipt: the claim, the source link, the date > 每项发现都形成凭证：包含主张、来源链接和日期

> then the gate that makes it real: a skeptic agent attacks every claim and tries to kill it… single-source hype gets labeled, contradictions get surfaced, only survivors pass > 然后那道让它成真的门出现了：一个怀疑代理会攻击每一条主张，试图将其扼杀……单一来源的热度会被标注，矛盾之处浮出水面，只有幸存者能过关

> verified findings land in the vault as pages, each one dated and linked, each one carrying an expiry date so stale knowledge announces itself > 经过验证的发现会以页面形式存入资料库，每条都带有日期和链接，每条都附有有效期限，这样过时的知识会自行显现

and the exact stack i run it on: 而我所用的确切技术栈是：

> last30days powered by ScrapeCreators (scrapecreators.com): one skill that sweeps reddit, X, youtube, instagram and tiktok for the last 30 days of practitioner talk on any topic> last30days 由 ScrapeCreators（scrapecreators.com）驱动：一项技能，它能扫描 Reddit、X、YouTube、Instagram 和 TikTok 上过去30天内关于任何话题的从业者讨论

> the official X MCP (api.x.com/mcp): live posts, threads and bookmarks straight from the source> 官方 X MCP（api.x.com/mcp）：来自源头的实时帖子、串帖和书签

> youtube transcripts with yt-dlp (github.com/yt-dlp/yt-dlp): any walkthrough or tutorial becomes text the agent can mine> 使用 yt-dlp 获取 YouTube 转录文本（github.com/yt-dlp/yt-dlp）：任何操作指南或教程都能转化为代理可挖掘的文本

> instagram and tiktok content through ScrapeCreators, because short-form is where new workflows surface first > 通过 ScrapeCreators 获取 Instagram 和 TikTok 内容，因为短格式内容往往是新工作流最先涌现的地方

> Perplexity deep research (perplexity.ai): the cited long-read pass across the web> Perplexity 深度研究（perplexity.ai）：对网络内容进行带引用的长文深度检索

> Firecrawl (firecrawl.dev): pulls the full text of every page worth keeping, as clean markdown> Firecrawl (firecrawl.dev)：提取每篇值得保留页面的完整文本，生成干净的 Markdown 格式

the skeptic is what separates research from rumor collection: fresh-context checkers outperform a model reviewing its own work, so the attack always comes from an agent that didn’t do the research 怀疑精神是区分研究与谣言收集的关键：基于新语境的验证者比模型自我审查更有效，因此攻击始终来自未参与研究的智能体

run this on your niche once a week and the vault fills with verified, dated, sourced intelligence your competitors are missing 每周在你的细分领域运行一次，知识库就会积累竞争对手所缺失的、经核实的、带日期与来源的智能情报

all of it worthless, though, if reading the vault costs more than it returns 不过，如果查阅知识库的成本高于其回报，这一切便毫无价值

read it without burning money 无成本阅读

a vault only works long-term if reading it is cheap, and this is the leak in almost every setup 长期有效的知识库，关键在于阅读成本低廉，而这正是几乎所有方案中的漏洞

the mental model: the context window is an expensive room, and everything that enters it gets paid for in tokens… the chunks of text every AI bill is counted in 核心思维模型：上下文窗口是一间昂贵的房间，每进入其中的内容都要按令牌付费…AI账单就是按这些文本块计算的

> your CLAUDE.md, the instruction file the agent reads at the start of every session, loads automatically every single time… that’s the always-paid tax… keep it under 200 lines pointing at the vault, never containing it > 你的CLAUD.md文件——智能体每次会话开始时自动读取的指令文档，每次都会自动加载…这是永远要付的税…请将其控制在200行以内，只需指向知识库，绝不要包含库本身

> everything else is pay-per-read: the agent checks INDEX.md, follows links, greps for keywords, and opens only the pages the trail points at… a full-folder sweep is the one move that never happens > 其他内容均为按篇计费：代理程序会查阅 INDEX.md、跟踪链接、grep 关键词，并只打开线索指向的页面……整文件夹扫描是永远不会出现的操作

> for big questions, send a worker: a subagent reads fifty pages in its own separate context and returns one paragraph of conclusions to your session… the expensive room holds decisions, not the library > 应对复杂问题，派遣工作线程：子代理在独立的上下文中阅读五十页内容，将一段结论返回至您的会话中……昂贵的对话空间只承载决策，而非整座图书馆

wire it into everything you build 与你构建的一切相连

a vault that only stores things is a filing hobby… this one feeds every project you run 仅用于存储资料的仓库只是归档癖好……而这里的数据仓库将滋养你运营的每一个项目

point any project at it with three lines in that project’s CLAUDE.md: 将任何项目指向它，只需在该项目的 CLAUDE.md 中添加三行内容：

## knowledge
- before starting, read the relevant pages from ~/vault/entities/ and ~/vault/concepts/
- ground every claim about our business, clients or audience in a vault page

and the outputs change immediately: 输出便会立即改变：

> marketing: campaign briefs grounded in your actual audience pages and competitor history, not generic personas > 营销：基于你真实受众页面和竞争对手历史的营销活动简报，而非泛泛的用户画像

> content: drafts that cite your own past research and match your voice profile > 内容：引用你过去研究并符合你个人风格特征的草稿

> coding: the agent keeps living architecture notes per project in the vault, so no session ever starts blind > 编码：代理在每个项目的存储库中持续维护动态架构笔记，因此每次会话都不会从零开始

> client work: every deliverable opens with the full relationship history behind it > 客户工作：每份交付成果都附有背后完整的合作历史记录

then the second half: the vault itself becomes product 下半部分则转向：存储库本身成为了产品

the research pages become articles and guides, the concept pages become courses, the client pages become case studies… you’re not creating from a blank page anymore, you’re packaging what the machine already verified 研究页面变成了文章和指南，概念页面变成了课程，客户页面变成了案例研究……你不再是从空白页开始创作，而是在整理机器已经验证过的内容。

the warning that saves your vault: sync is where vaults die 一个能救你资料库的警告：同步是资料库的葬身之地。

run a single sync system… if the agent writes files while iCloud syncs them, you get conflicted copies and scrambled folders 只运行一个同步系统……如果智能体在iCloud同步时写入文件，你就会得到冲突副本和混乱的文件夹。

git, the save-point system programmers use, works as the checkpoint layer… it locks in a version only when you tell it to, and that setup survives 程序员使用的版本控制工具Git可以作为检查点层……只有在手动确认时它才会锁定一个版本，而这样的设置不会崩坏。

the card 卡片

the whole build, in order… copy this: 整个构建流程，按顺序…复制以下内容：

> create the vault: raw/, entities/, concepts/, and an INDEX.md > 创建资料库：raw/（原始资料）、entities/（实体）、concepts/（概念），以及一个 INDEX.md（索引文件）

> write the four rules into your CLAUDE.md: one lesson per file, update don’t duplicate, delete what’s wrong, never touch raw/ > 将四条规则写入你的 CLAUDE.md：每份文件只记录一条经验、更新而非复制、删除错误内容、绝不修改 raw/ 目录

> dump everything you own into raw/: transcripts, bookmarks, notes, client folders > 把所有东西都丢进 raw/：转录文本、书签、笔记、客户文件夹

> run the /goal backfill with pasted proof and a stop clause > 运行 /goal backfill，粘贴证据并设置终止条件

> schedule the loops: session hook, nightly compile on the cheap tier, weekly lint, one premium synthesis pass > 安排循环任务：会话钩子、低成本层级的夜间编译、每周代码检查、一次优质综合处理

> run the weekly research sweep: fan out, let the skeptic attack, land survivors as dated pages > 执行每周研究扫描：广泛发散，让怀疑者攻击，最终将幸存内容保存为带日期标记的页面

> add the three knowledge lines to every project’s CLAUDE.md > 将三条知识线添加到每个项目的 CLAUDE.md 文件中

ihe model in the driver’s seat will change again… the vault survives every swap, and the feedback written into it makes it smarter every week no matter who’s driving 驾驶座上的模型会再次更换……但知识库在每次替换中都得以保留，而写入其中的反馈会使其每周都变得更智能，无论谁在驾驶

the smallest version takes an hour: one folder, ten files about your business, and an agent told to read them first 最小版本只需一小时：一个文件夹，十个关于你业务的文件，以及一个被告知先阅读这些文件的智能体

your outputs will tell you the rest 剩下的部分，你的输出会告诉你

join the best AI community in the world: weeklyaiops.com加入全球顶尖AI社区：weeklyaiops.com