Farzapedia：把个人数据变成可导航的个人维基百科

ID: 62057772
原文链接: https://x.com/karpathy/status/2040572272944324650
作者: Andrej Karpathy (@karpathy)
日期: 2026-04-05
分类: knowledge-management
标签: ai-tools, personal-wiki, llm, knowledge-management, file-over-app, BYOAI
质量评分: 5/5
抓取时间: 2026-06-27T20:40:00

中文翻译

背景

Farzapedia，Farza 的个人维基百科 —— 这是继 Karpathy 之前那条 "Wiki LLM" 推文之后的一个优秀落地案例。

Karpathy 表示，他非常喜欢这种个人化方案。相比"AI 用得越多就越聪明"这种默认模式，这种方案有以下几点关键优势：

四大核心优势

1. 显式（Explicit）

记忆产物是一个明确、可导航的 wiki。

你可以精确看到 AI 知道什么、不知道什么
可以检查和管理这个知识库
即使你本人不参与写作（LLM 代劳）
关于你的知识不是隐含的、不可见的，而是显式的、可查看的

2. 属于你（Yours）

你的数据属于你自己，在你自己的本地电脑上
不在某个 AI 厂商的系统里、无法导出
你对自己的信息拥有完全控制权

3. 文件优于应用（File over app）

记忆是通用格式（图片、Markdown）的简单文件集合
数据是可互操作的：你可以用各种工具 / CLI 处理这些信息
智能体可以应用整个 Unix 工具栈来处理它们
任何类型的数据都可以作为输入导入文件，任何类型的界面都可以作为输出查看
例如你可以用 Obsidian 查看它们，或自己 vibe code 一些工具
搜索 "File over app" 找到相关文章了解这套理念

4. 自带 AI（BYOAI）

你可以用任何你想要的 AI 来"插上"这些信息 —— Claude、Codex、OpenCode，随意
你甚至可以拿一个开源 AI 在你的 wiki 上做微调 —— 原则上这个 AI 可以"认识你"在其权重里，而不仅仅是从你的数据中检索
让 AI 公司时刻保持警觉 😄

Karpathy 总结

这种个人化方案把你放在完全控制的位置：

- 数据是你的

- 通用格式

- 显式且可检查

- 可用任何你想要的 AI

让 AI 公司时刻保持警觉！

当然，这并不是让 AI 认识你的最简单方法 —— 它确实需要你管理文件目录等。但智能体（agents）也让它变得相当简单，它们能帮上很多忙。Karpathy 预想会有不少产品出来让这一切变得更容易，但在他看来——

"智能体熟练度"是 21 世纪的一项核心技能

这些是极其强大的工具 —— 它们会说英语，并为你完成所有计算机相关的工作。抓住这个使用智能体的机会。

互动数据

帖文发布时间：2026-04-05 07:28 AM
浏览量：1,400,000+
引用的原帖来自 @FarzaTV：让 LLM 从 2,500 条日记、备忘录、iMessage 记录中提取，生成 400 篇关于他朋友、初创公司、研究领域、甚至他最喜欢的动漫及其影响的长文

社区共鸣（精选回复）

"Holy crap this is epic. ... Currently adding a /pai URL path to PAI's new Pulse system, which will be a web front end for managing all: Documentation, Knowledge (ideas, people, companies, etc), Learnings, Etc."

—— Daniel Miessler

"File over apps 💯"

—— Muratcan Koylan（"The File System Is the New Database: How I Built a Personal OS for AI Agents" 作者）

"This is a brilliant step toward true AI personalization. Farzapedia turns your personal data into clean, ownable, backlinked files that any agent can use without vendor lock-in."

—— Lonely

"The explicit memory artifact point is the key: 2,500 diary entries into 400 linked pages gives agents real retrieval paths instead of fuzzy chat memory. That file-over-app approach is what makes personalization portable."

—— Nyk（"Claude + Obsidian: The Memory Stack That Compounds" 作者）

*来源：Andrej Karpathy (@karpathy), 2026-04-05*

English Original

Farzapedia: turning personal data into a navigable personal Wikipedia

Farzapedia — Farza's personal wikipedia. A great follow-up example to my Wiki LLM tweet. I really like this approach to personalization in a number of ways, compared to "status quo" of an AI that allegedly gets better the more you use it:

The four advantages

1. Explicit

The memory artifact is explicit and navigable (the wiki). You can see exactly what the AI does and does not know and you can inspect and manage this artifact, even if you don't do the direct text writing (the LLM does). The knowledge of you is not implicit and unknown, it's explicit and viewable.

2. Yours

Your data is yours, on your local computer. It's not in some particular AI provider's system without the ability to extract it. You're in control of your information.

3. File over app

The memory here is a simple collection of files in universal formats (images, markdown). The data is interoperable: you can use a very large collection of tools/CLIs or whatever you want over this information because it's just files. The agents can apply the entire Unix toolkit over them.

Any kind of data can be imported into files as input, and any kind of interface can be used to view them as the output. E.g. you can use Obsidian to view them or vibe code something of your own. Search "File over app" for an article on this philosophy.

4. BYOAI

You can use whatever AI you want to "plug into" this information — Claude, Codex, OpenCode, whatever. You can even think about taking an open source AI and finetuning it on your wiki — in principle, this AI could "know" you in its weights, not just attend over your data.

Karpathy's takeaway

This approach to personalization puts you in full control:

- The data is yours

- In universal formats

- Explicit and inspectable

- Use whatever AI you want over it

Keep the AI companies on their toes! :)

Certainly this is not the simplest way to get an AI to know you — it does require you to manage file directories and so on, but agents also make it quite simple and they can help you a lot. I imagine a number of products might come out to make this all easier, but imo "agent proficiency" is a CORE SKILL of the 21st century.

These are extremely powerful tools — they speak English and they do all the computer stuff for you. Try this opportunity to play with one.

Engagement

Posted: 7:28 AM · Apr 5, 2026
Views: 1,400,000+
Quoted from @FarzaTV: "I had an LLM take 2,500 entries from my diary, Apple Notes, and some iMessage convos to create a personal Wikipedia for me. It made 400 detailed articles for my friends, my startups, research areas, and even my favorite animes and their impact on me."

Community signal

Daniel Miessler (PAI): "Holy crap this is epic. Currently adding a /pai URL path to PAI's new Pulse system, which will be a web front end for managing all: Documentation, Knowledge, Learnings, Etc."
Muratcan Koylan ("The File System Is the New Database"): "File over apps 💯"
Lonely: "This is a brilliant step toward true AI personalization. Farzapedia turns your personal data into clean, ownable, backlinked files that any agent can use without vendor lock-in."
Nyk ("Claude + Obsidian: The Memory Stack That Compounds"): "The explicit memory artifact point is the key: 2,500 diary entries into 400 linked pages gives agents real retrieval paths instead of fuzzy chat memory."

*Source: Andrej Karpathy (@karpathy), April 5, 2026*