Загрузка...

This AI Never Forgets Anything I Save

📩 Collab & sponsorship enquiries: ssktechy@gmail.com

👉 Want to automate your business with AI? Visit our website : https://replixlab.com
─────────────────────────────
All Resourcs: https://docs.google.com/document/d/1nBhG0welrl4P00ifqU2PWlgcY0ymjucNW0eb4iW6O7Y/edit?usp=sharing

Most “AI knowledge systems” today rely on RAG, embeddings, and vector databases. That works, but it adds complexity, cost, and a lot of moving parts.

A few days ago, Andrej Karpathy shared a different idea:
skip the infrastructure and just use markdown files + an LLM.

So I built my own version of that system using Claude Code and Obsidian.

The result is a self-organizing second brain:

I drop raw content into a folder
Claude processes it into structured wiki pages
everything gets linked automatically
I can ask questions, and the system keeps improving over time

In this video, I show the full setup step by step:

the exact folder structure (raw vs wiki)
how Claude builds and maintains the knowledge graph
live ingest demo with real-time graph growth
how querying works without vector databases
the compounding loop that makes the system smarter
and 4 key lessons I learned building this

This is not a theoretical system. It’s my actual working setup for research, content, and business workflows.

If you want a simple, powerful alternative to RAG, this is worth trying.

🔔 Subscribe for more AI stuff: https://www.youtube.com/@ssktechy
Connect with me: 📸 Instagram: https://www.instagram.com/ssktechy
💼 LinkedIn: https://www.linkedin.com/in/dssk

Видео This AI Never Forgets Anything I Save канала ssktechy
Яндекс.Метрика
Все заметки Новая заметка Страницу в заметки
Страницу в закладки Мои закладки
На информационно-развлекательном портале SALDA.WS применяются cookie-файлы. Нажимая кнопку Принять, вы подтверждаете свое согласие на их использование.
О CookiesНапомнить позжеПринять