RAG è già vecchio? Il nuovo modo di costruire Knowledge Base per AI con LLM Wiki
Join my AI Academy: https://www.rizzoaiacademy.com/ Want to develop advanced AI solutions? https://inferentia.xyz IG: / simorizzo_ai In this video, I'll show you a new approach to building an AI knowledge base using LLM Wiki. Let's start with the basics and take a comprehensive look: What is RAG (Retrieval Augmented Generation) and what are its limitations? How does Agentic File Search work? Why LLM Wiki represents a significant evolution for AI agent memory? How to structure a persistent wiki in Markdown. How to install and use it in practice. Step-by-step demo. The key idea is this: instead of retrieving fragments from scratch each time, knowledge is compiled, updated, and maintained over time. This is a very interesting paradigm shift for those who want to build more reliable AI agents, with more useful, navigable, and cumulative external memory. The video also covers: practical differences between RAG, Agentic Search, and LLM Wiki knowledge base structure ingest, query, and maintenance workflows real-world use cases for research, teams, studios, and more advanced AI systems References: Karpathy's post: https://x.com/karpathy/status/2039805... Official repo/gist: https://gist.github.com/karpathy/442a... Let me know in the comments if you'd like me to show you how to build a local LLM Wiki, integrate it with an agent, or compare it to a classic RAG system in the next video. 00:00 Introduction to LLM Wiki and Andrej Karpathy 01:22 Why LLMs need external storage 02:03 Gen 1: What is RAG (Retrieval Augmented Generation) 05:05 Gen 2: Agentic File Search and File System 07:25 Limitations of RAG and Advantages of Agents 09:56 Gen 3: LLM Wiki and Karpathy's Vision 11:46 The Technical Structure: Raw, Wiki, Index, and Log 13:11 Obsidian: The Knowledge Base Interface 14:26 How AI Explores and Responds Using the Wiki 16:43 MARP Format and Linting Phase (Health Check) 18:00 The Three Levels of Architecture 21:07 Tutorial: Installing Obsidian and Web Clipper 22:36 AI Agent Configuration and Initial Prompt 24:39 Practical Example: Ingesting Data from the Web 27:36 Creating Real-Time Knowledge Base 30:10 Querying: Querying Your Wiki 32:38 Case Study: Inserting and Parsing PDF Documents 37:32 Generating Automatic Presentations with MARP 39:36 Automatic Wiki Maintenance (Linting) 40:36 Final Comparison of RAG, Agentic, and LLM Wiki 42:31 Conclusions and Final Thoughts #agenticai #llm #aiagent

Is MCP already obsolete? Are AI Agents moving to CLIs?

Creo un Sistema RAG da Zero: Nessuna Libreria, Solo Python Purissimo!

Karpathy's LLM Wiki - Full Beginner Setup Guide

I created a Second Brain (LLM Wiki) in 15 minutes with Claude + Obsidian (Karpathy method)

Is RAG Still Needed? Choosing the Best Approach for LLMs

Ho implementato LLM Wiki di Karpathy (ed è geniale)

Ho Provato a Trasformare 100€ in 1000€ con l'AI

Come far gestire all'agente la tua Knowledge Base: LLM Wiki + QMD | tutorial completo

The Best Way to Build an AI Knowledge Base in 2026

The Secret of AI Agents' Memory
![Special Episode with an Anthropic Researcher [Gian Segato]](https://i.ytimg.com/vi/_I54OFKJyZU/hqdefault.jpg?sqp=-oaymwEjCNACELwBSFryq4qpAxUIARUAAAAAGAElAADIQj0AgKJDeAE=&rs=AOn4CLAhaAX0-cFTygLeLs1T_WDFk-tMJg)
Special Episode with an Anthropic Researcher [Gian Segato]

Qwen 3.5's new FREE local AI is INSANE!

DeepSeek 4.0: What NVIDIA Feared Has Come True!

Build An AI Second Brain Knowledge Base (Step-By-Step)

Come Addestrare un LLM con i Tuoi Dati Personali: Guida Completa con LLaMA 3.2

Convert PDFs to LLM Datasets in Minutes! (The Guide No One Told You)

How To Build LLM Wiki In Obsidian? 🧠 A Memory Layer For Any Agentic AI

Finetuning o RAG? La Verità sull'Addestramento dei Modelli AI (E Come Scegliere!)

Claude Code + Obsidian: Ho creato un sistema che RICORDA TUTTO (Ti mostro come)

