Creo un Cerebro Inversor con Claude Code (Método Revelado por Anthropic)

#claude #claudecode #investments #wallstreet #warrenbuffett #karpathy #obsidian #claudebrain #AI ​​#artificialintelligence #investing #money #stockmarket 👉 Continue your education at: https://jfpartners.net/domina-la-bolsa 👉 Access my free investment course: https://jfpartners.net/curso-bolsa-e-... 👉 Follow me on Instagram:   / javieerdelvalle   👉 Karpathy Guide: https://gist.github.com/karpathy/442a... Here's an optimized description for search engine ranking. It's designed so that the first two lines (what you see before "show more") sell the click, and the rest loads keywords naturally: Claude Code, second brain, Obsidian, Karpathy, AI for investing, Warren Buffett. I'm building a second investing brain with Claude Code and Obsidian, following the method shared by Andrej Karpathy (Anthropic, co-founder of OpenAI): a persistent wiki that synthesizes your documents once and accumulates knowledge forever. In this video, I show you the complete step-by-step process, applied to a real-world case: ingesting 17 annual letters from Warren Buffett and converting them into a network of interconnected concepts (margin of safety, float, Mr. Market, economic goodwill, etc.) that the AI ​​can consult instantly, with traceable citations for each letter. The problem with using AI the way we all do: you dump the files on it, the conversation ends, and the context dies. Every new question starts from scratch. Karpathy's solution is a 3-tier architecture—raw files, an AI-managed wiki, and a CLAUDE.md file containing your rules and investment philosophy—that builds knowledge layer upon layer instead of repeating work. What you'll learn in this video: ▸ Why AI context gets lost (and how to make it persistent) ▸ Karpathy's LLM-Wiki pattern explained simply: raw → wiki → CLAUDE.md ▸ How to set up Claude Code step by step (desktop app, no terminal required) ▸ How to structure the index and log so AI doesn't "hallucinate" data ▸ Building the Buffett brain live: pilot, progressive intake, and review ▸ How to visualize the network of concepts with Obsidian (connection graph) ▸ Real-world applications for investors: company analysis, earnings tracking, industry and sector research ▸ The Obsidian Web Clipper trick for feeding the brain with articles This method works for anything you need to study in depth: a company (10-K, 10-Q, annual reports), an industry, academic literature, or your own business. The key: AI only uses YOUR sources, with citations, without padding with the current consensus. Chapters: 00:00 Karpathy's Method (from Anthropic) 01:45 The Problem of AI Today 03:47 The 3-Layer Architecture 08:00 Case Study: Warren Buffett 17:17 The Brain in Obsidian 20:26 How to Scale It The content, information, and opinions provided and expressed in all our videos, texts, and/or images are for informational and educational purposes and within the framework of the authors' free interpretation of the subject matter. The content of all our videos, texts, and/or images does not constitute an offer or recommendation to buy or sell financial instruments. The recipient of this content should be aware that the financial instruments and values ​​discussed may not be suitable for their profile or specific investment objectives. Therefore, the recipient must make their own investment decisions, seeking any specialized advice they deem necessary and assuming full responsibility for both the means and the results. The authors are not responsible for the use or interpretation of this content, nor for any damages or losses the recipient may suffer by entering into transactions or making any decisions based on this content.