karpathy

LLM Wiki

A pattern for building personal knowledge bases using LLMs.

This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.

The core idea

Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.

qmacro

*&---------------------------------------------------------------------*
*& Report  UI5_REPOSITORY_LOAD
*&
*&---------------------------------------------------------------------*
*&
*& This report implements the up- and download of a SAPUI5 application
*& into an UI5 Repository.
*&
*& Furthermore it is possible to delete a UI5 Repository
*&

bevacqua

var spawn = require('child_process').spawn;

gulp.task('npm', function (done) {
  spawn('npm', ['publish'], { stdio: 'inherit' }).on('close', done);
});

banaslee

eXtreme Go Horse (XGH) Process

Quelle: http://gohorseprocess.wordpress.com

Übersetzung ursprünglich von https://gist.github.com/Neffez/f8d907ba8289f14e23f3855011fa4e2f

1. Ich denke, also ist es nicht XGH.
In XGH wird nicht gedacht, es wird das erste gemacht, was in den Sinn kommt. Es gibt auch keine zweite Option, die erste ist schneller.

2. Es gibt 3 Wege ein Problem zu lösen: den richtigen Weg, den falschen Weg und den XGH Weg, welcher exakt wie der falsche ist, aber schneller.