Skip to content

Instantly share code, notes, and snippets.

View rbergman's full-sized avatar

Bob Bergman rbergman

View GitHub Profile
#!/usr/bin/env ruby
require 'ftools'
require 'fileutils'
require 'rubygems'
require 'RMagick'
include Magick
require 'open3'
def merge( files = [] )
@rmanalan
rmanalan / wc-rest-next.js
Created September 30, 2010 21:53
wc-rest.next.js
/*
* A more dynamic API for WebCenter
* Rich Manalang / @rmanalan
*
* This is an attempt to make a better Javascript wrapper for the WebCenter REST API.
* Goals:
* - Dynamic object creation
* - Callbacks receive proper objects from prior call
* - Concurrent request support
* - Beautiful API
@ramnathv
ramnathv / gh-pages.md
Created March 28, 2012 15:37
Creating a clean gh-pages branch

Creating a clean gh-pages branch

This is the sequence of steps to follow to create a root gh-pages branch. It is based on a question at [SO]

cd /path/to/repo-name
git symbolic-ref HEAD refs/heads/gh-pages
rm .git/index
git clean -fdx
echo "My GitHub Page" > index.html
@hcarver
hcarver / redis-migrate.rb
Created October 6, 2016 20:08
For migrating between Redis databases on Heroku, when the source database doesn't support the sync Redis command
# For migrating between Redis databases on Heroku, when the source database doesn't support sync
# This is useful for migrating away from Redis Cloud, for example.
require "redis"
# You need to have created a new database first, with
# heroku addons:create heroku-redis:hobby-dev -a $APP
# OR
# heroku addons:create heroku-redis:premium-0 -a $APP

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.