gogonaft

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.

	#!/bin/bash

	#
	# @author: Hernan Garcia <hernandanielg@gmail.com>
	# https://gist.github.com/hernandanielg/430f3adb8e297f37ef6f0efb45a51bdc
	#
	# usage: ./delete_iam_user.sh [options] <user>
	# options:
	# -d\|--dry-run dry run mode
	#

	# This example assumes the NGINX proxy is on the same host as the Plex Media Server.
	# To configure Plex Media Server to serve requests without requiring authentication,
	# ensure that your LAN subnet is correctly added to the advanced server setting called
	# "List of IP addresses and networks that are allowed without auth". Example:
	# 192.168.0.1/24

	upstream plex-upstream {
	server 127.0.0.1:32400;
	}

	# Security groups that contain 0.0.0.0/0 rules
	aws ec2 describe-security-groups --filters Name=ip-permission.cidr,Values=0.0.0.0/0 --output=text \| grep SECURITYGROUPS

	# Security groups for ElasticSearch
	aws ec2 describe-security-groups --filters Name=ip-permission.from-port,Values=9200 --output=text \| grep SECURITYGROUPS

	# Search last 10,000/1MB of CloudTrail logs for 'AccessDenied' (removed AWS account number from stream name)
	aws logs get-log-events --log-group-name CloudTrail/DefaultLogGroup --log-stream-name 000000000000_CloudTrail_eu-west-1 \| grep AccessDenied

	# Get number of AWS API calls in time period (assumes a Cloudwatch Logs 'catch-all' filter and metric has been created against CloudTrail logs)