GeorgeYan

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.

Go (Golang) GOOS and GOARCH

All of the following information is based on go version go1.17.1 darwin/amd64.

GOOS Values

GOOS	Out of the Box
`aix`	✅
`android`	✅

Example of deployment process which I use in my Nuxt.js projects. I usually have 3 components running per project: admin-panel SPA, nuxt.js renderer and JSON API.

This manual is relevant for VPS such as DigitalOcean.com or Vultr.com. It's easier to use things like Now for deployment but for most cases VPS gives more flexebillity needed for projects bigger than a landing page.

UPD: This manual now compatible with nuxt@2.3. For older versions deployment, see revision history.

Let's assume that you have entered fresh installation of Ubuntu instance via SSH. Let's rock:

	#!/usr/bin/env zsh
	brew install lrzsz

	recv='/usr/local/bin/iterm2-recv-zmodem.sh'
	send='/usr/local/bin/iterm2-send-zmodem.sh'

	curl -sSL https://github.com/RobberPhex/iTerm2-zmodem/raw/master/iterm2-recv-zmodem.sh -o "$recv"
	curl -sSL https://github.com/RobberPhex/iTerm2-zmodem/raw/master/iterm2-send-zmodem.sh -o "$send"

	chmod +x "$recv" "$send"

	# type(<scope>): <subject>

	# <body>

	# <footer>

	# type 字段包含:
	# feat：新功能（feature）
	# fix：修补bug
	# docs：文档（documentation）

	package main

	/*
	http代理测试


	* [Mocking a HTTP access with http.Transport in Golang - oinume journal](http://oinume.hatenablog.com/entry/mocking-http-access-in-golang)
	* [Go http访问使用代理](http://www.cnblogs.com/damir/archive/2012/05/06/2486663.html)
	* [GO HTTP client客户端使用 - 海运的博客](https://www.haiyun.me/archives/1051.html)
	* [Making Tor HTTP Requests with Go \| DevDungeon](http://www.devdungeon.com/content/making-tor-http-requests-go)

	# Version key/value should be on his own line
	PACKAGE_VERSION=$(cat package.json \
	\| grep version \
	\| head -1 \
	\| awk -F: '{ print $2 }' \
	\| sed 's/[",]//g')

	echo $PACKAGE_VERSION

	#!/bin/sh
	# Mounts PWD (pom.xml must be at root) on lightest Maven image and extracts the project version
	VERSION=`docker run -v $(pwd):/usr/src/app \
	-w /usr/src/app \
	maven:3.6.0-jdk-8-alpine mvn org.apache.maven.plugins:maven-help-plugin:3.1.0:evaluate -Dexpression=project.version \
	-q \
	-DforceStdout`
	docker tag $IMAGE_NAME $DOCKER_REPO:$VERSION
	docker push $DOCKER_REPO:$VERSION

	{
	"globals" :
	{
	"alwaysShowTabs" : true,
	"defaultProfile" : "{09dc5eef-6840-4050-ae69-21e55e6a2e62}",
	"initialCols" : 120,
	"initialRows" : 30,
	"keybindings" :
	[
	{

	## Consumer Throughput: Single consumer thread, no compression
	## Consumer Throughput: 3 consumer thread, no compression

	bin/kafka-consumer-perf-test.sh --topic benchmark-3-3-none \
	--zookeeper kafka-zk-1:2181,kafka-zk-2:2181,kafka-zk-3:2181 \
	--messages 15000000 \
	--threads 1