Marco Tse Marcotsept

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.

>>> docker exec -it CONTAINERID /bin/sh
/app # telnet
/bin/sh: telnet: not found

/app # apk update
fetch http://dl-cdn.alpinelinux.org/alpine/v3.7/main/x86_64/APKINDEX.tar.gz
fetch http://dl-cdn.alpinelinux.org/alpine/v3.7/community/x86_64/APKINDEX.tar.gz
v3.7.0-243-gf26e75a186 [http://dl-cdn.alpinelinux.org/alpine/v3.7/main]
v3.7.0-229-g087f28e29d [http://dl-cdn.alpinelinux.org/alpine/v3.7/community]

Kubernetes on DigitalOcean with CoreOS

Let's look at an example of how to launch a Kubernetes cluster from scratch on DigitalOcean, including kubeadm, an Nginx Ingress controller, and Letsencrypt certificates.

Overview

Environment

We'll be creating a four-node cluster (k8s-master, k8s-000...k8s-002), load balancer, and ssl certificates.

A rooted device (Google if you need help here).
A computer with [ADB][0] installed, and a working knowledge of ADB.
A copy of an Action Launcher 3 APK on your computer. [APKMirror.com][1] is a good place to grab a copy from. All you're doing here is setting Action 3 as a system app, so any version of the app will work. From there, the Action 3 will update via the Play Store like a typical pre-installed app (such as Gmail or YouTube).

	version: '3'
	services:
	mongo1:
	hostname: mongo1
	container_name: localmongo1
	image: mongo
	restart: always
	expose:
	- 27017
	ports:

	axios({
	url: 'http://localhost:5000/static/example.pdf',
	method: 'GET',
	responseType: 'blob', // important
	}).then((response) => {
	const url = window.URL.createObjectURL(new Blob([response.data]));
	const link = document.createElement('a');
	link.href = url;
	link.setAttribute('download', 'file.pdf');
	document.body.appendChild(link);

	There is no way to store an empty object/array/null value.
	There are also no actual arrays. Array values get stored as objects with integer keys.
	(If all keys are integers, it will be returned as an array.)

	Basically, it's one giant tree of hashes with string keys.
	Simply write a value to any location, and the intermediary locations will automatically come into existance.

	── Classes ──

	DataSnapshot : Container for a subtree of data at a particular location.

	FROM php:5.6-cli

	RUN apt-get update \
	&& apt-get install -y \
	libfreetype6-dev \
	&& rm -rf /var/lib/apt/lists/* \
	&& docker-php-ext-install iconv \
	&& apt-get remove -y \
	libfreetype6-dev \
	&& apt-get install -y \

	/*==================================================
	= Bootstrap 3 Media Queries =
	==================================================*/

	/========== Mobile First Method ==========/

	/* Custom, iPhone Retina */
	@media only screen and (min-width : 320px) {

	}

Marco Tse Marcotsept

LLM Wiki

The core idea

Kubernetes on DigitalOcean with CoreOS

Overview

Environment

Table of Contents

Default

Powerlevel10k