Viktor Taranenko viktortnk

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.

Get Mountain Lion and Homebrew to Be Happy

1) Install XCode 4.4 into /Applications

Get it from the App Store.

2) Install Command Line Tools

In XCode's Preferences > Downloads you can install command line tools.

	cask "starcraft-mass-recall" do
	version "8.0.1"
	sha256 "76750def5b63cf58afeb63e9948b5efec23adb436964eb730fe5d800d5a89e3b"

	resources = {
	mod: ["3246208", "SCMR_v#{version}.zip"],
	assets: ["3239552", "SCMRassets_v#{version}.zip", "9277d51c89e3cc473a3b5d588c2f3cdb"],
	cinematics: ["2819339", "SCMRcinematics_v7.2.zip", "f19ea12979eb8eaf703ed00ef33b7912"],
	redux: ["3246278", "Enslavers_Redux_v#{version}.zip", "b61129ef7df2004faa4c08480e447f31"],
	}

	import torch
	import torch.nn as nn
	import numpy as np
	import torch.optim as optim
	from torch.autograd import Variable

	# (1, 0) => target labels 0+2
	# (0, 1) => target labels 1
	# (1, 1) => target labels 3
	train = []

	/**
	* This is not meant to cover ALL possible ways of creating service clients; instead it focuses on the simplest way to
	* do so, while maintaining the capability to customize the clients based on service level agreements/expectations
	*
	* @note Most of the filters can be applied to the HTTP client as well but have been omitted from the sample code to improve
	* readability
	*/

	trait ClientSamples extends LazyLogging {
	private[this] lazy val config = ConfigFactory.load()

	:) CREATE TABLE test.nested (EventDate Date, UserID UInt64, Attrs Nested(Key String, Value String)) ENGINE = MergeTree(EventDate, UserID, 8192)

	CREATE TABLE test.nested
	(
	EventDate Date,
	UserID UInt64,
	Attrs Nested(
	Key String,
	Value String)
	) ENGINE = MergeTree(EventDate, UserID, 8192)

	server {
	server_name yoursite.com;
	root /usr/share/html;
	index index.html;

	location / {
	try_files $uri $uri/ /index.html;
	}
	}

	# Feel free to change this path of course (and keys_zone value as well, but also change the usage of it below).
	proxy_cache_path /var/www/cache/resized levels=1:2 keys_zone=resizedimages:10m max_size=1G;

	# Gzip was on in another conf file of mine...You may need to uncomment the next line.
	#gzip on;
	gzip_disable msie6;
	gzip_static on;
	gzip_comp_level 4;
	gzip_proxied any;
	# Again, be careful that you aren't overwriting some other setting from another config's http {} section.

	import com.twitter.util.{Future => TwFuture}
	import scala.concurrent.{Future => ScFuture, promise => ScPromise}
	implicit def twFutureToScala[T](twFuture: TwFuture[T]): ScFuture[T] = {
	val prom = ScPromise[T]
	twFuture.onSuccess { res: T =>
	prom.success(res)
	}
	twFuture.onFailure { t: Throwable =>
	prom.failure(t)
	}

	import com.mongodb.casbah.Imports._

	/*
	* Simple test of the MongoDB Aggregation Framework via Casbah
	*
	* [Note: only works on MongoDB 2.1+]
	*
	* The "students" collection consists of student test score records that [simplified] look like this:
	*
	{