Arnab Chakraborty arnabc

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.

What I Wish I'd Known About Equity Before Joining A Unicorn

Disclaimer: This piece is written anonymously. The names of a few particular companies are mentioned, but as common examples only.

This is a short write-up on things that I wish I'd known and considered before joining a private company (aka startup, aka unicorn in some cases). I'm not trying to make the case that you should never join a private company, but the power imbalance between founder and employee is extreme, and that potential candidates would

A quick "how to" on what you need to do to both setup AND recover a single-server PostgreSQL database using WAL-E

WAL-E: https://github.com/wal-e/wal-e
Assuming Ubuntu 12.04 LTS ("Precise")
We'll be using S3. Make sure you have an IAM in a group with GetObject, ListBucket and PutObject on the bucket you want to use (and that it's not public).

Setup:

These packages:

	ruby '2.7.1'

	gem 'rails', github: 'rails/rails'
	gem 'tzinfo-data', '>= 1.2016.7' # Don't rely on OSX/Linux timezone data

	# Action Text
	gem 'actiontext', github: 'basecamp/actiontext', ref: 'okra'
	gem 'okra', github: 'basecamp/okra'

	# Drivers

	#!/usr/bin/env node

	/**
	* After prepare, files are copied to the platforms/[platform] folder.
	* Lets clean up some of those files that arent needed with this hook.
	*/
	var fs = require('fs');
	var path = require('path');

	var rootdir = process.argv[2];

	/**
	* Fancy ID generator that creates 20-character string identifiers with the following properties:
	*
	* 1. They're based on timestamp so that they sort after any existing ids.
	* 2. They contain 72-bits of random data after the timestamp so that IDs won't collide with other clients' IDs.
	* 3. They sort lexicographically (so the timestamp is converted to characters that will sort properly).
	* 4. They're monotonically increasing. Even if you generate more than one in the same timestamp, the
	* latter ones will sort after the former ones. We do this by using the previous random bits
	* but "incrementing" them by 1 (only in the case of a timestamp collision).
	*/

	/*
	* Copyright (c) 2015 Fran Montiel
	*
	* Licensed under the Apache License, Version 2.0 (the "License");
	* you may not use this file except in compliance with the License.
	* You may obtain a copy of the License at
	*
	* http://www.apache.org/licenses/LICENSE-2.0
	*
	* Unless required by applicable law or agreed to in writing, software

	/*
	* Copyright 2014 Johannes Homeier
	*
	* Licensed under the Apache License, Version 2.0 (the "License");
	* you may not use this file except in compliance with the License.
	* You may obtain a copy of the License at
	*
	* http://www.apache.org/licenses/LICENSE-2.0
	*
	* Unless required by applicable law or agreed to in writing, software

	package your_package;

	import android.animation.AnimatorSet;
	import android.animation.ObjectAnimator;
	import android.app.Activity;
	import android.content.Context;
	import android.graphics.Bitmap;
	import android.graphics.Canvas;
	import android.graphics.Color;
	import android.graphics.Paint;

	def toCamelCase(String string) {
	String result = ""
	string.findAll("[^\\W]+") { String word ->
	result += word.capitalize()
	}
	return result
	}

	afterEvaluate { project ->
	Configuration runtimeConfiguration = project.configurations.getByName('compile')