Florian Kammermann floriankammermann

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.

Migrating from Ubuntu on WSL to ArchLinux on WSL2

Obsolete notice

This document was created back in 2020 and might not be actual nowadays. It is not supported anymore, so use thise information at your own risk.

Upgrading to WSL 2

Download WSL2 Kernel
run wsl --set-default-version 2 in windows command line, so that all future WSL machine will use WSL2.

How to create star history url's from cncf landscape

get the landscape: curl https://raw.githubusercontent.com/cncf/landscape/master/landscape.yml > landscape.yml
split into subcategories: csplit --digits=2 --quiet --prefix=outfile landscape.yml "/- subcategory:/+1" "{*}"
remove outfile00, it has no information: rm outfile00
rename to subcategory: for f in `ls | grep outfile`; do new_name=$(head $f | grep "^\s\s\s\s\s\s\s\sname:"| sed 's/.*name:\s//' | sed 's/\s&\s/_/' | tr ' ' '_' | tr '/' '_'); mv $f $new_name; done
create url per subcategory: for f in `ls | grep -v landscape.yml`; do params=$(cat $f | grep repo_url | sed 's/^.*github.com\///' | sed 's/.$//' | tr '\n' '&' | sed 's/.$//'); echo $f: https://star-history.t9t.io/#$params; done

After installing mitmproxy run it (just type mitmproxy) in a terminal session and quit. This will create the necessaries certificates files at ~/.mitmproxy.
Extract the certificate to .crt format:
openssl x509 -in ~/.mitmproxy/mitmproxy-ca.pem -inform PEM -out ca.crt
Trust the certificate into CA:
sudo trust anchor ca.crt
Run the mitmproxy again

Stevey's Google Platforms Rant

I was at Amazon for about six and a half years, and now I've been at Google for that long. One thing that struck me immediately about the two companies -- an impression that has been reinforced almost daily -- is that Amazon does everything wrong, and Google does everything right. Sure, it's a sweeping generalization, but a surprisingly accurate one. It's pretty crazy. There are probably a hundred or even two hundred different ways you can compare the two companies, and Google is superior in all but three of them, if I recall correctly. I actually did a spreadsheet at one point but Legal wouldn't let me show it to anyone, even though recruiting loved it.

I mean, just to give you a very brief taste: Amazon's recruiting process is fundamentally flawed by having teams hire for themselves, so their hiring bar is incredibly inconsistent across teams, despite various efforts they've made to level it out. And their operations are a mess; they don't real

	import javax.net.ssl.*;
	import java.io.IOException;
	import java.net.*;
	import java.security.cert.X509Certificate;
	import java.util.Iterator;
	import java.util.List;
	import java.util.Map;

	public class URLTest {

	package main

	import (
	"net/http"
	)

	func main() {
	finish := make(chan bool)

	server8001 := http.NewServeMux()

	import javax.net.ssl.SSLParameters;
	import javax.net.ssl.SSLSocket;
	import javax.net.ssl.SSLSocketFactory;
	import java.io.*;

	/** Establish a SSL connection to a host and port, writes a byte and
	* prints the response. See
	* http://confluence.atlassian.com/display/JIRA/Connecting+to+SSL+services
	*/
	public class SSLPoke {

	#!/bin/bash
	# bash generate random alphanumeric string
	#

	# bash generate random 32 character alphanumeric string (upper and lowercase) and
	NEW_UUID=$(cat /dev/urandom \| tr -dc 'a-zA-Z0-9' \| fold -w 32 \| head -n 1)

	# bash generate random 32 character alphanumeric string (lowercase only)
	cat /dev/urandom \| tr -dc 'a-z0-9' \| fold -w 32 \| head -n 1