thetechoddbug (José María Gutiérrez) TheTechOddBug

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.

Full desktop shell in WSL2 using WSLg (Wayland)

Note

If you want a more complete experience and already tried the Xvnc tutorial, maybe you should try the WSLg (Xwayland) instead.

Warning

This is a work in progress tutorial. Things done here may break existing functionality, so be careful!

In this tutorial, we will install and use a full GNOME Desktop environment in WSL2, without any external software. The only requirement is a working WSLg installation. At the moment, the instructions are only for Ubuntu (20.04, 22.04 and 24.04) distros and GNOME, but you can request me to test other distros and desktop environments.

Converting JSON to OpenAPI / Swagger spec

Exploring tools that allow converting a JSON response automagically into an OpenAPI / Swagger spec.

Test JSON Response

{
    "accounts": {
        "default": {

Usage

Set a system environment variable AUTHZ_JSON_URL that points to a JSON file with the following structure:

{
  "admins": [
    "groupA", 
    "userA"
 ],

This is a snippet that will create a new user in jenkins and if security has been disabled , it will enable it :)

import jenkins.model.*
import hudson.security.*

def instance = Jenkins.getInstance()

def hudsonRealm = new HudsonPrivateSecurityRealm(false)
hudsonRealm.createAccount("MyUSERNAME","MyPASSWORD")
instance.setSecurityRealm(hudsonRealm)

	# Symbolic Representation of Prompt

	# Initialization: Define Universal State
	Ψ(t) ∈ H # Ψ(t): State vector in Hilbert space H

	# Field Configuration Space
	M = { (g, φ) \| g ∈ G, φ ∈ Φ } # G: Symmetry group, Φ: Field space
	μ : M → ℝ^+ # Measure on configuration space

	# Complexity Operator

	# Step 1: Represent Universe State
	Initialize Ψ(t) in Hilbert space H

	# Step 2: Define Field Configurations
	Define configuration space M with measure μ
	For each (g, φ) in M:
	Represent fields as algebraic structures (groups, rings, etc.)

	# Step 3: Complexity Operator
	Define operator T acting on Ψ(t) to extract complexity

	import openai # pip install openai
	import typer # pip install "typer[all]"
	from rich import print # pip install rich
	from rich.table import Table

	"""
	Webs de interés:
	- Módulo OpenAI: https://github.com/openai/openai-python
	- Documentación API ChatGPT: https://platform.openai.com/docs/api-reference/chat
	- Typer: https://typer.tiangolo.com

	# Source: https://gist.github.com/c7cdfef142bd65cc744789d3c1e90170

	###########################################
	# Talos Linux: OS Designed For Kubernetes #
	# https://youtu.be/iEFb2Zg4xUg #
	###########################################

	# Additional Info:
	# - Talos Linux: https://www.talos.dev/
	# - How To Create, Provision, And Operate Kubernetes With Cluster API (CAPI): https://youtu.be/8yUDUhZ6ako

	#!/bin/bash

	gource \
	-s .03 \
	-1280x720 \
	--auto-skip-seconds .1 \
	--multi-sampling \
	--stop-at-end \
	--key \
	--highlight-users \