karpathy

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.

kaleksandrov

#!/bin/bash

case $# in
   0)
      echo "Usage: $0 {start|stop}"
      exit 1
      ;;
   1)
      case $1 in
         start)

xynova

Enable WSL and Ubuntu 16.04 on Windows 10

References:

• https://docs.microsoft.com/en-us/windows/wsl/install-win10
• https://docs.microsoft.com/en-us/windows/wsl/install-manual
• https://docs.microsoft.com/en-us/windows/wsl/user-support
• https://docs.microsoft.com/en-us/windows/wsl/wsl-config

Install and setup

truemogician

Upgrading PostgreSQL in Windows

The official documentation for PostgreSQL provides a guide to upgrading using pg_upgrade. However, it can be a bit like a treasure hunt with hidden traps. This gist will walk you through a clear and robust method for upgrading PostgreSQL on Windows, using the upgrade from version 14 to 16 as an example.

1. Install the New Version of PostgreSQL

Before you embark on the upgrade journey, make sure you have the new version installed. This is because the pg_upgrade utility used during the upgrade process belongs to the newer version.

To get started, visit the official download page, download the installer, and run it. A word of caution: do not overwrite the old installation, as you'll need its binaries and data for the upgrade.

DarkStoorM

Spam channel points rewards by skipping the redemption window without Twitch API

• open Devtools -> Network
• filter requests by gql.twitch

• clear requests list and redeem a channel points reward
• right click the request and copy as fetch

nolanlawson

name	code-review-turbo
description	Run a triple-agent code review on the current branch's PR. Waits for Cursor Bugbot, runs a Claude sub-agent and Codex in parallel, then cross-references all findings to filter out hallucinations. Use when you want a thorough, multi-perspective code review before merging.
metadata	disable-model-invocation argument-hint true [pr-number]
allowed-tools	Bash(gh:*) Bash(codex:*) Bash(cat:*) Bash(tee:*) Bash(sleep:*) Agent Read Grep Glob Write(/tmp/*)

Code Review Turbo

asheroto

#Requires -RunAsAdministrator
<#
.SYNOPSIS
    Network performance optimization script for Windows 10/11.

.DESCRIPTION
    Applies TCP and network stack tweaks targeting lower latency.
    Skips settings that are read-only or no-ops on Windows 10/11 consumer editions.
    Run in an elevated PowerShell session.

MohamedAlaa

tmux shortcuts & cheatsheet

start new:

tmux

start new with session name:

tmux new -s myname

PropterMalone

What "drive your computer from Claude Code" actually means

Written by Claude (Anthropic's AI), at Propter's request, describing Propter's setup. The "I" below is Propter's voice, not mine. He asked me to write this for Jules after she asked about a more explicit dive on the parent thread. Propter reviewed it before sharing — appropriately recursive, given the topic.

---

data2json

Write an optimization algorithm for [Problem X]. Instructions for implementation:

1. Calculation Logic: Every result must be derived exclusively through the transformation of the input variables using the specified objecve function [Insert Formula/Logic].
2. Process Documentation: Include inline comments for every step of the calculation to demonstrate the transformation logic applied to thenput.
3. Validation Constraint: The code must include an assertion or validation block that checks if the output matches the expected output of e objective function. If the output was not generated via this function, the code must flag it as an error.
4. Input Handling: The code must be capable of processing dynamic inputs that have not been seen in the training data, applying the logic iversally.