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.

karpathy

"""
The most atomic way to train and run inference for a GPT in pure, dependency-free Python.
This file is the complete algorithm.
Everything else is just efficiency.

@karpathy
"""

import os       # os.path.exists
import math     # math.log, math.exp

JeanOlivier

# Trying to do a PID for the daikin
#
# This is a WIP prototype. Only heating was tested.
# Using log.error because I can't seem to be able to find how to read other log levels.

from random import choices

ENABLED = True  # Disabled if False. If true, the PID takes over when the Daikin is on.

target_temp = 22.5

Blackshome

blueprint:
  name: Bathroom Humidity Exhaust Fan
  description: >
    # 🚿 Bathroom Humidity Exhaust Fan
    
    **Version: 2.7**
    
    
    Step into the future of freshness - customize it your way and experience the convenience of automated humidity control! 🌿🚿
    

DennisFederico

tags: #ksqldb/tutorial

---

Provisioning ksqlDB Cluster

Cluster Data

environment=env-0x3135

achetronic

#!/bin/bash

SHA256_HASH="5bb4faffc8b35e2702b2ffa78e982b979d7b66db29bd55b0c58de8fa745df661"

for i in {1..1000}
do
   echo "Looking into page: $i"

   curl "https://registry.hub.docker.com/v2/repositories/apache/superset/tags/?page=$i" \
    | jq '.results[] | select(.["images"][]["digest"] == "sha256:'${SHA256_HASH}'")'

NAR8789

# explicitly define host-ip mappings
address=/myapp.local/172.16.1.2
# dnsmasq entries are always wildcard entries, so this maps both myapp.local and *.myapp.local

# (yes, it's fine for this to be your entire dnsmasq config. the defaults are pretty sensible)

tomdaley92

Proxmox - SPICE client setup for MacOS

1. Install a working (and compiled) version of virt-viewer. You may view the homebrew package's upstream source on GitHub.

   brew tap jeffreywildman/homebrew-virt-manager
   brew install virt-viewer

2. Once that's installed should be able make a call remote-viewer with a pve-spice.vv file downloaded from proxmox web interface

egg82

Proxmox

Find the proper driver at the NVidia website.

Note: Make sure to select "Linux 64-bit" as your OS

Hit the "Search" button.

burhan

# Generate new CSR and private key
openssl req -out CSR.csr -new -newkey rsa:2048 -nodes -keyout privateKey.key

# Generate new self-signed certificate
openssl req -x509 -sha256 -nodes -days 365 -newkey rsa:2048 -keyout privateKey.key -out certificate.crt

# Generate a new CSR from existing private key
openssl req -out CSR.csr -key privateKey.key -new

# Generate new CSR from certificate (must have private key)