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.

Stonesjtu

import gc

import torch

## MEM utils ##
def mem_report():
    '''Report the memory usage of the tensor.storage in pytorch
    Both on CPUs and GPUs are reported'''

    def _mem_report(tensors, mem_type):

chaadow

-- TerminalVim.app
-- This creates a shim Application that will enable you to open files from the Finder in vim using iTerm

-- To use this script: 
-- 1. Open Automator and create a new Application
-- 2. Add the "Run Applescript" action
-- 3. Paste this script into the Run Applescript section
-- 4. Save the application as TerminalVim.app in your Applications folder
-- 5. In the Finder, right click on a file and select "Open With". In that window you can set TerminalVim as a default
on run {input, parameters}

sos4nt

# A xterm-256color based TERMINFO that adds the escape sequences for italic.
#
# Install:
#
#   tic xterm-256color-italic.terminfo
#
# Usage:
#
#   export TERM=xterm-256color-italic
#

MohamedAlaa

tmux shortcuts & cheatsheet

start new:

tmux

start new with session name:

tmux new -s myname

hellerbarde

Latency numbers every programmer should know

L1 cache reference ......................... 0.5 ns
Branch mispredict ............................ 5 ns
L2 cache reference ........................... 7 ns
Mutex lock/unlock ........................... 25 ns
Main memory reference ...................... 100 ns             
Compress 1K bytes with Zippy ............. 3,000 ns  =   3 µs
Send 2K bytes over 1 Gbps network ....... 20,000 ns  =  20 µs
SSD random read ........................ 150,000 ns  = 150 µs

Read 1 MB sequentially from memory ..... 250,000 ns = 250 µs

jboner

Latency Comparison Numbers (~2012)
----------------------------------
L1 cache reference                           0.5 ns
Branch mispredict                            5   ns
L2 cache reference                           7   ns                      14x L1 cache
Mutex lock/unlock                           25   ns
Main memory reference                      100   ns                      20x L2 cache, 200x L1 cache
Compress 1K bytes with Zippy             3,000   ns        3 us
Send 1K bytes over 1 Gbps network       10,000   ns       10 us
Read 4K randomly from SSD*             150,000   ns      150 us          ~1GB/sec SSD