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.

Aldaviva

Download

1. Go to https://store.rg-adguard.net.
2. Search for the HEVC Video Extensions from Device Manufacturer (Microsoft.HEVCVideoExtension) app by entering the following store URL into the search box and then clicking the ✔ button.

   https://www.microsoft.com/en-us/p/hevc-video-extensions-from-device-manufacturer/9n4wgh0z6vhq
   

   • Do not use the normal HEVC Video Extensions app URL, because that can't play HEVC videos in Windows Media Player for some inscrutable reason.
3. Download the AppxBundle file for the version you want.
   • You may have to right click › Save Link As because the URL scheme is http, not https, if your browser is set to enforce HTTPS-only mode.

• If it tries to save as a filename that's just a GUID, you may copy the correct .AppxBundle filename and save it as that instead.

zorth64

// Variables used by Scriptable.
// These must be at the very top of the file. Do not edit.
const LUNAR_MONTH = 29.530588853;
let RAD = Math.PI / 180,
dayMilliseconds = 86400000,
J2000 = 2451545,
J1970 = 2440587.5,
E = RAD * 23.4397;

const getJulianDate = (date = new Date()) => {

k16shikano

name	japanese-tech-writing
description	日本語の技術文書・書籍原稿の文章規範。整形（一文一行、引用ブロック、脚注、コラム記法）、段落と論証の構成（パラグラフライティング）、論証の厳密さ（ツッコミどころの除去）、読み手の負荷の管理、視点と語り、演出の抑制、LLM っぽい空句の禁止、冗長の排除を定める。日本語で技術書の章、草稿、記事、解説文を書くとき、または推敲・リライトするときに使用する。

日本語技術文書の文章規範

日本語で技術的な原稿（書籍の章、記事、解説文）を書く・推敲するときは、以下の規範に従う。

整形

SaltyAom

You are the orchestrator.

You think, design and plan architectural decision only. Write only essential technical specs and reasoning. Prefer not to write code yourself unless all executors fail.

Never assume the user's intent. Ask for clarification for ambiguous situations.

Executors:
- fellow (Fable): second opinion for architectural or complex decisions. For hardest problem only.
- mechanical (Sonnet): mechanical for fast/small redundant tasks.
- thinker (Opus): reasoning-heavy, help you review/verify your thinking along side Codex gpt-5.6-sol below.

devinschumacher

Skool Video Downloader v2.0.0 | How to Download Skool Classroom Video Content for Offline Viewing

🔗 Links

Estarossa97

/*
 * Abstraction.js 2020 Edition
 *
 * Copyright (c) 2020 "Cowboy" Ben Alman
 * Licensed under the MIT license.
 * http://benalman.com/about/license/
 */

let $elseif, $else, $if = state => state ? (
  $elseif = () => () => {},

Flynsarmy

shader_type canvas_item;

uniform vec4 color : source_color = vec4(1.);
uniform float thickness : hint_range(0.1, 1.) = .4;
uniform float ratio : hint_range(0., 1.) = .5;
uniform vec4 segment_color : source_color = vec4(1., .886, .412, 1.);
uniform int segment_count : hint_range(1, 12, 1) = 1;
uniform int segment_selected : hint_range(0, 12, 1) = 0;

// "Ring - exact" signed distance function - https://iquilezles.org/articles/distfunctions2d/