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.

steven-tey

import { BlurImage, YouTube } from "@dub/ui";
import { nFormatter } from "@dub/utils";
import { Eye, UserCheck, Video } from "lucide-react";
import { Suspense } from "react";

export function YoutubeChannel({ id }: { id: string }) {
  return (
    <Suspense fallback={<div className="not-prose grid gap-4"></div>}>
      <YoutubeChannelRSC id={id} />
    </Suspense>

jh3y

:root {
    --none: linear(0, 1);
    --power1-in: linear( 0, 0.0039, 0.0156, 0.0352, 0.0625, 0.0977, 0.1407, 0.1914, 0.2499, 0.3164, 0.3906 62.5%, 0.5625, 0.7656, 1 );
    --power1-out: linear( 0, 0.2342, 0.4374, 0.6093 37.49%, 0.6835, 0.7499, 0.8086, 0.8593, 0.9023, 0.9375, 0.9648, 0.9844, 0.9961, 1 );
    --power1-in-out: linear( 0, 0.0027, 0.0106 7.29%, 0.0425, 0.0957, 0.1701 29.16%, 0.2477, 0.3401 41.23%, 0.5982 55.18%, 0.7044 61.56%, 0.7987, 0.875 75%, 0.9297, 0.9687, 0.9922, 1 );
    --power2-in: linear( 0, 0.0014 11.11%, 0.0071 19.24%, 0.0188 26.6%, 0.037 33.33%, 0.0634 39.87%, 0.0978 46.07%, 0.1407 52.02%, 0.1925 57.74%, 0.2559 63.49%, 0.3295 69.07%, 0.4135 74.5%, 0.5083 79.81%, 0.6141 85%, 0.7312 90.09%, 1 );
    --power2-out: linear( 0, 0.2688 9.91%, 0.3859 15%, 0.4917 20.19%, 0.5865 25.5%, 0.6705 30.93%, 0.7441 36.51%, 0.8075 42.26%, 0.8593 47.98%, 0.9022 53.93%, 0.9366 60.13%, 0.963 66.67%, 0.9812 73.4%, 0.9929 80.76%, 0.9986 88.89%, 1 );
    --power2-in-out: linear( 0, 0.0036 9.62%, 0.0185 16.66

wesbos

import dotenv from "dotenv";
import { readFile, writeFile } from 'fs/promises';
dotenv.config();
import { Configuration, OpenAIApi } from "openai";

const configuration = new Configuration({
  apiKey: process.env.OPENAI_API_KEY,
});
export const openai = new OpenAIApi(configuration);

sindresorhus

Pure ESM package

The package that linked you here is now pure ESM. It cannot be require()'d from CommonJS.

This means you have the following choices:

1. Use ESM yourself. (preferred)
   Use import foo from 'foo' instead of const foo = require('foo') to import the package. You also need to put "type": "module" in your package.json and more. Follow the below guide.
2. If the package is used in an async context, you could use await import(…) from CommonJS instead of require(…).
3. Stay on the existing version of the package until you can move to ESM.

rauschma

Strings

String.prototype.*

None of the string methods modify this – they always return fresh strings.

• charAt(pos: number): string ^ES1

  Returns the character at index pos, as a string (JavaScript does not have a datatype for characters). str[i] is equivalent to str.charAt(i) and more concise (caveat: may not work on old engines).

rhukster

#!/bin/bash
# Creator: Phil Cook
# Modified: Andy Miller
#
# >>> IMPORTANT: Moved to: https://github.com/rhukster/sphp.sh
# >>> Kept here for legacy purposes
#
osx_major_version=$(sw_vers -productVersion | cut -d. -f1)
osx_minor_version=$(sw_vers -productVersion | cut -d. -f2)
osx_patch_version=$(sw_vers -productVersion | cut -d. -f3)

rauschma

This content moved here: https://exploringjs.com/impatient-js/ch_arrays.html#quickref-arrays

cecilemuller

{
  "version": "0.2.0",
  "configurations": [
    {
      "name": "Example",
      "type": "node",
      "request": "launch",
      "runtimeExecutable": "node",
      "runtimeArgs": ["--nolazy", "-r", "ts-node/register/transpile-only"],

schnerd

const CDP = require('chrome-remote-interface');
const argv = require('minimist')(process.argv.slice(2));
const file = require('fs');

// CLI Args
const url = argv.url || 'https://www.google.com';
const format = argv.format === 'jpeg' ? 'jpeg' : 'png';
const viewportWidth = argv.viewportWidth || 1440;
const viewportHeight = argv.viewportHeight || 900;
const delay = argv.delay || 0;