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.

ZPascal

require "formula"
require_relative "lib/private_strategy"

class Hoge < Formula
  homepage "https://github.com/yourcompany/hoge"
  url "https://github.com/yourcompany/hoge/releases/download/v0.1.0/hoge_v0.1.0_darwin_amd64.tar.gz", :using => GitHubPrivateRepositoryReleaseDownloadStrategy
  sha256 "6de411ff3e4b1658a413dd6181fcXXXXXXXXXXXXXXXXXXXX"
  head "https://github.com/yourcompany/hoge.git"
  version "0.1.0"

sibelius

// @flow
import { ConnectionHandler } from 'relay-runtime';
import { isObject, isArray } from 'lodash/fp';

export function listRecordRemoveUpdater({ parentId, itemId, parentFieldName, store }) {
  const parentProxy = store.get(parentId);
  const items = parentProxy.getLinkedRecords(parentFieldName);

  parentProxy.setLinkedRecords(items.filter(record => record._dataID !== itemId), parentFieldName);
}

lenkavon

This is the response to lightning talk CFP at ReactiveConf.

Sharing code between web and mobile using React + ReactNative + Redux

When you stand in front of the project that aims to come with mobile and web application, going for combo React.js & ReactNative is easy choice. But to decide how to handle the data flow is a different story.

My team made this decision and after couple of talks and earlier experience with graphQL + Relay, we went for Redux. Redux enabled us to share most of the bussines logic even though the apps differ in usecases. I would like to share the challenges we had to overcome with you.

robertknight

import { options, render } from 'preact';

import { ATTR_KEY } from 'preact/src/constants';

/**
 * Returns the input properties that were passed to an element when it was
 * rendered.
 *
 * See `shallowRender` documentation for usage example.
 */

sebmarkbage

import { Component } from "React";

export var Enhance = ComposedComponent => class extends Component {
  constructor() {
    this.state = { data: null };
  }
  componentDidMount() {
    this.setState({ data: 'Hello' });
  }
  render() {

staltz

The introduction to Reactive Programming you've been missing

(by @andrestaltz)

---

This tutorial as a series of videos

If you prefer to watch video tutorials with live-coding, then check out this series I recorded with the same contents as in this article: Egghead.io - Introduction to Reactive Programming.

---

michalskop

# -*- coding: utf-8 -*-

# calculates 18 groups

import pickle
import csv

with open("responses0", 'rb') as f:
  responses = pickle.load(f)

ogrrd

Never touch your local /etc/hosts file in OS X again

To setup your computer to work with *.test domains, e.g. project.test, awesome.test and so on, without having to add to your hosts file each time.

Requirements

• Homebrew
• Mountain Lion -> High Sierra

Install

fprochazka

<?php

/**
 * Třída existuje, aby se vůbec neukládala session, tam kde není potřeba.
 * Například v API, nebo v Cronu se různě sahá na session, i když se reálně mezi requesty nepřenáší.
 *
 * @internal
 */
class ArraySessionStorage extends Nette\Object implements Nette\Http\ISessionStorage
{