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.

lunks

---
version: 2.1
aliases:
  - &install_yarn_dependencies
    run:
      name: Install yarn dependencies
      command: yarn install --frozen-lockfile --ignore-engines
  - &update_yarn
    run:
      name: Update yarn

ryansimms

Deploying to Elastic Beanstalk via CircleCi 2.0

I got to here after spending hours trying to deploy to an Elastic Beanstalk instance via CircleCi 2.0 so I thought I'd write up what worked for me to hopefully help others. Shout out to RobertoSchneiders who's steps for getting it to work with CircleCi 1.0 were my starting point.

For the record, I'm not the most server-savvy of developers so there may be a better way of doing this.

Setup a user on AWS IAM to use for deployments

• Add user here

yossorion

What I Wish I'd Known About Equity Before Joining A Unicorn

Disclaimer: This piece is written anonymously. The names of a few particular companies are mentioned, but as common examples only.

This is a short write-up on things that I wish I'd known and considered before joining a private company (aka startup, aka unicorn in some cases). I'm not trying to make the case that you should never join a private company, but the power imbalance between founder and employee is extreme, and that potential candidates would

yefim

{
  "AWSEBDockerrunVersion": "1",
  "Image": {
    "Name": "<AWS_ACCOUNT_ID>.dkr.ecr.us-east-1.amazonaws.com/<NAME>:<TAG>",
    "Update": "true"
  },
  "Ports": [
    {
      "ContainerPort": "443"
    }

spalladino

# Backup
docker exec CONTAINER /usr/bin/mysqldump -u root --password=root DATABASE > backup.sql

# Restore
cat backup.sql | docker exec -i CONTAINER /usr/bin/mysql -u root --password=root DATABASE

jeffweiss

defmodule MyApp.Mixfile do
  use Mix.Project

  def project do
    [app: :my_app,
     version: get_version,
     elixir: "~> 1.0",
     elixirc_paths: elixirc_paths(Mix.env),
     compilers: Mix.compilers,
     build_embedded: Mix.env == :prod,

RobertoSchneiders

I am deploying with this IAM using Codeship and Circle CI to Elastic Beanstalk. I had a lot of trouble with this config. I talked to the aws support for about 6 hours until this worked properly, so, I guess it is worth to share.

UPDATE: In the end, I have to use the AWSElasticBeanstalkFullAccess policy. My custom policy keep breaking every week with some new added permission or some EB internal change. Anyway, the IAM I was using is below.

This works for me with CircleCI and EB Cli.

{
    "Version": "2012-10-17",
    "Statement": [
        {

iamralch

import (
	"archive/tar"
	"fmt"
	"io"
	"os"
	"path/filepath"
	"strings"
)

func tarit(source, target string) error {

magnetikonline

AWS Elastic Beanstalk deploy user restricted IAM policy

An IAM user policy document to give minimal rights for deploying an Elastic Beanstalk application.

Where:

• REGION: AWS region.
• ACCOUNT_ID: AWS account ID.
• APPLICATION_NAME: Desired target Elastic Beanstalk application name(space).
• IAM_INSTANCE_PROFILE_ROLE: The instance profile (IAM role) Elastic Beanstalk EC2 instaces will run under.