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.

samselikoff

Remix's useFetcher doesn't return a Promise for any of its methods (like fetcher.submit()) because Remix doesn't want you to explicitly await anything so they can handle things like cancellation for you. Instead, they recommend adding a useEffect and performing whatever logic you need to after the fetcher is in a particular state.

I found using an effect to run some logic after a submission to be too indirect, and there seem to be plenty of cases where you want to submit a form and then perform some other work on the client (sometimes async, like requesting the user's permission for their location), and I'd rather just do that after a submission in the event handler rather than an effect.

So here's a proof of concept hook that wraps Remix's useFetcher and returns a version of submit that is a promise, and resolves with the data from the action:

function useFetcherWithPromise() {
  let resolveRef = useRef();
  let promiseRef = useRef();

brunneis

1. systemctl stop docker
2. Create the file /etc/docker/daemon.json with the following content:

{
  "storage-driver": "overlay2",
  "storage-opts": ["overlay2.override_kernel_check=true"]
}

3. systemctl start docker

guillemcanal

Build PHP5 iconv extension from source for Alpine

If you plan to use iconv() to transliterate string in you PHP project, please, take this:

FROM alpine:3.4

RUN apk add --update php5-cli wget build-base php5-dev autoconf re2c libtool \

satyadeepk

#Ref: https://github.com/circleci/android-cloud-test-lab/blob/master/circle.yml

export DIRECTORY="/var/jenkins_home/GoogleCloudSDK/google-cloud-sdk/bin"
if [ ! -d "$DIRECTORY" ]; then
  # Control will enter here if $DIRECTORY doesn't exist.
	cd /var/jenkins_home
	wget https://dl.google.com/dl/cloudsdk/release/google-cloud-sdk.zip -O google-cloud-sdk.zip
	unzip -o google-cloud-sdk.zip -d ./GoogleCloudSDK/
	./GoogleCloudSDK/google-cloud-sdk/install.sh
	

michiel

#
# Wide-open CORS config for nginx
#
location / {

     if ($request_method = 'OPTIONS') {

        add_header 'Access-Control-Allow-Origin' '*';
        
        #