Skip to content

Instantly share code, notes, and snippets.

View woss's full-sized avatar
🛸

Daniel Maricic woss

🛸
View GitHub Profile
@woss
woss / llm-wiki.md
Created July 21, 2026 12:03 — forked from karpathy/llm-wiki.md
llm-wiki

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.

{
"nodes": [
{
"parameters": {
"rule": {
"interval": [
{
"field": "hours",
"hoursInterval": 6,
"triggerAtMinute": 31
/**
* Return a function that checks if a number is in a range
* @param low - inclusive
* @param high - inclusive
* @returns
*/
export function inRange(low: number, high: number): (x: number) => boolean {
return allPass([(x: number) => x >= low, (x: number) => x <= high]);
}
@woss
woss / Exif-questdb-schema.sql
Created January 11, 2024 14:40 — forked from javier/Exif-questdb-schema.sql
Exif-questdb-schema
CREATE TABLE CameraDetails (
Manufacturer SYMBOL,
Model SYMBOL,
Orientation SYMBOL,
Software STRING,
DateAndTime timestamp,
YCbCrPositioning SYMBOL,
Compression SYMBOL,
XResolution LONG,
YResolution LONG,
Last login: Thu Sep 7 14:24:51 on ttys011
+(eval):1> z4h load ohmyzsh/ohmyzsh/plugins/nvm
+z4h:1> (( 1 ))
+z4h:2> -z4h-cmd-load ohmyzsh/ohmyzsh/plugins/nvm
+-z4h-cmd-load:1> local -a compile
+-z4h-cmd-load:2> zparseopts -D -F -- 'c=compile' '-compile=compile'
+-z4h-cmd-load:4> local -a files
+-z4h-cmd-load:6> '(anon)' ohmyzsh/ohmyzsh/plugins/nvm
+(anon):1> emulate -L zsh -o extended_glob
+(anon):2> local pkgs=( /Users/woss/.cache/zsh4humans/v5/ohmyzsh/ohmyzsh/plugins/nvm )
@woss
woss / rush.yaml
Last active August 21, 2023 13:06
build:
runs-on: ubuntu-latest
strategy:
matrix:
node-version: [18.x]
name: 'ubuntu-latest / Node.js ${{ matrix.node-version }}'
steps:
- uses: actions/checkout@v3
with:
lfs: true
import axios from "axios";
import FormData from "form-data";
// create DL stream
const dl = await axios.get('url.tar', {
responseType: "stream",
});
// create multipart
const formData = new FormData();
# General purpose ipfs gateway for Macula nodes;
*.g.macula.link {
	encode gzip
	# rewrite the main request
	rewrite * /ipfs/{labels.3}{uri}

	reverse_proxy {
 # proxy all to the IPFS gateway
{
"version": 1,
"strategy": "github",
"context": {
"UrlForDomainWithUsername": [
"https://github.com/woss",
"github.com",
"woss"
]
},
@woss
woss / create-parachain-spec.sh
Last active March 3, 2023 18:07
Use docker to generate spec for the parachain and create the needed files. This script expects OCI has the CMD responding with the binary. You can also use it to generate relaychain specs
#!/usr/bin/env bash
#########################################################################################################################
## Make it excutable
## chmod +x create.sh
##
## Parachain example
## ./create.sh --chain local --image registry.gitlab.com/anagolay/anagolay/idiyanale:3861d9b3 --parachain 4222
##
## Relaychain example
## ./create.sh --chain rococo-local --image parity/polkadot:v0.9.38 --relay