Chris rot26

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.

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:

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.
If the package is used in an async context, you could use await import(…) from CommonJS instead of require(…).
Stay on the existing version of the package until you can move to ESM.

Using Multus CNI in K3S

By default, K3S will run with flannel as the CNI and use custom directories to store CNI plugin binaries and config files(You can inspect the kubelet args K3S uses via journalctl -u k3s|grep cni-conf-dir). So you need to configure that properly When deploying Multus CNI.

For example given the official Multus manifests in https://github.com/intel/multus-cni/blob/36f2fd64e0965e639a0f1d17ab754f0130951aba/images/multus-daemonset.yml, the following changes are needed:

volumes:
 - name: cni

Comparison of ZSH frameworks and plugin managers

Changelog

update 1: add a FAQ section
update 2: benchmark chart and feature comparison table
update 3:
- improve the table with missing features for antigen
- new zplg times result

TLDR

Plain JavaScript objects are better than classes when they can be used, and many popular modern frameworks have adopted their use.

Consider that in React a component can be created as either a class or as an object.

// using a class
class Welcome extends React.Component {
  render() {
 Hello, {this.props.name}

	blueprint:
	name: Sensor Light
	description: >
	# 💡 Sensor Light

	Version: 8.5


	Your lighting experience, your way - take control and customize it to perfection! 💡✨

	'''
	Place this file in your lambda package next to the handler file.

	Setting the log formatter on lambda loses the request id out of the context.
	But, you often want to use a logger across a few files and name them
	or change the log format.

	This factory will ensure that the lambda context id is available to any
	log format you choose. It also has the ability to filter out strings
	from the logs using a regular expression.


	'''
	Unit test setup automatically called
	'''
	from unittest import mock
	import pytest

	# Fixture, autouse
	@pytest.fixture(scope='session', autouse=True)
	def patch_boto3(request):

	# Default npm flag settings
	# Put this `.npmrc` file in your npm project root next to `package.json`

	# Always save packages that get installed to avoid "missing dependencies"
	save = true

	# Always save the EXACT version that you ran and installed.
	# upgrade to latest version on each new version (put `npm outdated` in your CI)
	# `npm outdated && yarn upgrade --latest && npm outdated`
	save-exact = true

	# convert .xlsx to .xlsx.csv using https://github.com/dilshod/xlsx2csv
	pip install xlsx2csv
	# (shell-fu from http://stackoverflow.com/a/12965604)
	find . -iname "*.xlsx" -exec sh -c 'xlsx2csv "$1" > "$1.csv"' x {} \;

	# grep .csv files
	brew install ripgrep
	rg -i -g "*.csv" "waldo"

	# ...or plain ole grep