Særgeir serpro69

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.

If you're using a high-end bluetooth headset on your Macbook Pro it's likely your mac is using an audio codec which favors battery efficiency over high quality. This results in a drastic degradation of sound, the SBC codec is the likely culprit, read more about it here.

Find out what codec you're using

Play a song on your headphones
Option (⌥) click the Bluetooth button at the top of your screen
If you're using AAC or aptX, you can stop here—those are the highest quality codecs.

Change your codec to AAC or aptX

2022

Daily Journaling

const dayInMonth = moment().format("D");
const currentMonth = moment().format("M");
var i = 1;
while (currentMonth > i) {
    var entries = entriesSum(i);
    var max = daysInMonth(i);
    var perc = entries + "/" + max;

Usage

Download and use the capacities-cards.css as a snippet
Copy the raw value of the capacities-index.md file and paste it in your obsidian note

3. Edit the callout to your liking

Callouts

capacities-card: Base container for the card

capacities-index: Container for side index

	#!/usr/bin/env python3
	"""
	cc-sessions — Fast CLI to search, browse & resume Claude Code session history

	OVERVIEW
	--------
	Claude Code stores all conversation history locally in ~/.claude/projects/ as JSONL files.
	This tool indexes those sessions for fast search and provides a clean CLI interface to:
	- Search by keyword across all conversations
	- Filter by date, branch, or project

	/*
	This DataviewJS View helps to add functionality to a Dataview Table.
	THIS IS A BETA RELEASE - there are bugs and its not feature-complete
	Version: v0.6.0

	Given a Dataview Query, this view will
	- Execute the query
	- Extract the headers/columns of the result
	- Render filters to adjust the Dataview table results
	- Render the table results, filtering and re-rendering on filter changes

	```dataviewjs
	// find dates based on format [[YYYY-MM-DD]]
	const findDated = (task)=>{
	if( !task.completed ) {
	task.link = " " + "[[" + task.path + "\|*]]";
	task.date="";
	const found = task.text.match(/\[\[([12]\d{3}-(0[1-9]\|1[0-2])-(0[1-9]\|[12]\d\|3[01]))\]\]/);
	if(found) task.date = moment(found[1]);
	return true;
	}

	<%*
	/*
	# Hypothes.idian a templater script for retrieving annotations from Hypothes.is
	Dev: TfTHacker https://twitter.com/TfTHacker

	# Prerequisites:
	+ Templater plugin by https://github.com/SilentVoid13/Templater
	+ Free Hypothes.is developer token from: https://hypothes.is/account/developer
	+ This script will prompt you for his token and save it to a file called "hypothesis config.md"
	+ This file store your configuration and can be located any where in your vault.

	/* ——————————————————————————————————————————— ╗
	STYLE SETTINGS
	╚ ——————————————————————————————————————————— */
	/* @settings

	name: Rainbow Tags by raisabelatrix
	id: rb-tags-rsbltrx
	settings:
	-
	id: rainbow-tags

	.card-view {
	--card-width:35em; /adjust this size if you want narrower or wider cards/
	/*unit has to be in em so that card background will scale along with your font
	when you zoom in and out*/
	}

	/The Card/
	.card-view .markdown-preview-section {
	background-color: var(--background-primary);
	margin: 3% auto;