Mart Sõmermaa mrts

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.

TL;DR

Install Postgres 10, and then:

sudo pg_dropcluster 10 main --stop
sudo pg_upgradecluster 9.6 main
sudo pg_dropcluster 9.6 main

Effective Modern CMake

Getting Started

For a brief user-level introduction to CMake, watch C++ Weekly, Episode 78, Intro to CMake by Jason Turner. LLVM’s CMake Primer provides a good high-level introduction to the CMake syntax. Go read it now.

After that, watch Mathieu Ropert’s CppCon 2017 talk Using Modern CMake Patterns to Enforce a Good Modular Design (slides). It provides a thorough explanation of what modern CMake is and why it is so much better than “old school” CMake. The modular design ideas in this talk are based on the book [Large-Scale C++ Software Design](https://www.amazon.de/Large-Scale-Soft

How to convert markdown to PDF:

This post reviews several methods for converting a Markdown (.md) formatted file to PDF, from UNIX or Linux machines.

Using `Pandoc`:

$ pandoc How_I_got_svg-resizer_working_on_Mac_OSX.md -s -o test1.pdf

web ok? here take this with you

https://github.com/pulls?user=matrix-hacks replace matrix-hacks with your own

Recently CSS has got a lot of negativity. But I would like to defend it and show, that with good naming convention CSS works pretty well.

My 3 developers team has just developed React.js application with 7668 lines of CSS (and just 2 !important). During one year of development we had 0 issues with CSS. No refactoring typos, no style leaks, no performance problems, possibly, it is the most stable part of our application.

Here are main principles we use to write CSS for modern (IE11+) browsers:

SUIT CSS naming conventions + SUIT CSS design principles;
PostCSS + CSSNext. Future CSS syntax like variables, nesting, and autoprefixer are good enough;
Flexbox is awesome. No need for grid framework;
Normalize.css, base styles and variables are solid foundation for all components;

	import java.util.stream.Collectors;

	import org.springframework.data.domain.Sort;
	import org.springframework.data.domain.Sort.Order;

	import com.vaadin.flow.data.provider.Query;
	import com.vaadin.flow.data.provider.SortDirection;

	public class SpringDataVaadinUtil {

	from bluepy.btle import Scanner, DefaultDelegate

	class ScanDelegate(DefaultDelegate):
	def __init__(self):
	DefaultDelegate.__init__(self)

	def handleDiscovery(self, dev, isNewDev, isNewData):
	if isNewDev:
	print "Discovered device", dev.addr
	elif isNewData:

	bind: 0.0.0.0:8000
	workers: 1
	worker_class: "example.worker:CustomWorker"
	timeout: 30
	ca_certs: ca.crt
	certfile: server.crt
	keyfile: server.key
	cert_reqs: 2
	do_handshake_on_connect: true

	/**
	Typora user config
	C:\Users\hava\AppData\Roaming\Typora\conf\conf.user.json */
	{
	"width" : null, // Integer - Window's width in pixels. Default is null (last window width)
	"height" : null, // Integer - Window's height in pixels. Default is null (last window height)
	"directWrite": true, // Boolean - Enables DirectWrite font rendering system on Windows. Default is true.
	"defaultFontFamily": {
	"standard": null, //String - Defaults to "Times New Roman".
	"serif": null, // String - Defaults to "Times New Roman".