Artem Belevich Artem-B

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.

Fix: Antigravity App Login Crash (BigInt Serialization Error)

The Problem

When attempting to sign in to the Antigravity desktop app via Google OAuth, the app fails to open the browser for authentication. No visible error is shown in the user interface, but the login process silently stalls.

Diagnosing the Issue

To understand what is happening behind the scenes, you can run the application with Node's inspector attached.

AGENTS.md

Purpose

This project uses LLM-assisted reverse engineering with Ghidra and Ghidra MCP to analyze a DOS game binary. The goal is to progressively recover meaningful program structure by tracing execution from the application entry point, identifying functions, variables, globals, data structures, and subsystem boundaries, and renaming only when there is high confidence.

This repository also maintains an ARCHITECTURE.md file that records confirmed subsystem discoveries and their relationships.

Accuracy matters more than speed. Never guess.

Which GGUF is right for me? (Opinionated)

Good question! I am collecting human data on how quantization affects outputs. See here for more information: ggml-org/llama.cpp#5962

In the meantime, use the largest that fully fits in your GPU. If you can comfortably fit Q4_K_S, try using a model with more parameters.

llama.cpp feature matrix

See the wiki upstream: https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix

For awhile now, each time I got a new Windows laptop I would dig up strasis gist on how to set up agent forwarding for SSH in WSL2 -- but recently I tried to point someone else at it and they were very confused by it, so this is my attempt at simpler instructions.

Installation

With Chocolatey, you must use an elevated PowerShell session. If there's no choco command found, it will fall back to winget for the npiperelay install. To force using Winget even if you have choco installed, you need to download it, so you can pass parameters to it.

Easy mode: just run this in PowerShell:

Tesla Key Card Protocol

Researched by Robert Quattlebaum darco@deepdarc.com.

Last updated 2020-02-03.

Foreward

This document was originally written several years ago. At the time I was working as an execution core verification engineer at Arm. The following points are coloured heavily by working in and around the execution cores of various processors. Apply a pinch of salt; points contain varying degrees of opinion.

It is still my opinion that RISC-V could be much better designed; though I will also say that if I was building a 32 or 64-bit CPU today I'd likely implement the architecture to benefit from the existing tooling.

Mostly based upon the RISC-V ISA spec v2.0. Some updates have been made for v2.2

Original Foreword: Some Opinion

The RISC-V ISA has pursued minimalism to a fault. There is a large emphasis on minimizing instruction count, normalizing encoding, etc. This pursuit of minimalism has resulted in false orthogonalities (such as reusing the same instruction for branches, calls and returns) and a requirement for superfluous instructions which impacts code density both in terms of size and

	# Author: Giovanni Mazzeo (github.com/thelinuxguy)
	# Script fetches your TeslaFi.com user data to allow importing into TeslaMate.

	# Updated 04/05/2025 to include more fields to normalize based on comments in gist.

	# My script fixes a couple bugs and issues seen by other people running the older script:
	# 1) "Invalid CSV delimiter" issue: https://github.com/teslamate-org/teslamate/issues/4569
	# 2) "battery_level" column integeter data change in 2024. https://github.com/teslamate-org/teslamate/issues/4477

	# You can thank me by buying me a coffee :)

	import requests, pickle
	from datetime import datetime
	import json, pytz
	import pandas as pd
	from influxdb import DataFrameClient

	# Set "static" variables
	login_url = "https://monitoring.solaredge.com/solaredge-apigw/api/login"
	panels_url = "https://monitoring.solaredge.com/solaredge-web/p/playbackData"
	DAILY_DATA = "4"

	import sympy

	def custom_latex_printer(exp, **options):
	from google.colab.output._publish import javascript
	url = "https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.3/latest.js?config=default"
	javascript(url=url)
	return sympy.printing.latex(exp, **options)

	sympy.init_printing(use_latex="mathjax", latex_printer=custom_latex_printer)