Ivan Nemytchenko inem

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.

Chat GPT "DAN" (and other "Jailbreaks")

This steps should help working remotely with Sublime Text. They are meant to be incremental, just setting up SFTP will go a long way.

My workflow

I have all my code on my laptop, edit locally and automatically push the files to my server. I never edit "tracked" files on the server directly. I sometimes modify untracked files on the server using rmate (see below).

Rails Active Storage Cheatsheet

Rails Active Storage Cheatsheet
- Install Active Storage
- Usage
Install this libraries first

A super simple bash script to check the status of a GitLab CI pipeline.

$ git push
...
$ git pipeline-status
Status of last pipeline for user/project on gitlab/master:
"pending"
...
$ git pipeline-status

A series of posts on css-in-js

0. styles as objects

First, an exercise. Can we represent all of css with plain data? Let's try.

let redText = { color: 'red' };

	L1
	Базовые знание диалекта Typescript, основных встроенных примитивов (Array, Object, Function, String, Date) и методов работы с ними, понимание контекста (this) и способов управления им
	Базовые знания фреймворка React
	Знание наиболее часто используемых Web API: fetch/XHR, Storage, PostMessage, DOM
	Умение писать простейшие алгоритмы (поиск, сортировка, сравнение объектов) без требований к их оптимальности.
	Навыки кроссбраузерной адаптивной верстки по макету с использованием существующего кода, знание CSS и препроцессора SCSS
	Навыки отладки разметки в браузерах Chrome/Safari
	Знание принципов работы HTTP, отличий между различными методами, принципов работы Cookies
	Практическое владение лучшими практиками структурирования кода: SOLID/DRY/KISS и т.д.
	Базовые знания git, понимание git flow

	# Create a 1GiB Swap File
	# Lightsail ubuntu image doesn't boot with swap.
	# If you run out of RAM, the system hangs OOM.
	sudo fallocate -l 1G /swapfile
	sudo chmod 600 /swapfile
	sudo mkswap /swapfile
	sudo swapon /swapfile
	sudo cp /etc/fstab /etc/fstab.bak
	echo '/swapfile none swap sw 0 0' \| sudo tee -a /etc/fstab

	FROM ruby:2.5.5-alpine3.9 as builder

	WORKDIR /app

	ARG RAILS_MASTER_KEY

	ENV BUNDLE_SILENCE_ROOT_WARNING=1 \
	BUNDLE_IGNORE_MESSAGES=1 \
	BUNDLE_GITHUB_HTTPS=1 \
	BUNDLE_FROZEN=1 \

	# lib/tasks/db_fixtures_export.rake
	namespace 'db:fixtures' do
	desc "generate fixtures from the current database"

	task :export => :environment do
	Rails.application.eager_load!
	models = defined?(ApplicationRecord) ? ApplicationRecord.descendants : ActiveRecord::Base.descendants
	models.each do \|model\|
	puts "exporting: #{model}"