turboBasic

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.

turboBasic

#!/usr/bin/env python3
"""List GitHub organization teams and print matching team names.

Usage:
  export GITHUB_TOKEN="<token-with-read:org>"
  python3 list_github_org.py --org your-org-name find-teams-with-members --user-names user1,user2

Notes:
- For private org/team visibility, the token needs at least `read:org` scope.
- Fine-grained tokens must include organization members/teams read permissions.

turboBasic

Code snippets for querying web pages in a browser's console

Select contents of all elements of <span>1234567</span> format

[...document.querySelectorAll('span')]
  .filter(el => /^\d+$/.test(el.textContent.trim()))
  .map(el => el.textContent.trim())

turboBasic

About

This is an awesome list or resources on sashiko, kantha, visible mending and other related techniques.

Overview

• basic concepts of Japanese aesthetics in Ukrainian

Techniques related to sashiko

turboBasic

#!/usr/bin/env zsh

## Save initial state

exec {saved_stdin}<&0   # save initial stdin
exec {saved_stdout}>&1  # save initial stdout


## Redirect initial stdin and stdout to terminal

turboBasic

# Usage: compare-locales LC_TIME POSIX de_DE.UTF-8 'date'
function compare-locales() {
    local -r category="$1"
    local -r locale1="$2"
    local -r locale2="$3"
    local -r command="$4"

    #echo "=== Comparing '$command' in $locale1 and $locale2 ==="
    printf "Output in %-15s %s\n" "$locale1:" "$(
        export "$category=$locale1"

turboBasic

# Calculate the size of a git commit in bytes.
function git-commit-size() {
  # This script calculates the size of a git commit in bytes.
  # Usage: git-commit-size.sh <commit hash>
  #
  # Example:
  #   git-commit-size.sh 1234567
  #
  # Output:
  #   The total size of the commit in bytes, formatted with thousands separators.

turboBasic

# List running EC2 instances filtered by `Name` tag
#
# params:
#   $1 - optional string. Filters output by instances with `Name` tag containing the string
function aws-ec2() {
    aws ec2 describe-instances \
        --filter "Name=instance-state-name,Values=running" \
        --filter "Name=tag:Name,Values=*$1*" \
        --query "Reservations[].Instances[].{
            id: InstanceId,

turboBasic

#!/bin/sh

item=$1
field=$2

bw get item "$1" \
|   jq --raw-output '.fields[] | select(.name == "'"$field"'") | .value'

turboBasic

##  Global vars
export LESS='--buffers=128 --HILITE-UNREAD --ignore-case --LONG-PROMPT --max-back-scroll=15 --no-init --quiet --quit-at-eof --quit-if-one-screen --RAW-CONTROL-CHARS --status-column --tabs=4 --window=-4'
export FZF_DEFAULT_COMMAND='rg --files --no-ignore --hidden --follow --glob "!.git/*"'
export FZF_DEFAULT_OPTS='--height 50% --ansi --info=inline'
export EDITOR='/usr/local/bin/nvim'


##  zinit
source ~/.zinit/bin/zinit.zsh