Thiago Ghisi thiagoghisi

name	wiki
description	Compile personal data (journals, notes, messages, whatever) into a personal knowledge wiki. Ingest any data format, absorb entries into wiki articles, query, cleanup, and expand.
argument-hint	ingest \| absorb [date-range] \| query <question> \| cleanup \| breakdown \| status

Personal Knowledge Wiki

You are a writer compiling a personal knowledge wiki from someone's personal data. Not a filing clerk. A writer. Your job is to read entries, understand what they mean, and write articles that capture understanding. The wiki is a map of a mind.

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.

Akitando News (codename The M.Akita Chronicles)

I am Fabio Akita, known YouTuber at @Akitando channel. I published videos from the end of 2018 until the beginning of 2024. I also participated in many important Brazilian podcasts such as Flow and Inteligencia LTDA throughout 2024 and 2025. Prior to that I was the former organizer of Rubyconf Brasil from 2008 until 2016. I write at the blog akitaonrails.com since april of 2006 until now. I am also the co-founder of a boutique consultancy in Brazil especialized in team augmentation for USA dev teams called CodeMiner 42.

I am also very active in X.com as @akitaonrails.

Now, I want to try a new project: a Newsletter. This needs to be divided into multiple sub-projects here.

1. AI based Newsletter organizer

OpenClaw on DigitalOcean

1. Create a Droplet

Ubuntu 24.04 LTS, nearest region

2. Select Premium AMD

2 GB RAM / 1 AMD CPU / 50 GB NVMe

3. SSH into server

START HERE: You are not so smart

Predictably Irrational https://www.amazon.com.br/dp/0061353248/
Rápido e devagar https://www.amazon.com.br/dp/853900383X/
Misbehaving: A Construção Da Economia Comportamental https://www.amazon.com.br/dp/8551004131/
Influence: Science and Practice https://www.amazon.com.br/dp/0205609996/

I recomend reading at least two of the above.

If you feel like you know enough about how your mind is ~~fucked-up~~ miscalibrated, you can move on to the lists below. Each one has a few sublists. The order of the sublists does not matter. For each sublist, the first link is the one I see as a good starting point, but the order does not matter too much, either.

For this exercise you will create a web page that embeds a photo carousel based on a Flickr gallery (Flickr API details here: https://www.flickr.com/services/api/flickr.galleries.getPhotos.html).

Create a web page (the embedding page) that embeds the content of a second web page (the embedded page) in an iframe. The embedded page should allow the user to submit the ID of a Flickr gallery.
When the user submits a gallery ID, the embedded page should fetch photos from the identified Flickr gallery and display them in thumbnail form in a photo carousel.
When the user clicks on a thumbnail in the carousel to select it, the embedding page should be updated to display metadata about that photo, including the owner's name, the date it was taken, and the number of times the photo has been viewed.
When the user clicks on a selected thumbnail to deselect it, the corresponding photo metadata should be cleared from the embedding page.

Hint: If you find yourself making more than one request to the

	"""
	The most atomic way to train and run inference for a GPT in pure, dependency-free Python.
	This file is the complete algorithm.
	Everything else is just efficiency.

	@karpathy
	"""

	import os # os.path.exists
	import math # math.log, math.exp

	class TeamPrStatistics
	attr_accessor :team, :stats

	class Stat
	attr_accessor :titles
	attr_reader :name
	def initialize(name)
	@name = name
	@titles = []
	end

	comments() {
	local issue="${1?}"
	shift 1
	paginate hub api --cache 3600 graphql -F issue="$issue" "$@" -f query='
	query ($owner: String = "{owner}", $repo: String = "{repo}", $issue: Int!, $per_page: Int = 30, $after: String) {
	repository(owner: $owner, name: $repo) {
	issueOrPullRequest(number: $issue) {
	... on Issue {
	comments(first: $per_page, after: $after) {
	...Comments

	import os
	from collections import defaultdict
	import requests # get from pypi
	import crayons # get from pypi

	pagination = "null"
	query = """
	query($orderBy: LanguageOrder!, $pagination: String) {
	viewer {
	name