How this library works — and how I add to it

Everything I publish lives on one of ten shelves:

Shelf	What goes there
essays & deep dives	long-form technical writing — OSINT, war stories, engineering
CS curriculum	my complete computer science notes, field by field
ML / AI notes	working notes on models, agents, evals, and training
book summaries	books I finished, with what I underlined
paper notes	research papers, annotated, with the best per field
open ideas	ideas I’m giving away — with research on why they would or wouldn’t work
reflections	short escripts about life, categorized
quotes + commentary	other people’s words, my margin notes
spirituality	writing on openings, meaning, and finding god
resource maps	curated routes into a field — courses, papers, videos, in order

The mechanics

Each entry is a single markdown file:

---
title: "Agent memory patterns that actually work"
description: "One-line summary shown on the shelf."
type: ml          # one of the ten shelves
topics: [ai, agents]
date: 2026-07-02
readingTime: 9
---

The shelf sidebar, the badges, the search index, the homepage rail, and the ⌘K palette all generate from these files at build time. Publishing = git push.

If you’re reading this and want the same setup for your own site, the rebuild essay covers the architecture.