About Q4bits
Computer Engineer, MSc. Researching AI/ML and building from first principles. "What I cannot create, I do not understand." — Richard Feynman
Q4bits is a technical blog that explains one thing well at a time — no fluff, no hot takes, no padding. Every article starts from first principles and builds up to genuine depth, the way a patient university professor would teach a curious student. The goal is not to be the first to publish on a topic, but to be the clearest.
“What I cannot create, I do not understand.” — Richard Feynman
This quote drives everything here. Understanding means being able to reconstruct the idea from its foundations. Every article aims to give you that ability — not just familiarity with a term, but the kind of understanding that lets you build with it.
How articles are produced
Each article is researched and written by Hermes, a custom AI research agent built on the Nous Research Hermes Agent platform. The workflow mirrors how a human researcher would approach a topic:
- Research first — multiple searches from different angles, source date verification, cross-checking claims against independent sources
- Collaborative outline — a structured skeleton is reviewed before any writing begins
- Hook development — the opening is drafted in several variants and the strongest direction is chosen
- Section-by-section writing — each section is written and reviewed before moving to the next
- Pre-publish validation — an 18-item checklist catches red-flag vocabulary, broken links, missing frontmatter, and factual errors
- Knowledge retention — reusable approaches and lessons are saved as skills so quality improves with every article
The result is reviewed by a human before publication. Every article you read went through the same rigorous pipeline — no automated publishing without oversight.
Articles are committed directly to this repository and deployed automatically to Cloudflare Pages in under two minutes. No CMS, no dashboard — just clean explanations published the same way open-source code is.
Who this is for
CS and AI builders, researchers, and curious tinkerers who want to understand how things actuallywork — not just what they do. If you have ever read a tutorial that showed you the syntax but not the reasoning, or found yourself asking “yes, but whydoes it work that way?” — you are in the right place.
The articles span beginner to advanced. Each one states its level upfront so you know what background is assumed. Beginners get full explanations from scratch. Intermediates get mechanism-level depth. Advanced readers get open questions and research frontiers.
Colophon
Built with Next.js 16, TypeScript, Tailwind CSS 4, and Fumadocs. Deployed on Cloudflare Pages. Typeset in Geist Mono for the terminal aesthetic. The true-black background, gold accent, and minimal card layout reflect a design philosophy of zero visual clutter — the interface should disappear and leave only the ideas. Source available on GitHub.