February 16, 2026
From vectors to vibes
Show HN: Maths, CS and AI Compendium
Open AI textbook drops: 'Maths' vs 'Math', JAX code, missing outlines, and chatbot study hacks
TLDR: A free, intuition-first AI textbook dropped on GitHub with live math and machine learning chapters plus JAX code. Comments erupted over “Maths” vs “Math,” called for outlines on unfinished sections like SIMD, and shared chatbot study hacks—proof this guide already has learners fired up.
An open, intuition-first "Maths, CS & AI Compendium" just hit GitHub. The creator says seven years of notes helped friends land roles at DeepMind, OpenAI, and Nvidia, and now it’s public. The crowd is split between applause and eyebrow raises, but everyone loves the hands-on promise: Henry drops, "Code walkthroughs and exercises are included, in JAX" (a coding tool for math-heavy stuff). Early chapters on vectors, matrices, calculus, stats, probability, and machine learning are live.
Then the real battle: the title. One commenter launched a transatlantic grammar war with "Math, not Maths. You wouldn’t called it Econs 101 would you?" Brits clutched their "Maths", Americans waved "Math" like a flag. Memes flew—"Econs 101" became the day’s in-joke—and the name debate briefly overshadowed the syllabus.
Practical devs want structure: "SIMD for example… :D" begged for outlines on unfinished parts (SIMD means doing one operation over many pieces at once). Another commenter shared a study hack: use Gemini’s quiz mode to drill topics, pairing bots with books. So yes, we’ve got AI learning AI, JAX code, and a community ready to contribute—if the roadmap shows them the way.
Key Points
- •An open, intuition-first textbook on mathematics, computing, and AI is published on GitHub by Henry Ndubuaku.
- •Six chapters are available now: Vectors, Matrices, Calculus, Statistics, Probability, and Machine Learning.
- •Additional chapters on NLP, vision, audio, multimodal learning, systems, GPU programming, inference, and more are planned.
- •The material was developed over seven years and helped peers prepare for interviews at DeepMind, OpenAI, and Nvidia in 2025.
- •The project invites contributions via GitHub stars/watches, issues for topic suggestions, and pull requests for corrections.