09 — Resources Canon

“The problem isn’t a shortage of resources. It’s that 80% of them teach the same 20% — and the wrong 20% at that.”

The ML internet is full of courses, books, and blog posts written to make you feel productive while you’re still on page one. This folder is the antidote: a curated, community-verified set of resources mapped precisely to the 13-month roadmap. Every item here earned its place. Nothing is here because it’s popular. Everything is here because practitioners — working engineers, research scientists, and people who have shipped real systems — still recommend it in 2025-2026.

The Philosophy: 20% → 80%

The Pareto principle applies brutally to ML education. The resources below represent a deliberate hierarchy:

  1. Understand the math (books + foundational courses)

  2. Build intuition through implementation (code-first courses + papers)

  3. Read primary sources (papers — not blog summaries of papers)

  4. Use the right tools (libraries, not frameworks you’ll migrate away from)

  5. Validate against real data (benchmark datasets)

Most learners do step 1, skip step 2, never do step 3, and use whatever tool they found first for steps 4-5. Don’t be most learners.

Resource Type Map

Folder Contents

File

What It Contains

Priority

01_books_canon.md

10-12 books that form the foundation of ML knowledge

Start here

02_courses_canon.md

12+ courses ranked honestly by community signal

Month 1

03_papers_canon.md

25-30 seminal papers organized by category

Ongoing

04_tools_and_libraries_canon.md

Python ecosystem map with versions and pitfalls

Day 1

05_datasets_canon.md

15-20 benchmark datasets and when to work with them

Month 2+

What Most People Get Wrong

They treat resource selection as a one-time decision made in month one. It isn’t. The right book in month 1 is wrong in month 7. The papers section is organized by roadmap phase for exactly this reason. Revisit this folder at the start of each new phase and check whether your current resource stack still matches where you are.

One more thing: when a resource contradicts another, check the date and implementation. A 2016 textbook and a 2024 paper describe the same field, but not the same reality.

Return to 13-Month Roadmap Root · Resources folder — all files listed above