Recommended AI channels, courses, newsletters, and reports
A living Practicum page. Checked: 2026-08-18. Resources appear, move, and disappear. Check a source before you rely on it. These are not ads or affiliate links, and inclusion does not mean the author has reviewed every item in full.
This is a starting list for following AI beyond the book. It uses the same rule as the tools map: one direction and one line on why it may help. Do not try to follow everything. Pick one or two sources that fit your work, then return to this page once a quarter.
YouTube channels and podcasts
| Channel or podcast | Why follow it | Link |
|---|---|---|
| Andrej Karpathy | A neural network researcher explains how the systems work. The focus is mechanics, not daily news. | youtube.com/@AndrejKarpathy |
| Lex Fridman Podcast | Long interviews with researchers and AI company founders. Useful when you want context and competing views. | lexfridman.com/podcast |
| Two Minute Papers | Short, plain-language reviews of recent AI research papers. | youtube.com/@TwoMinutePapers |
Courses
| Course | Why take it | Link |
|---|---|---|
| DeepLearning.AI by Andrew Ng | Introductory and applied courses on AI and machine learning, including short practical courses. | deeplearning.ai |
| Practical Deep Learning for Coders by fast.ai | A hands-on course for readers who want to understand the systems, not only use finished tools. | course.fast.ai |
Newsletters and blogs
| Newsletter or blog | Why follow it | Link |
|---|---|---|
| One Useful Thing by Ethan Mollick | A practical, anti-hype view of AI at work from a Wharton professor. | oneusefulthing.org |
| Anthropic Blog | Anthropic's own product and research announcements. Read it for the company's claim, then check other evidence. | anthropic.com/news |
| OpenAI Blog | OpenAI's own product and research announcements. Apply the same evidence rule. | openai.com/news |
Primary places for further research
- Reports: World Economic Forum https://www.weforum.org/publications/, McKinsey Global Institute https://www.mckinsey.com/mgi, Stanford AI Index https://hai.stanford.edu/ai-index, and Epoch AI https://epoch.ai/.
- Vendor announcements: OpenAI https://openai.com/news/, Anthropic https://www.anthropic.com/news, and Google https://blog.google/technology/ai/.
- Primary series you can check yourself, without a press release in between: Cloudflare Radar https://radar.cloudflare.com/ for machine traffic, ARC Prize https://arcprize.org/ for whether models handle genuinely unfamiliar problems, Eurostat https://ec.europa.eu/eurostat for labour data, and the European Commission's digital strategy pages https://digital-strategy.ec.europa.eu/ for what AI regulation currently requires.
This list is deliberately short. Add a source only when you can say why it helps and when you last checked it. For tools, use the tools map. For work grounded in your own documents, open the NotebookLM guide.
Version: 2026-08-18.