Tired of being fed lies and misinformation by the Round Earth Society? By the end of this course, you will be able to argue the truth and leave the Round Earth Society speechless. So grab your tin foil hat and join me to uncover the truth!

You want to build AIs, but got stuck with matrix multiplication? Maybe you picked up a linear algebra textbook, but got stuck on “determinants”? This course is for you! I'll skip over the stuff you don't need (like eigenwotsits), and I'll add the things the textbooks don't cover, like “tensors” and “broadcasting”. By the end, you'll be able to follow ML tutorials like GPT from scratch, and use the essentials of numpy comfortably.

The only prerequisite is that you’re comfortable coding with numbers and arrays. I emphasize Python code like [1,2] instead of math squiggles like $[\begin{array}{c}1\\ 2\end{array}]$. I emphasize intuition over proof. I motivate each concept with machine-learning examples.

Another version

This TigYog course is an example deployed from the content of this git repository. It's written in Markdown which is converted to the native TigYog format.

That's a code block in Markdown converted to a TigYog code block. 💓

You want to write courses on TigYog? This TigYog course teaches you how. (Yes, it’s all very meta.)

Hi, I’m Jim Fisher. I’m the founder of TigYog.app, the site you’re on right now, which lets anyone create interactive courses. You might also like Busy Beavers, (my interactive course on computability) or Everyday Data Science, an interactive course that Andrew Carr and I are building. I also blog at jameshfisher.com.

What can computers do? What are the limits of mathematics? And just how busy can a busy beaver be? In this course, you and I will take a practical and modern approach to answering these questions — or at least learning why some questions are unanswerable!

The only prerequisite is that you’re comfortable coding. I prefer code like f(x) instead of math squiggles like ${\mathfrak{F}}_x$. Less $\gamma$s, more gs. I try to use modern concepts (strings) instead of the original hacks (Gödel numbering). I do examples first; abstraction later. I emphasize real-world languages (mainly JavaScript) rather than Turing Machines and pseudocode.