I'm here to ask you one question, and one question only.

I actually prefer the term macOS to OSX.

Prerequisites (In Order Of Importance)

If your buffer has empty lines, they are filled with annoying tildes. Vim 8 and Neovim finally fix this by allowing you to conceal them with the following command:

I have (what I find) a handy mnemonic for the law of total variance. It's also sometimes called the conditional variance formula, or (the reason for this mnenomic) Eve's law. At least if I'm quoting Wikipedia right.

I forget where I've seen this, but I like using a squiggly arrow to represent a function that returns a probability.

(n,) vs. (n, 1)

Algebraic manipulation is always a fun way to make something profound seem like mere trickery.

1 or 2 # returns 1, not True
(1 or 1) == True # returns False

Running plt.show() causes all sorts of issues on macOS. To fix them, switch your matplotlib backend to something besides MacOSX. I used Qt5Agg.

From I Want to be a Mathematician, Paul Halmos's autobiography.

Understanding the fundamentals in chapter 5, especially over and underfitting, really pays off. This chapter is really easy to read if you can even half-remember those earlier sections.

I'm glad I read this early on, because it's kept me going ever since.

This is a topic near to my heart, in the same way that an old bullet stuck in your chest is.

Why do we care about true statements?

Why are we privileging truth over falsity? There's nothing inherently better about inconsistent models.

Started shallow, going deep.