The Activation Functions of a Neural Network - Machine Philosopher

I think I read about activation functions "squashing inputs into outputs" five or six times before I finally started getting the gist of them. This was no fault of the material I was reading, they were just this abstract concept that I just accepted and shoved my inputs into when I was building a neural network. However, I finally feel that I have a much clearer picture of why they are used and how well each one performs for certain tasks. I hope somehow I can squash this concept into your head by the end of this article so something useful can come out in the future! I'm sure you will catch on much faster than I did.