This is amazingly cool and stimulating to think about. It's also a compelling reason to take another look at Mathematica. The code is clearly dense and hard to understand, but I wouldn't even know how to go about performing some of these computations in another language.
I really wish the licensing for Mathematica weren't as strict; it's a fantastic language that would be everywhere if it had a free "community version" or something. I used it very extensively in college and then took advantage of the $500 lifetime professional upgrade for new graduates. Now I do server-side development and data science at a startup, and I often work on problems that could solved efficiently with a few lines of Mathematica code. I'll use it for one-time internal scripts, but due to Wolfram's licensing you can't just bake it into your server code. Not to mention, it's nearly impossible to find engineers who have much experience with it since you can't exactly download it and hack on it like you can with the alternatives.
Don't get me wrong, it's a powerful tool and well worth it's price. But adoption is extremely low outside of academia, and I don't see that changing anytime soon unless Wolfram adopts more developer-friendly licensing. Which is a pity because it's definitely the "right tool for the job" in a lot of cases.
