Judging by the timestamps on my solutions, this is the culmination of 2 hours of actual gameplay. The timeline suggests that I started at 11 hours before submission close, working for one hour, returning at 7 hours before submission close, and finally finishing at 6 hours before. This does not count the time I spent thinking about this throughout the later half of the week, where I kept thinking, "oh, this'll be so much fun to make when I get around to it in a few hours." Why did this procrastination happen in the later half of the week? Because I procrastinated opening up the puzzle through roughly the same train of thought. The only reason why this solution wasn't submitted with even less time on the clock is because I couldn't guarantee that I'd wake up to un-procrastinate. Really, I need to step up my procrastination game.

There's *tons* of room for improvement here. Delivering 2 metal products per loop is disgusting, and pretty quickly fixed. 3 metal products per full cycle is doable in the exact same number of cycles per loop, and would probably cut down on total cycles. That would mean editing like three different timing things, though, so I'll do that later. If you're reading this, I procrastinated too hard on that. The quicksilver trash pipeline is... just horrible. With a little reconfiguration and input suppression, it doesn't even have to trash the quicksilver. That would actually be impacted by the prior change, though, so I'll probably wait to do that first. 

My original plans called for a sorting mechanism that used triplex bonding to determine if the template atom was fire, then otherwise elemental, then metal. Using animismus glyph instead meant that the elemental configurations were all treated the same, and I didn't need the Berlo. This immediately made it a very enticing option. Still, I feel like there might be some cool triplex trick that's possible. Hopefully someone a little more dedicated than me has made something awesome with that.