Wintergatan - Marble Machine (music instrument using 2000 marbles)

A few years ago (I think it popped up on HN and all over Reddit) Martin made the Marble Machine[0], which was Rube Goldberg-style machine inspired by mechanical clockwork music devices like clock-towers, music boxes, and street organs. A little less than a year, he announced that he would be making a newer version with much more design intent, better precision, and the ability to tear it down to take it on a world tour. I personally have been following the project through the weekly videos he puts up about the various design challenges he's been facing, and how his team has been growing from himself to around 10 people. The growth, scale, and skill seen from Martin has been amazing. The project has gotten to the point that he met with the developers of Fusion 360 because the model had so many parts it almost broke the program[1]. I figured that this video would be interesting to HN because it discusses the project management of such a massive endeavor, and I think the project as a whole might appeal to the hacker/maker side of HN. [0]: [this video] [1]: [link]

> It's like we're stuck in a codebase that was started centuries ago, and everyone is too afraid to commit a bugfix, let alone a major refactor. This describes everything about the classical music world. They're still playing the same old pieces by the same old composers, on the same old instruments. There's nothing wrong with doing old things, but it's somewhat baffling. As a home guitar player and producer, I can trivially get pedals and amps that practically let me make any sound I want with my guitar, as well as equipment for alternative techniques like the E-bow. If you jump over to the MIDI/DAW world, there's pretty much no limit to what you can do. Some stuff is already coming out that lets you translate humming into MIDI in real-time. Yet people trip over themselves to tell you why there shouldn't be pianos of different sizes, let alone (god forbid) pianos with different keyboard layouts. I can only imagine the reaction of the average classical performer if you were to suggest making a major overhaul to the design or performance of an instrument itself, if not adding in new instruments altogether. As an engineer, I've been doubly baffled by the adherence to doing things traditionally just because that's how it's always been done. The attitude I've perceived from people in the classical world that I know is that they pretty much don't believe there could be a better notation, technique, etc. Meanwhile in the software world, it feels like we have a new coding language e