For his thesis, Mariano Salcedo built a modular, real-time, interactive music visualizer. Leaning into his undergraduate background studies at MIT, he used machine learning (specifically neural cellular automata) to let simple local rules grow into complex, ever-changing visuals that respond live to music.
Going forward, he is hoping to dig deeper into self-organization: basically, how complexity and order can emerge from many simple things interacting without anyone directing them from the top down.
Salcedo believes there’s a lot to explore in how that idea could shape new tools for music, art, and beyond.