This project investigates big data (noise), and its counterpart artificial intelligence (signal), from a post human-centered composition and remix perspective. Experimental music is deeply rooted in technology and tools at a given time, and through its noises reflects an era in all its imaginary and techno-scientific shifts. This project embraces a history of Noise Music since the Futurists’ Art of Noises in 1913 until today, while proposing a non-human perspective on this heritage using Neural Networks. Hundreds of tracks, sliced into hundreds of samples, have been ordered by sonic proximity on a map that can now be navigated as a meta-remix, beyond tools, styles, countries and time. Here this history of Noise can be reshaped and contorted through a purely statistical perspective, somehow specific to the big-data era, generating infinite unheard connections.