I had a neural network hallucinate over the bible - the text is the input to generate the visuals, and the audio is a mix between text to speech and autoencoder-based processing of gregorian chants