After seeing the excellent Cosmic Trigger play on Friday and thinking I might use Robert Anton Wilson’s approaches as a framework for making art about machine learning, I thought it apposite to push RAW’s words through a neural network and see what came out the other side. I used the torch-rnn Docker image and trained it on the Illuminatus! and Schrödinger’s Cat trilogies along with some other books I could extract plain text from. I then recorded my Mac’s speech tool reading it out over eight hours.

In other words, exactly the same process I used for the Goodbye Wittgenstein book and audio. Now there’s a pair, which is nice. If I continue and it becomes a series maybe I’ll think about doing something more with it.

R(AW)NN exists in a number of formats: