I had an idea of the world as a cubist painting. Using a convolutional neural network, I combined compelling satellite imagery with the bold work of abstract expressionist painters. The results you see collected here show approximately what I'd hoped: that humans shape our planet in distinct, clear ways. These range from wispy traces of wildfire smoke to rigid geometric landscapes, and are still evident when perspective is doubly-flattened: first by the aerial view, and second by a busybody algorithm. These images show a wrenching aesthetic dilemma: sometimes our hammerblow of deconstruction on the land is beautiful to behold, or at least sublime.
I am not an artist; I make these pieces with some of my own sensibility expressed as a series of command line arguments — switches and weights. I’m amalgamating the true work that undergirds it: the paintcraft of Francis Picabia, the code of Justin Johnson, the scientists who built the neural network, and the engineers who flung satellites into orbit. But in a way, this is meta-cubism: aided by machines, I am pulling apart composed materials and putting them back together, slightly rearranged. I would hope that this would raise at least a smirk from Picabia before he would throw a drink in my face.
- Bill Morris