I experimented with Ganbreeder, an app that uses a machine learning technique to generate images. It uses a pair of networks, one that produces data and the other that classifies it. The user selects a number of images of objects/animals/people and Ganbreeder merges them all together to create new ones. The images created are strange, and do not make much sense which I think it my attraction to them. The idea of putting data into a network and it producing something feels like an interesting direction, and the occasionally warped outcome is significant when considering the flaws of the digital age. Perhaps I can experiment with this idea later, but using sound instead of visuals.