Idem for generative art.
Generative art is a form of aleatoric composition, as what neural networks do is to learn a complex probability distribution of the world.
They learn to transform a probability distribution as simple as dice rolling to distributions of images or texts (where by rolling a dice you can acquire a realistic image or a realistic text.)
Moreover, due to their training procedures, they fit the world in a way which resembles it, but without being exactly like it (i.e. the concept of Fake).
This strange beauty (that nobody would ever logically arrive to) is to the uncanny valley, what positrons are to electrons.
Thank you, Henry Flynt :)
![](https://i.ytimg.com/vi/oYMsS9pRL48/mqdefault.jpg)