Painting with Fire presents a vignette of transformation: the evolution of GANs from 2014 - 2023 seen through the eyes of fire. Using 14 vintage GAN models, Ionson articulates the emergence of synthetic image generation. Ionson explores fire as the quintessence of humanity's relationship to both innovation and destruction, embracing the unexpected abstractions of classic AI as a distinctive aesthetic statement.

GAN (generative adversarial network) technologies were developed by researchers and programmers but then utilized by artists. Though the technology was invented to make the images as realistic as possible, Brd subverts this to use it chaotically and abstractly; the opposite of what the AI researchers had in mind. This project celebrates the innovators and artists who have brought synthetic image generation to life with models such as the aiCAN, DCGAN, Stylegan 1, 2 & 3, bigGan, biGAN, pix2pix, AttnGan and more.
https://bardionson.com/painting-with-fire/
a Sovrn Curated release

Painting with Fire: A History in GANs by Bard Ionson

Epoch 33 the model drives into purple patterns the dark energy surrounding fire as it is often photographed in the dark. To see the light best one must surround it with darkness. Will Ai exist best in the darkest places?
Channeling the fire myths of the world, Prometheus, the Phoenix, the gods of the sun and flame. A story of technology brought to humans; to myths of destruction and rebirth. To be immortal; to be a god; it is always a paradox of fire and technology. The morality of the fire is always in question but we desire it and fear it. And yet it has changed us, we have learned so much since we first touched fire. We used it to destroy the wood, smelt the metal and build the chip. And now we teach and show the fire what we know; what we create and our reflection is there telling us our faults. The GAN is but a single neuron in a blazing star; a greater fire is coming. 
One of 200 unique pieces of art in Painting With Fire: A History in GANs. Exploring the history of generative adversarial networks GAN as represented by fire.  https://bardionson.com/painting-with-fire
DCGAN - Radford, Alec; Metz, Luke; Chintala, Soumith . "Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks"
Anna Ridler, The Fall of the House of Usher II|2017|https://annaridler.com/fall-of-the-house-of-usher-ii