What You May Have Missed #12
On AI art, trad artists, and the clash between the two
Before we start, here’s an idea for you: Christmas is around the corner. I’m quite bad a coming up with ideas for gifts, and I know I’m not alone! That’s why I want to remind you that you can gift a TAB subscription to someone who may like it as much as you!
Generative AI grows and grows
This week's WYMHM is focused on AI art. The events that are taking place right now in the space will play a vital role in the future of Generative AI as a whole.
2022 has been the year of “text-to-everything,” as Daniel Bashir from “Last Week in AI” deemed it. Generative models based on text prompts have been an ever-increasing source of interest for researchers, investors, and users alike.
The AI art community—built around text-to-image models like DALL-E, Midjourney, and Stable Diffusion—keeps finding new intriguing concepts at the research level. One recent example is Riffusion, a novel approach that leverages Stable Diffusion to make music in real-time using text prompts (demo on Gradio).
Generative AI has also reached the consumer level. Ready-to-use models emerge every month: Lensa is at the top of the Apple store—although it’s not free of problems—and Sharif Shameem just released Lexica Aperture.
People come up with more use cases in search of market fit. And the open-source industry seems to be cutting artists some slack:
However, not everything is good news for AI art.
The conflict between traditional artists (that is, anyone who's a visual artist of some kind but doesn't use AI in their work) and advocates of AI's applications on visual art is growing beyond repair.
This week it’s reached the tipping point.