AI image generation is here in a big way. A newly released open source image synthesis model called Stable Diffusion allows anyone with a PC and a decent GPU to conjure up almost any visual reality they can imagine. It can imitate virtually any visual style, and if you feed it a descriptive phrase, the results appear on your screen like magic.
Some artists are delighted by the prospect, others aren’t happy about it, and society at large still seems largely unaware of the rapidly evolving tech revolution taking place through communities on Twitter, Discord, and Github. Image synthesis arguably brings implications as big as the invention of the camera—or perhaps the creation of visual art itself. Even our sense of history might be at stake, depending on how things shake out. Either way, Stable Diffusion is leading a new wave of deep learning creative tools that are poised to revolutionize the creation of visual media.
Yes, Dall•E 2, Midjourney, stable diffusion and google Imagen are all deep neural nets (DNNs) that allow text → image generation.
These systems are pretty straight-forward extensions of years of work in artificial vision. Basically these systems have been trained on huge labelled data sets of human-created images and photos. With some advances in text understanding linked to the image network, and a nice dose of random starting conditions, an almost infinite set of novel creative images can be generated.
The argument that these systems are not truly creative belies the point that human artists also depend on their cultural upbringing and exposure to millions of images as they grow up…
As impressive as it is from a technical standpoint, it truly is – am I the only one here thinking it’s a stupid idea to build machines smarter, faster, stronger than humans?
Some (few) generations down the road this stuff will make nuclear fission look like a pretty safe and easy to control technology. I’m understating here.
We are still a long way away for general artificial intelligence (i.e. an intelligent agent who is as flexible and adaptable across many domains as a human, or even most animals). However this is an area where the ethics and philosophical implications are being actively studied and carefully developed.
Nick Bostrum is an interesting (if over-foreboding IMO) example:
For a critique of general AI and these deep networks, see the works of Gary Marcus, e.g.:
But in the shorter term, there will be a practical impact on many middle class jobs (including the creative arts as these AI demonstrate, or management, or cancer detection in scans etc.), where such domain-specific AIs can work “better” than humans can, and will slowly erode the workspace available…
