Recently, I have been really impressed by the potential of diffusion models like Dall-E 2, Midjourney, and Stable Diffusion. These powerful AI tools have the ability to generate unique and highly stylized artwork, opening up a whole new realm of creative possibilities.
I am planning on integrating these diffusion models into my own creative workflow, as well as using them for prototyping and experimentation. While they provide an incredible source of inspiration and input, there are also many ethical considerations to take into account. For example, there are issues of copyright and fair payment for artists, as well as concerns about how AI-generated art might impact the ability of artists to sell and develop their own styles.
Ultimately, it will be important to find ways to balance the benefits of these technologies with the need to support and protect the creative community. By carefully considering these ethical aspects, we can ensure that these diffusion models are used in a way that helps artists to create new and innovative works of art, rather than simply reproducing the work of others.
AI Art
Laser Cutting
Using AI to generate laser cut patterns would be an application that has received little attention.
In Midjourney, you can see how great the application is for generating jewelry and other artwork.
Training a suitable network would be one of many projects I would like to work on.
As a starting point I have tried a model from PiyarSquare (https://huggingface.co/PiyarSquare/stable_diffusion_silz) among others, but I am not always completely satisfied with the results.
Woodworking for Festival Installations
Another application I'm very interested in is designing festival installations.
Here, too, you can see how well generative AI can be used to achieve impressive results, even if they are not always realizable on a 1:1 basis.
For this, too, it would be exciting to train networks that realize already finished patterns, which are then cut out by a CNC machine and assembled manually.
Generative Avatare
Even if there is no direct use for it, I find the generation of people who do not exist a peculiarly exciting phenomenon.
With "This Person does not exist" (https://this-person-does-not-exist.com/de), there is already the possibility of very quickly creating images of a wide variety of people.
However, specially trained networks such as Realistic Vision create even higher quality results, as can be seen in the gallery.