AI Art Weekly #13
Hello there my fellow dreamers, welcome to issue #13 of AI Art Weekly 👋
I have some exciting news to share! Besides the usual AI goodies, we found our first sponsor 🥳. Looking Glass was kind enough to support this and the next issue of AI Art Weekly. They produce 3D holographic canvases that make it possible to display 3D images and animations without the need for a VR headset. As of now, most AI art is only two dimensional, but depth maps can be used to convert images into 3D scenes. This makes the Looking Glass Portrait one of the coolest ways to display your art. If you want to support the newsletter and get yourself or a loved one a cool present, the link below will get you $40 off plus some additional freebies.
With all that being said, thank you so much for reading AI Art Weekly 🙏
Let’s start with this weeks highlights:
- Imagen 3D , Text-To-Motion , Video outpainting and Stable Diffusion spectograms that can get turned into music
- Interview with painter and AI artist Ilya Shkipin
- Promising Dreambooth and Automatic1111 WebUI alternatives.
Sponsored
– Turn your AI generations into incredible 3D holograms with the Looking Glass Portrait. For the month of December, we are giving away 100 free conversion credits and a carrying case with every purchase + free shipping (for the US & Japan). Get an additional $40 off by clicking the link above or below (discount applied at checkout).
Cover Challenge 🎨
This weeks “ snowfall ” challenge got a wooping 101 submissions from 58 artists 🤯. The community decided on the final winner –congratulations to @GuyP for winning the final round with his giant snowman ☃️. And as always a big thank you to everyone who contributed!
The theme for the next challenge should be obvious, it’s “ christmas ” 🎅🧑🎄. Think Santa Claus, reindeers, Christmas elves, Krampus etc. Prize is another $50. Rulebook can be found here and images can be submitted here.
I’m looking forward to all of your beautiful artworks 🙏
If you want to support the newsletter, this weeks cover is available for collection on objkt as a limited edition of 10 for 2.75ꜩ a piece. Thank you for your support 🙏😘
Reflection: News & Gems
Imagine 3D
Luma AI launched a new Text-To-3D product called Imagine this week. Generation is only available to a few lucky individuals at the time of writing. But there is a waitlist you can join. I definitely signed up and can’t wait to give it a try. In the meantime you might want to take a look at what the community has generated – the results look amazing.
Text-to-Motion
Speaking of 3D – there is a LOT of research not only going into generating 3D objects, but also into animating them. PhysDiff, MoFusion and MLD are three different proposals for Human Motion Diffusion which allow the generation of unique animation patterns from pure text prompts. Just imagine your 3D generated avatar sitting in your Looking Glass Portrait and a Motion Diffusion model generating animations on the fly based on your voice, gestures, the weather outside or any other input you can come up with.
MAGVIT: Masked Generative Video Transformer
Image outpainting and inpainting are mind blowing features in itself, but what if you could inpaint or outpaint videos? Well, MAGVIT makes it possible. But that’s not where the fun stops. The model is also capable of turning a single image into an animation, create future frame predictions and compress videos by a rate of 600x compared to the original footage (but don’t tell Hooli).
RIFFUSION
RIFFUSION (riff + diffusion) is a Stable Diffusion v1.5 model fine-tuned on images of spectrograms paired with text. The model produces spectograms based on text prompts which then can get turned into audio clips. Pretty smart. I had some trouble using their web app, but luckily enough there is also a HuggingFace space.
An interesting short documentary by @kylevorbach faking his life over a month by Dreambooth’ing his face and fabricating a synthetic life on Social Media. The mad-lad.
@DrJimFan dives into the @Google Robotics RT-1 announcement. The new robot is equipped with a 7-DOF arm, a 2-fingered gripper, and a wheeled base. It can retrieve lost sunglasses, organize condiments, restock and prepare snacks, and close drawers left open in Google’s kitchens.
@ammaar combined ChatGPT and MidJourney to create an illustrated children’s book trying to introduce the concept of AI to kids. This is certainly not the first of its kind, but it’s always fun to see how people use AI to create physical products.
Imagination: Interview & Inspiration
This week we talk to artist Ilya Shkipin. Ilya is a great example of a traditional artist who embraces the possibilities of AI generative art in his own way and combines the new with the old to create something unique and new. Drawing inspiration from the likes of Frank Auerbach and Francis Bacon, Ilya’s work is gritty, weird and disturbing. His most recent project, “Eyes Wide Shut”, is a collection of girl portraits created with a combination of AI and painting that explores the line between sensuality and the grotesque. Let’s jump in.
[AI Art Weekly] Hey Ilya, what’s your background and how did you get into AI art?
As a child, I knew I wanted to be an artist. My grandfather and father were both artists, so the desire to create came naturally to me. I have always been fascinated by art history, and some of my influences include painters who create gritty, weird, and disturbing imagery such as Frank Auerbach, Richard Diebenkorn, Francis Bacon.
About a year and a half ago, I was introduced to neural networks by Nikita Panin and slowly began to embrace this new medium as it continued to develop. This is when my career took a different path: neural net art seemed like the most natural fit for me. I didn’t care about mediums or “purity” anymore; the most important thing was achieving my own aesthetic and voice.
The originality of AI art lies in the relationship between the digital and the traditional, a form that is still not fully resolved. AI poses questions about what constitutes true creativity, technique, and the role of artists in society. It also raises the issue of standing out in a world where anyone can create anything.
[AI Art Weekly] Do you have a specific project you’re currently working on? What is it?
I just released a drop on Foundation this called “Eyes Wide Shut”. It is a collection of girl portraits created through a combination of AI and painting, ranging in gradation from sensual to disturbing, from pain to pleasure. The series will explore the line between sensuality and the kind of grotesque shock that is usually reserved for train wrecks or those disastrous moments when you realize that you forgot your girlfriend’s birthday.
[AI Art Weekly] What does your workflow look like?
AI artists guard their workflow carefully, as revealing secret ingredients will ruin the magic of it. My workflow varies heavily, but my main objective is to leave a trace of my own touch in almost every piece I make. Even if AI can produce any technique and brushwork that you desire, I am still a painter at heart, and the painting process itself is enjoyable for me. I don’t know if it makes my art more valuable, but it makes it more “me” rather than the product of AI.
I can’t share prompts for the reason stated above, but I would say that lately a prompt is just the initial stage, after which the work passes through many other stages such as image-to-image processing.
[AI Art Weekly] How do you imagine AI (art) will be impacting society in the near future?
I think that AI art is just the beginning, and people who are concerned about the state of art are missing the big picture. AI will revolutionize our society on the same scale as the industrial revolution did before. Think about how it will affect the running of governments, laws, progress in medicine, technology, books, and film making. As far as art goes, one interesting aspect is how AI democratizes the image-making process. There are both positive and negative aspects to this. On the one hand, artists who have trained for 20 years to become professional painters and illustrators may not be happy about it, as their jobs can now be done by interns. On the other hand, artists who have a knowledge or understanding of aesthetics can utilize AI much better than those who don’t.
[AI Art Weekly] Who is your favourite artist?
I don’t have a single favourite artist, but I enjoy looking at a lot of art. I love going to museums and attending art exhibitions. Currently, I am planning to visit a new exhibition at the Dolby Chadwick gallery in San Francisco that showcases some of my favourite artists, such as Emilio Villalba.
[AI Art Weekly] Anything else you would like to share?
Lately, I’ve been focusing more and more on music making with my band, The Laytcomers. If you like noisy and freaky post-punk music, they might be worth checking out!
Each week we share a style that produces some cool results when used in your prompts. This weeks style is based upon a finetuned Stable Diffusion model trained to produce analog film style images (join me in figuring out keywords that produce similar results in MidJourney). Trained with Dreambooth by @wavymulder.
Creation: Tools & Tutorials
These are some of the most interesting resources I’ve come across this week.
I tried the WebUI several weeks ago, and I think it was even before I started writing this newsletter. It was the first one to run on an Apple M1 Macbook, and I must say it has come a long way. If you’re looking for a more user-friendly and stable alternative to the Automatic1111 WebUI, you might want to give InvokeAI a try.
There is a new Stable Diffusion animation notebook in town created by @RiversHaveWings. It’s not as elaborate as Deforum when it comes to the settings, but the output looks cool.
Seek.art MEGA is a general use “anything” model trained on nearly 10k high-quality digital artworks that significantly improves on Stable Diffusion v1.5 across dozens of styles.
LoRA is a fine-tuning method in the domain of LLM and this repository promises a faster and more lightweight fine tuning alternative to Dreambooth. There are both a HuggingFace Space and a Google Colab notebook for you to try out.
If you don’t want to think about all the technical aspects of fine tuning and don’t mind shelling out a few bucks to create your checkpoints, this @Gradio HuggingFace space by @abhi1thakur is probably the easiest way to do so (I feel like I say this every week lately).
And that my fellow dreamers, concludes yet another AI Art weekly issue. Please consider supporting this newsletter by:
- Sharing it 🙏❤️
- Following me on Twitter: @dreamingtulpa
- Leaving a Review on Product Hunt
- Using the Looking Glass link at the top or one of our affiliate links at https://aiartweekly.com/support
Reply to this email if you have any feedback or ideas for this newsletter.
Thanks for reading and talk to you next week!
– dreamingtulpa