AI Art Weekly #1
Welcome to the first Issue of AI Art Weekly. A newsletter by me (dreamingtulpa), to cover some of the latest happenings in the AI Art world.
Each issue will contain three main sections:
- Reflection : We take a look at what’s new in the AI art space. What are the latest updates and what new features got released this week.
- Imagination : AI art wouldn’t be much without imagination. In this section we share prompts for you to try out and highlight people from the AI art community.
- Coordination : After reflection and imagination comes coordination. In this section we share tools and tutorials to help you turn your dreams into art!
Without further ado, let’s get into it.
There were quite a few updates this week. Midjourney released a new feature which lets users create seamless repeating patterns by just adding the
--tile option to their prompts. Seem like a great way to create patterns for websites, home decor, gift wrappings and more. Websites like https://www.pycheung.com/checker and https://bardotbrush.com/repeat let you upload your creations and scale them according to your liking.
Stability.AI released a new in-painting / out-painting feature to their DreamStudio web service. In a similar fashion to Dalle2, it’s now possible to modify existing pictures by using a brush to erase parts of images you want regenerated. Outpainting now allows you to extend images beyond their borders and imagine what could be outside of the frame. Karen Cheng on Instagram created a cool behind the scenes video using outpainting to create a 3D instagram filter called “Museum of Future” that’s worth checking out.
And last but not least, Dalle2 announced that they support uploading faces again. Apparantly the team is now confident enough with their safety system to minimize the potential of harmful deep fakes. Great news!
Each issue of AI art weekly will feature a short interview with an artist from the AI art community. The first spot takes WeavingWithAI aka @GanWeaving on Twitter which I stumbled upon through his now famous Memento Mori tweet.
[AI Art Weekly] @GanWeaving, what is your favourite prompt when creating art?
One prompt that I really like is
petting zoo with alien creatures (example).
Another one is
boys/girls fixing their robotic toys (girls, boys, video animation).
[AI Art Weekly] Who is your favourite artist?
I don’t have one favourite artist. I won’t mention the very well known ones, instead I’ll list a couple of probably lesser known ones: Mattias Adolfsson, Enki Bilal and Shaun Tan.
[AI Art Weekly] What does your workflow look like?
Since I’m mostly doing StableDiffusion these days, I’m usually quickly iterating ideas on playgroundai.com, playing around with prompt variations, and mixing different artists to see if there’s a style coming out of this that I really like. And then I will use those prompts in the Deforum notebook on Google Colab and experiment a lot with its parameters. Either for still images or animations.
[AI Art Weekly] Anything else you would like to share?
Just play around with these wonderful new tools. Brush up your art history! Look up historical artists and art styles. Mash things together! Feed paradoxes to the machine and curate the outputs. And above all, don’t forget: Everything is a remix! Let no one tell you differently. Especially not that AI art is just pushing a button or copy-pasting. Only ignorant people will claim this. Educate them!
Each week we share a prompt that produces some cool results when used in combination with other prompts. This weeks featured prompt is
Yoji Shinkawa, a japanese artist best known as the lead character and mecha designer for the Metal Gear franchise.
AI art is still very much in its early stages. Although generative art has existed for a while now, improvements in user experience have not yet catched up. They will eventually, but until then, Google Colab and the command line tool for consumer GPUs is what we have to deal with. In this section of the newsletter I share some interesting tools and tutorials for you to play around with.
If you come across or developed a new tool and want it featured here, let me know on Twitter.
First is a Google Colab notebook for creating video animations. If you haven’t tried this yet, please do, it’s amazing. There are different modes for 2D, 3D, video input or interpolation with each providing different results depending on what you want to achieve.
Last week I created a video input experiment with the Deforum colab above. People asked me how I did it, so I put together a quick Twitter tutorial on how to replicate this style of content.
A worlds first one click installer for StableDiffusion for Macs that have a M1/M2 chip. I haven’t tried this yet, but will coming next week.
Inspired by @mattdesl, user @radamar created a Hugging Face space which lets you create color palettes using text prompts
AI Art Weekly isn’t just about image and video related resources. Dance Diffusion is the first in a suite of generative audio tools for producers and musicians to be released by Harmonai. It lets you generate, regenerate and interpolate audio samples. Might be cool to incorporate and combine with other tools like the Deforum notebook.
Looking for prompt references? Lexica is a search engine that lets you search through millions of images generated by StableDiffusion AI. It also offers an API for more tech savvy people.
And that my friends, concludes the first AI Art weekly newsletter. Please consider subscribing and sharing if you liked it and let me know on Twitter if you have any feedback. The more people get to see this, the longer I can keep this up and the more resources I can put into it.
Thanks for reading and see you next week!