ChAIos Media News - Inaugural edition!

Hello and welcome to the very first edition of the ChAIos Media newsletter. With the holidays around the corner, this edition is one of the small scale variety but we wanted to make sure we got the first one out of the door today to tell you a little bit more about where we intend to go with this newsletter.

This newsletter is operated by Ben Nash and Marco van Hylckama Vlieg. We are both seasoned designers and avid AI community members and we are extremely excited to offer you our thoughts on what’s going on in the AI space every week, starting today.

Let’s start with the first burning question everyone will undoubtedly have on their mind: “Why another AI newsletter?”

The answer to this question is simple. The generative AI scene is incredibly overwhelming. I’m sure we’ve all felt like it’s impossible to keep up with everything that’s going on in this space. That’s where this newsletter comes in. First of all our focus is creatives using AI. We deliberately do not focus on business / enterprise, engineering and other non-creative fields. There are plenty of other excellent newsletters that cover those. Secondly, even in the creative AI arena the amount of things that are happening on a weekly basis can still be quite overwhelming. This is where we hope to provide our core value. Our newsletter will inform you of the handful of things we feel you MUST know about every week and nothing more. Look at it as a curated ‘essence’ of what happened this week in generative AI for creatives. Because of this, don’t expect this newsletter to contain every single tidbit of AI news but rather a selection of ‘must know about’ items.

This week’s edition is really just the warm-up up for us. After the holidays things will really start taking off. We appreciate you being here with us this early!

You may also wonder what the ChAIos Media brand is all about. While we can’t reveal everything just yet, we promise you will see a lot more very exciting things coming from us in 2024. This newsletter is only just the beginning. We are cooking up some absolutely amazing things and we can’t wait to tell you more.

We are excited to have you on board from the very start and embark on this journey with us!

Amazing creations from the community

Every week we will feature some of the most awesome videos we’ve spotted in on social media. If you would like for us to consider your video, feel free to contact @AIandDesign or @bennash on X and send us a link to your masterpiece!

The Christmas Rifle, by Chrissie Pressman

Reindeer Revenge by Ethereal Gwirl

Ben & Marco’s 12 Days of Christmas in AI

Tips and Tricks

RunwayML text-to-video now supports multiple aspect ratios

RunwayML quietly launched a small but significant update to their gen-2 interface. You can now generate pure text-to-video scenes in aspect ratios other than the standard 16:9 one. This will greatly help people who need different formats. Getting a format other than 16:9 has always been possible when using image-to-video but until just now it was not possible with pure text-to-video.

Enhance your RunwayML gen-2 videos

Marco has performed some research this week to try and confirm what we have already suspected for a while: images upscaled in the amazing tool Magnific.AI actually make your RunwayML gen-2 generations look visibly better. In order to make the comparison as fair as possible, two sets of images were fed into RunwayML gen-2. A raw Midjourney render and a version upscaled in Magnific. The difference turned out to be quite striking. If you need to squeeze every bit of performance out of RunwayML gen-2, Magnific will deliver. Highly recommended!

Check it out for yourself in Marco’s post on X.

Check out Marco’s Visit Mars reel to see an entire video created using this pipeline. The seed images were created with Leonardo AI, upscaled with Magnific.AI and rendered into video with RunwayML gen-2.

Recent Movers and Shakers

Pika Labs has started rolling out to all users on the wait list

Pike Labs has started rolling out their amazing new model and web based UI to all users on the wait list. And the best part? Pika is still FREE! If you’re on the wait list you should have received a welcome email or you will soon. If you aren’t using Pika Labs yet: get on the wait list right here!

Additionally, for an in-depth interview with Matan Cohen-GRumi, Creative Director at Pika Labs, check out the recording of our X Space from December 3 2023.

Deforum Art has recently introduced paid plans and a new web app for current wave of beta users.

After their initial beta launch as a discord bot as a straightforward and user-friendly introduction to Deforum, they now have a beta web app with the ability to create videos from text and browse feeds. Deforum is an AI animation tool that can be used as an extension to the Stable Diffusion Web UI. The Discord bot and web app aim to reach more users by providing a hosted service. The beta sign up is currently closed. We’re hoping for more signups soon.

Deforum Art Discord Server

Deforum Art Website

Check out Ben’s “New York City in December” video made with Deforum Art

Introducing DomoAI, a new tool for images and videos.

DomoAI is an AI tool that combines multiple high-quality models, including Stable Diffusion, to generate images and videos. It’s currently available on discord. https://discord.gg/22kEnzdc

[generated with DomoAI text-to-image called Words to Pictures]

DomoAI Key Features

1. Words to Pictures - more than 10 models, specializing in Anime and Realism

2. Img2img - Pictures to Anime, Anime to Real life pictures, Pictures to Avatars, Sketch to pictures and more...

3. Image-to-Video Generation -  Generate Short Animation from Pictures.

4. Some great examples:

   https://twitter.com/bennash/status/1736789171602088364

   https://twitter.com/ai_for_success/status/1736407871674630296

A great open source LLM

Open Source AI is all the rage. “Regulate this!”, we say.
A very capable open source LLM has been doing the rounds. It’s called Mixtral 8x7b MOE and it’s pretty capable, surpassing GPT-3.5 in most areas. Some other notable things about it are the fact that you can actually run this on your own local machine provided you have the required hardware and the fact that it has way fewer restrictions and censorship in place than ChatGPT, Bard or Claude. This LLM will happily write explicit song lyrics for you, just to name one example. Grok, take note! You can play with it for free here on Neets.ai.

Midjourney V6 any day now

The wait for Midjourney V6 has been quite excruciating but it appears David and his crew will drop a massive Christmas present for all of us any time now. A “Rating party” in which users were to rate pairs of images started this past weekend. We are obviously just as excited as you are about these developments. And hey, now we all have something fun to do when the inevitable annoying politics discussions begin at the Christmas dinner table!

Previews for next week

Neural Frames just concluded a contest that was sponsored by ChAIos Media

The amazing Nicolai Klemke, creator of Neural Frames organized the first Neural Frames video contest. We can’t wait to see the amazing work submitted to this contest.

Our own Ben who is and expert Neural Frames creator, will be one of the judges in this contest. The contest has closed and therefore the winners will be announced very soon! We will feature them in the next edition of this newsletter.

Wanna learn more about Neural Frames? Listen to the recording of our latest Space in which we had Nicolai as our guest.

See you next week!

And it’s a wrap. Our first edition is done. We’re sure there’s a lot to improve but we’re learning as we are doing this. We hope you enjoyed reading it and we’re excited to bring you more next week!

Ben and Marco