ChAIos Media News - February 12 Edition

Hello, fellow generative AI enthusiast. This is the first ChAIos Media News for February. We had to skip last week because we were recovering from pulling some all-nighters working on our RunwayML gen:48 projects. As always it was great fun and we will highlight some stunning creations from community members in this edition. Let’s dig right in!

Awesome creations from the community

Black Llama Corporation

Marco has created a mock SuperBowl ad for a service that provides AI powered protection for anyone walking around town wearing an Apple Vision Pro headset. It has been entered into superbowl ad contest by FBRC.ai.

The Apple Vision Pro Face Replacer App

This entry by Creative AIgency had us laugh out loud. Just like Marco’s film this is a parody / mock advertisement related to the Apple Vision Pro. By far the funniest gen:48 entry we’ve seen.

Chrissie impresses with her gen:48 submission “The Unseen Artist”

EVen though this entry did not make thelist of gen:48 finalists, it sure managed to capture the hearts of many by storm, including Chris Anderson’s, the head of TED, who awarded this film with a shiny new Apple Vision Pro headset. Congratulations, Chrissie!

Mice City - a gen:48 entry

Shane and Reza impressed us with their awesome 1980s themed gen:48 submission “Mice City”. You’re going to want to watch this one! We think this entry is particularly amazing because it uses both RunwayML gen-1 and gen-2 in a comprehensive piece.

FinalFrame releases V2 to the public with new UI and features

The indie project FinalFrame is progressing, introducing a comprehensive update with its Version 2 release, which showcases a completely redesigned user interface. This update brings an array of new features, including capabilities for converting text to image, text to video, and image to video. It also introduces a timeline feature for assembling videos from shorter clips and a function that allows videos to be extended from their final frame. Upcoming additions to their toolkit include an image upscaler and a native video upscaler for 2K and 4K resolution videos. We are eager to explore the capabilities of this new video upscaler, as the video quality appears to be competitive with other Stable Diffusion video generators currently available.

More info https://finalframe.ai/ and https://twitter.com/FinalFrameAI

Check out this experiment from Ben where an image from Midjourney was used to create this video in FinalFrame V2

The Runway Gen:48 film contest is currently open for voting

The GEN:48 film contest is currently in its voting phase for its second edition. The voting period will remain open until 11:59 pm ET on February 14. During this time, viewers have the opportunity to vote for up to three films that they believe deserve recognition. With several hundred films submitted, Runway has carefully selected 40 films to be presented for public voting. It's important to spread the word about your favorite films and encourage others to participate in the voting process. This is a great chance to support talented filmmakers and help decide the winners of the contest.

Vote here: https://gen48.runwayml.com/vote

Runway’s announcement:

In case you missed it, Ben and Marco dropped a special show in which they discuss nominees and non-nominees on the day the list of finalists was published.

AI has gone mainstream with the focus in Super Bowl commercials

During this year's big game, several commercials featured AI as a prominent topic or product feature. One notable example was the Minions, who were shown using an AI text-to-image generating tool to create images. The commercials also parodied AI image hallucinations and the concept of generating extra limbs through AI.

If you're interested in exploring this further, we recommend checking out Rowan Cheung's thread, where he has compiled a collection of these “Super" commercials. It's fascinating to see how AI is being integrated into popular culture and advertising.

Canva launches Magic Studio, it’s most ambitious AI project yet

Canva has recently launched its most significant product launch yet: Magic Studio. This suite of AI-powered tools aims to empower users, from individuals to large organizations, by bringing the power of artificial intelligence into a single platform. Magic Studio offers features such as Magic Design, which generates sophisticated designs based on user prompts or uploaded media; Magic Switch, an innovative tool that transforms content across various formats and languages with just one click; and Magic Grab, a time-saving image editing feature that makes any photo editable like a Canva template.

Magic Design for Video helps users create engaging professional videos in seconds, streamlining the process by bypassing manual editing tasks. Magic Design for Presentations allows users to generate presentations with unified storytelling, outlines, and content using just a few words as input. In addition, Magic Media enables users to generate perfect visual assets at their fingertips, even for objects that don’t exist yet.

To ensure safety and security while using AI in the workplace, Canva has introduced Canva Shield, an industry-leading set of safety, privacy, and security controls. Role-based permissions enable administrators to control team access to AI-powered tools as their organization evolves and grows. Magic Studio empowers users by simplifying complex design tasks, accelerating creative workflows, and providing an intuitive platform for generating high-quality content with confidence.

Google releases Gemini, it’s latest and greatest multi modal LLM

Google has recently unveiled Gemini Ultra, marking a significant milestone in the realm of large language models (LLMs). Sundar Pichai, CEO of Google and Alphabet, emphasized the transformative potential of AI in various aspects of human life and innovation. Gemini Ultra, a product of Google DeepMind, is designed to be multimodal, meaning it can process and integrate various types of data, including text, images, audio, and video, to perform a wide array of tasks efficiently. This model is part of the Gemini family, which also includes Gemini Pro and Gemini Nano, catering to different needs and computational capabilities​​​​.

Gemini Ultra stands out for its exceptional performance, surpassing current benchmarks and even human experts in certain domains. It has demonstrated superior capabilities in understanding natural images, audio, video, and complex problem-solving, such as mathematical reasoning and coding tasks. Its flexibility and efficiency make it a powerful tool for developers and enterprises, aiming to revolutionize how AI applications are built and scaled. The model's state-of-the-art performance is evident in its ability to outperform the previous benchmarks set by models like GPT-4 in a variety of tasks​​.

For users and developers interested in exploring Gemini Ultra's capabilities, Google has integrated the model into its services, including Google Workspace apps like Docs, Slides, Sheets, and Meet. We are particularly excited about this integration as it had the potential to provide unique value by having access to our personal data on Google’s servers.

The model will also be accessible through a new $20 Google One tier, offering additional features like 2TB of storage. Moreover, Gemini Ultra will power the rebranded Google Bard, previously known as Gemini, and will be available in more than 150 countries, initially in English, with plans to add more languages in the future​​.

For more detailed information on Gemini Ultra's capabilities and the broader vision of Google's AI advancements, you can visit the official announcements on Google's blog and DeepMind's page.

Last Sunday, we discussed Gemini in detail during our weekly space on X. Among others, we had our friend Alex Northstar as our guest to discuss his first impressions of this exciting new chatGPT competitor.

Listen to the recording of our X space on Sunday, February 11, 2024

And that’s a wrap

That’s it for this week. Thank you for your support and see you again next week! Got a tip, want to be featured, interviewed, or have something you think should make it into next week's edition? Reach out to Marco or Ben on X, or drop us an email at [email protected]. Your input is always welcome!