AI Sandbox
The AI experiments below were completed by July 10, 2024. As AI is progressing rapidly, I decided to date-stamp these explorations on my website to provide context in case they quickly become outdated. I have included notes below each experiment.
The purpose of these AI experiments is to keep up with the ever-changing technology and to be prepared for the moment when AI could produce high-quality imagery at the required level.
Approach
Using Leonardo, Photoshop, Luma Dream Machine, Krea.ai, and mapping the outputs in After Effects, I was able to put together various concepts for visuals that would entertain Las Vegas onlookers gazing at the majestic Orb courtosy of Sphere Entertainment Co. Basically, I daisy-chained one AI to another AI to another AI, taking advantage of the unique strengths of each.
Control
As of writing of this post the Image to video options have only been available for one week. Currently, Luma only allows the upload of 2 key frames. Once we’re able to use 10 or more key frames, we’ll start to see improved control. So will see how things progress from this point.
Functionality
In terms of being able to generate quick ideas and try them in edit, AI is actually quite useful. At this moment, there's not enough control for it to impact productions as avertised, but I do think AI can make a significant impact between 10 and 35% (depending on the type of production).
Caveats
Technically, this experiment wouldn’t work as final outputs unless a designed seem was created. The Sphere requires outputs to be 360 video similar to if you were designing for VR. Nonetheless, as concept pieces and getting out quick ideas these are cool to see. Above and below are the eperiments unwrapped.
In 2011, my former colleague Clayton Douglas wrote a screenplay for an ambitious film project. Despite its creative promise, the film was never produced due to resource constraints. However, recent advancements in AI video technology have reignited interest in Clayton's work. On July 2, 2024, I created a proof-of-concept video showcasing the first 8 seconds of Clayton's film using AI-generated visuals. This experiment explores the potential of AI to bring previously unrealized creative visions to life.
