This video was made almost entirely by AI. I used ChatGPT to write a script, Midjourney to create reference images, Runway Gen-1 to apply the style of the images to my source video, and Boomy AI for the music. Workflow breakdown w/ comparisons in thread. 🧵

Before we get into it, here is a comparison shot of the original footage and the Gen-1 output. And yes, that's a roll of toilet paper lol

Step 1 - Generate a script in ChatGPT I asked ChatGPT to "Write a script for a 9 second video composed of three 3 second clips. The story should feature a man in his living room and be sci-fi themed" This was what it gave me 👇

Step 2 - Make reference image in Midjourney Based on the script, I asked MJ to /imagine sci-fi film still, medium shot, centered, side-angle view, a man sitting on a chair, holding a glowing orb in his hands, living room, new york, 4k --ar 16:9 This was the image I picked 👇

Step 3 - Capture footage on my iPhone I rearranged some furniture in my living room to better match the reference image in hopes of having cleaner outputs later on. This was the original shot of the first clip 👇

Step 4 - Upload video clips & reference image to Gen-1 Gen-1 is a tool that lets you "apply the composition and style of an image or text prompt to the structure of your source video." I didn't use a text prompt, but I did add --interpolate as a parameter for smoother output 👇

Step 5 - Bring clips into a video editor I just used iMovie for this, nothing fancy. I imported the clips, trimmed some up, and strung them together.

Step 6 - Source AI-generated music from Boomy Boomy lets you create generative music based on style and mood. The track I picked is called Trippy Harvester by Peaky Chicky. I thought it paired well with the mood of the scene.

Final step - Add music and export Luckily I didn't have to edit much cause the clips aligned pretty nicely with the music. The final export 👇

Follow me @nickfloats YouTube vid coming this week (link in bio) More content you might like 👇

BTW this took me about 3 hours in total. Most of that time was spent reshooting clips on my iPhone and testing out a bunch of different parameters in Gen-1 to try and get the smoothest output.