Will AI kill the camera?

Preposterous!

I’m reading about AI today and clicking the links in the article. Trying out generating AI video. The use case is painfully boring and generic corporate videos. (They suggest adding an AI avatar “talking head” to a PowerPoint presentation.)

But it’s not difficult to extrapolate from this. I thought of my old Teletrivia videos. Now I could create a fake host, with their own voice, and animate the video using a script and some tweaks. I could say the script out loud and then cause it to be. Given the capabilities we’re already seeing in the new generative AI in just a few months of release, what’s realistic when we look ahead five years, say?

If I trained an AI using a high-resolution corpus of every movie in existence, matched to its script, how long before I could create an entire movie just by writing a screenplay?

I can already conjure beautiful, photorealistic images of people, places and things that don’t exist with simple text description. What happens when we run applications that can do that in real time on quantum computers powered by nuclear fusion?