Humans perform at least two tasks:
- Idea generation
- Idea communication
Generative models can help or replace both of these, but to some extent humans will want to continue finding pleasure in these activities.
One way to view communication is to synthesize smaller micro-ideas into larger macro-ideas. For example, an author of a novel might jot fragments in their notebook over years, then finally combine and polish those fragments into a published novel. Generative AI will ease the burden of synthesis so humans can focus on conjuring fragments of ideas. For example:
- You have an idea for a YouTube video. You will write down ideas, a sketch of a narrative, fragments of videos indicating certain gestures you want to use to communicate the idea, and inspiration clips from other videos. Then a model will produce several candidate videos from those ideas and you will recursively provide feedback until the video meets their standard. Complex video-editing software will no longer be used.
- You want photos of your wedding. Instead of hiring a professional photographer to take and edit photos of the important moments, you’ll hire people to film as much as possible in high-res 8k video, then a generative model will stitch these videos together into the photos or videos (or animated 3d scenes) you want of your wedding.
- You want to write memoirs of your life. Simply keep a journal.