A few weeks ago, I realized something strange.
I wasn't saving the videos I generated anymore.
I was saving the prompts.
Recently I've been spending a lot of time experimenting with AI video creation after work.
Not because I need videos for a project.
Mostly because I'm curious about how different prompts change the final result.
My current workflow is surprisingly simple:
-
Start with a single idea
-
Write a short scene description
-
Generate a 5-second test video
-
Adjust the camera movement
-
Repeat until something feels right
One of my recent experiments was:
"A quiet coffee shop on a rainy evening. Warm lights. Slow camera push-in. People reading near the window."
The first result looked okay.
The second result felt more cinematic.
The third one finally captured the mood I had imagined.
That's when I started saving prompts instead of videos.
Because I realized the real creative asset isn't the output.
It's the idea behind it.
Lately I've been using Wan 2.7 for these experiments:
https://www.jxp.com/wan/wan-2-7
What I like most is the ability to quickly test visual concepts without opening complicated editing software. Wan 2.7 supports text-to-video, image-to-video, reference-based generation, and fast iteration, which makes experimentation much easier.
Some of my favorite scenes so far:
🌧 Rainy city streets
☕ Coffee shop interiors
🌙 Late-night urban environments
🎬 Cinematic character moments
I'm still learning every day.
But AI video generation has become less about technology and more about collecting visual ideas before they disappear.
And honestly, that's probably why I keep coming back to it.