Sabitlenmiş Tweet
I am not new to this game at all.
This is in fact my third account that I have spun up as an outlet for these explorations. Each iteration has taught me something different.
Each burnout and restart has peeled back another layer of what's actually happening here, beneath the hype and the breathless X threads.
And here's what I keep coming back to: I feel inspiration again when working with Gen AI video tools.
Not the kind of inspiration that comes from watching someone else's demo reel, but the real thing—the kind that wakes you up at 3am with an idea you need to test immediately.
The kind that makes you chain together six different tools in ways their creators never intended.
This shot, for example, was the result of a multi-step workflow combining a mix of image and video generation work. I started with a rough concept, generated a base image in Midjourney, brought it into Ideogram for further iteration as an image prompt, then across to Leonardo Motion for video, using a modified version of my initial MJ prompt.
None of these tools alone could have achieved what I wanted. But stacked together, with intention? That's where the magic lives. Stacking like this also adds a level of nuance and character that single model gens can’t quite grasp.
These are still early days, and all kinds of pragmatic workflows are not talked about in the chase for the newest and best generation from the latest model.
We lose sight that there are very straightforward pathways to dependable and creative video generation. Current media can be used, remixed, repurposed.
Old photos, magazine images, screenshots from films on Tubi that you bash together.
//It was never about the prompts.//
Not really. It is about the process and the thrill of discovery:
Creativity
Experimental Mindset
Workflow Hacking
Investigation
Knowledge of film / literature, language and genre
The best prompt in the world won't save you if you don't understand visual storytelling. And the worst prompt can still yield something interesting if you know how to work with what you get.
//Where This All Goes//
Long form content might not be the place for AI generated video. At least not yet. Maybe we need to adapt to tell real stories in different mediums?
Perhaps the future isn't AI replacing traditional filmmaking, but AI enabling entirely new forms. Hyper-personalized stories, impossible worlds rendered in real-time, interactive narratives that shift based on viewer input. Or maybe it's something smaller and stranger: visual poems, experimental shorts, tone pieces that exist somewhere between film and dream.?
I don't have the answers yet. That's why this is my third account, and probably won't be my last.
There's something here worth digging for.
And I intend to keep digging.
English
