I treated AI video as a toy for other people. My picture of it was a studio somewhere, a render farm, a team of people who knew what they were doing. I run IBRA alone from a desk in Baghdad, so I filed the whole thing under "not for me." That changed this week, when I tried Google's Veo inside Gemini and watched it turn one clip of my face into three short films with me acting in them. I was wrong about the cost of entry.
Uploading my own face was the entire setup
One clip in, no crew, no lights.
The tool asked for a short video of my face. It asked for a sample so it could match my voice. That was the whole preparation. No camera rig, no editing timeline, no second take. I gave it what it asked for and then typed a few short prompts describing the scene I wanted. The part I'd assumed would take days took a single upload.
Each film came back in about a minute
A minute is faster than I could film one shot.
I made three clips. Each one finished in roughly a minute from prompt to result. To put that against the old way: a single real shot on a phone still needs setup, framing, a retake or two, then trimming. Call that ten minutes at best for one usable second of footage. Veo handed me a full short scene, with me in it, in about sixty seconds. The gap between those two numbers is the whole story.
The voice match is what unsettled me
Close to mine is a strange thing to hear.
The generated voice came back close to my own. Not a robot reading my words. Something near enough that I recognized myself in it. Handing over a face is one thing, since photos of me already exist online. Handing over a voice sample and hearing it played back inside a character I never performed is a different feeling. It worked well. That's exactly why it gave me pause.
Think of it like a stamp, not a camera
The tool prints the scene, it doesn't record it.
A camera captures something that happened in front of it. Veo works more like a stamp cut from my face and voice, pressed onto scenes that never took place. The person in the clip is me, and also isn't. Once I saw it that way, the speed made sense. Nothing was filmed, so nothing had to be. The tool builds the image from the sample instead of recording the world.
What this shortcut costs
Speed this cheap has a bill attached. A face and a voice that can be cloned in one minute can be cloned by anyone who gets a clip of me, not only by me. I traded a real barrier for a real convenience. The same tool that let me star in three films could let someone else put words in my mouth I never said. I don't have a fix for that. I only have the awareness that both sides came in the same box.
I made three short films of myself this week without owning a camera. One tool, one afternoon, one face. I still don't know where the line sits between using this and being used by it. For now I'm keeping the clips, and keeping my guard up.
