One person made a 13-minute AI film. 100,000 people watched it. Then the comments quietly found what it can't fake yet.
The piece
One person made a thirteen-minute film with AI, and a hundred thousand people watched it. It's called Make Believe, by Montlouis Mane, made for the Higgsfield Global Film Festival. He says it's entirely computer-generated, and designed not to look it. It's called Make Believe, by Montlouis Mane, made for the Higgsfield Global Film Festival. He says it's entirely computer-generated, and designed not to look it.
It's called Make Believe, by Montlouis Mane, made for the Higgsfield Global Film Festival. He says it's entirely computer-generated, and designed not to look it. It's called Make Believe, by Montlouis Mane, made for the Higgsfield Global Film Festival. He says it's entirely computer-generated, and designed not to look it. It's called Make Believe, by Montlouis Mane, made for the Higgsfield Global Film Festival. He says it's entirely computer-generated, and designed not to look it.
If you've been told a machine is coming for the thing you do, don't take my word for where it's at. Listen for yourself. And I want to be straight with you: the visuals are good. Really good. One person did this. Anyone telling you it obviously looks fake hasn't watched it.
So here's where it actually breaks. And I'm not going to describe it to you. Thirty years in, I've sat in a booth re-recording one line forty times to get the breath right. Listen to what happens when nobody does that. [SILENT, source_audio evidence beat]
That's it. That's the gap. Not the lighting. Not the resolution. It's the timing. Nobody's listening to anybody. One viewer put it better than I can: the effects work is incredible, but the acting's tougher to nail down, like a Lifetime movie.
Another viewer sent the director continuity notes, a button that switches sides between shots. The director wrote back and thanked him. I love that. That's just filmmaking. And this gap narrows. It's not where it stays.
So put reaction on tape. Next reel, next self-tape, open with a scene that's all listening and timing, not a monologue. That's the part the machine still can't do.