Tim Simmons interviews Javi Lopez about a 17-minute Backrooms film developed over roughly six months from about twelve hours of generated footage. Lopez describes writing detailed prompts for individual shots, actions and dialogue, then selecting and editing the outputs rather than generating a finished film in one step.
Long handheld camera movements were especially difficult: some shots took many attempts, while shorter cuts were easier to obtain. Lopez describes working with lower-resolution drafts before upscaling chosen footage, an approach that reduced the cost of discarded generations but did not eliminate manual iteration.
Post-production remained central. Lopez discusses editing in DaVinci Resolve and building sound from recordings and libraries rather than relying on generated audio alone. He argues that collaboration across editing, voice and sound can improve future projects. The account illustrates one filmmaker's workflow, not a guarantee that AI removes the need for a production team.
Watch on YouTube
