Google detailed an internal experiment in which teams of AI agents produced short films, testing whether agents could collaborate outside software work. The exercise involved 10 crews, each made up of three agents with distinct roles, working through a filmmaking process inside Scion, Google's open-source multi-agent orchestration testbed. Across pilot rounds and the competition, the project produced more than 25 films totaling about 44 minutes. Each crew included an Idea Person, a Technical Lead, and an Editor, with a Team Coach agent supervising checkpoints and a Coordinator agent scheduling the contest in five waves over about 21 hours. The crews followed a seven-step production process based on traditional filmmaking. A control system requiring one agent to check another's work was added after problems in earlier trials, including a pilot where a team reported finishing a film, but the output was a 94-byte placeholder file. Human reviewers fed comments from pilot rounds back into the process, covering issues including overlapping audio, character inconsistency, and hard-to-follow stories. The toolchain combined command-line tools written in Go with Python batch automation, using Gemini image generation and Veo 3.1 for video clips, typically in segments of four to eight seconds at 720p.