What OpenAI shipped
OpenAI highlights a workflow where developers can use a personal dot agent to triage on-call incidents. By allowing the dot to investigate issues and spin up a Codex thread, engineers can steer complex debugging directly from the ChatGPT app.
Demonstrating autonomous workflows requires balancing the speed of the machine with the comprehension of the viewer. When the agent handles the heavy lifting at 3am, the visual narrative must clearly separate the AI's background investigation from the human's mobile steering.
How the motion works
To establish the urgency of a 3am pager alert, a designer might use a rapid sequence of hard-cuts between a dark environment and the stark brightness of a phone screen. This immediately sets the context and stakes without relying on heavy exposition or voiceover.
When demonstrating the dot investigating the issue, motion can guide the viewer's eye through the triage process. A slow punch-in on the ChatGPT app interface could emphasize the autonomous work the agent is doing, drawing focus entirely to the generated Codex thread while cropping out mobile OS distractions.
For UI interactions, managing hold-time is critical. When the engineer steers the work from their phone, holding on the final text prompts before execution allows the audience to read the command. Proper easing on the scrolling text then mimics the natural rhythm of human review, contrasting with the instant generation speed of the AI.
What to steal from it
- Use stark lighting contrasts to establish time and urgency without relying on voiceover.
- Employ a slow punch-in to direct focus toward autonomous UI actions and crop out unnecessary interface elements.
- Pace UI scrolling to match human reading speed rather than machine generation speed.
- Rely on deliberate hold-times to ensure complex code or text prompts are legible on mobile screens.
Want a video like this for your product?
Impractical cuts one from a single prompt — the same motion craft, in about twenty minutes.
