A creative Chinese engineer has just built a futuristic prototype where hand gestures trigger specific actions within an AI agent's interface, as shown in the video.

✌️ Two fingers up = run the program
🤌 Rotating fingers = generate a new idea
🤏 Pinch (bringing two fingers together) = perform heavy-duty tasks

The interface uses a webcam, MediaPipe, and TouchDesigner to capture hand and finger movements in real time.

The example in the video demonstrates the precise synchronization and ultra-low latency between the movement and the agent's reaction.