Apple researchers introduced Minions, a design that distributes AI inference and embedding jobs across idle Apple Silicon Macs. It uses a central work queue, PostgreSQL to prevent duplicate assignments, and storage for data transfer, reassigning jobs that stop responding for a specified period. The paper describes some jobs starting within minutes instead of waiting roughly ten hours for a central GPU, but this is not a like-for-like comparison of raw computing speed. The system distributes independent jobs that each machine can run; it should not be construed as memory pooling for a single enormous model or as support for training.