About this reportLarge models and finetuned derivative models will power the application layerRunning AI on the edge could push inferencing (and training) costs to the user.
