Amazon, wrist-bands and AI
Amazon is king of optimizing workflows. From 1-click book purchasing, to same day delivery, the optimization is tangible. Many of these gains are fruits from the massive data grove they have grown over time.
After reading recent reports of a wristband requirement for Amazon warehouse workers, I thought of course, more data! However, much of the coverage focuses around workers rights: “Amazon wants to wring out more from their low paid transient workers.” While Amazon's extremely capitalistic motivations are clear, I believe it misses a much bigger point. Amazon can’t scale if it simply improves efficiency on the backs of human workers. Knowing when someone works or how utilized they are may be side-effect of the wrist bands, but not the primary motivator.
Amazon wants to train robots on how to pack more efficiently. They want to know which motions and movement patterns relate to highest quality box packing. They want to know what general and motions people are making. Where are the robots you may ask? Instead of building expensive robots and training them on the job with high failure rates, why not train in software and back test against real world data? I bet these wrist bands have more use as training data for different types of networks. The robots are coming.