Amazon spent millions putting a lost dog in front of a hundred million people during the Super Bowl. The ad was thirty seconds long, emotionally bulletproof, and a perfectly accurate depiction of a networked AI surveillance system activating across an entire residential neighborhood. They just made sure you were crying when you saw it.Meanwhile, in Tucson, an 84-year-old woman had been missing for over a week. Her Nest doorbell footage was declared unrecoverable by local law enforcement. Then the FBI pulled images from what Google called "residual data located in backend systems," which is a very polite way of saying "deleted" doesn't mean what any of us thought it meant.This piece traces the architecture underneath the doorbell camera, from Ring's Search Party feature to Flock Safety's twenty billion monthly license plate scans to Discord requiring 200 million users to submit their face or their papers, and asks the question nobody in the Super Bowl ad wanted you to think about: who else do the cameras see when they're looking for the dog?If you want the full reporting, the sourcing, and the deeper infrastructure story, everything is on the Substack below. That's also the best way to support the work so I can keep doing this.Substack: thedreydossier.substack.comFollow me onTikTok: / thedreydossierInstagram: / thedreydossierIndependent journalism only exists if people can actually see it, so if this connected some dots, share it with someone. Thanks for being here.