Unrelated but maybe not
Not long ago I deployed IP cameras around my building - high traffic area with lots of tourists and others. I had the rest of the infrastructure (Cisco VSM) and my platform is ONVIF compatible, so I went to Amazon and bought some fairly generic ONVIF compatible cameras, rather than paying Cisco tax. The cameras work as needed and are actually nice, but the bundled software was amazingly bad from a security standpoint - will only run on windows, must be run from a browser, browser must be Internet Exploiter, turn off ALL security for the session with the camera app, install these plugins, trust the camera app to do lots of things it should not need to do ever....and then you are able to blow a new IP address into the camera. However, the camera was configured out of the box to connect to various "Free" services automagically and had factory settings that would have put the camera right on the internet and likely checking in with some CiC location when first plugged in if the user had used the default settings in their consumer grade router.
No way to simply log into the camera and set IPs as I would expect, you had to deploy the craplication to configure the camera at all.
A sandboxed VM was used to re-IP the cameras for the PoE subnet (and subsequently deleted), the camera switches are on a discreet switches with their own dedicated subnet and an invalid gateway, and the firewall supporting the does not show anything unexpected. But the out-of -box experience caused me to realize that this sort of IoT crap can be an entirely new attack vector.
Biting the hand that feeds IT
