I dont know if this is a 'hard' problem, but to me its a problem that has not been 'easily' solved:<p>Dynamic Digital [media] Endpoints: (think digital signage)<p>When I was at Cognitive Networks, there was a function where you could tell the TV what to display (for dynamic commercials)... But here is the premise of what I mean when I talk about a Dynamic Digital Media Endpoint:<p>power on device --> registers with CONTROL --> supplies its manifest of functions if new, else provides its ID as online --> CONTROL gives the device an API endpoint to register with --> Device contacts api endpoint --> API endpoint streams slated media to device.<p>So, for example, I can have a MPV media server running, with an app that any of my devices can be told to register with, declaring their manifest of what media they can accept (play a video, webpage, audio, signage) - and then you just assign content to the device. Whenever it either boots (like a smart TV) it will check in then play what the control tells it. If its a tablet (kiosk) it will just load up that webpage. In your home - you can have a central home management platform and all your devices register to it and you can push whatever content you want to any screen in your home.<p>Ive done digital signage, wayfinding, routing, geotracking stuff in the past, and I am not sure the current state, but IME previously - this was never elegantly solved.<p>With the Magic-Wormhole threads and such a lot of really cool things are on the horizon with people being able to build API-Legos via being able to quickly iterate little API widgets with GPT helpers these days... is so fn neat!.