> Everything runs on the router itself. If the rest of your infrastructure is on fire, this still works, which was most of the point. And then a moment later:

> Everything runs on the router itself. If the rest of your infrastructure is on fire, this still works, which was most of the point. And then a moment later: > You need about ten minutes, a router you own, and a card on file with an LLM provider. So it depends on a working internet connection and an external model server. No hate on you for making this - it’s a fun idea, but the first bit is pretty misleading