I didn't back it either, but from looking at the Kickstarter page it seems like they had a viable product that was more or less working (Raspberry Pi clone running the stock Minecraft server, pre-configured to work with a dynamic DNS service that the company hosts) but then kept tacking on features after the project ended to make everything "perfect" rather than shipping. They added a web admin interface, a variety of custom Minecraft servers, seamless UPNP support, redirecting the server traffic through a VPN since they couldn't get UPNP working reliably, etc. They should have just added a FOTA capability and worked on that stuff after the fact if they got enough continued sales to warrant the R&D work.

I think the other mistake they made was offering a WiFi option rather than just making it ethernet only for the initial launch. It seems like they had a lot of trouble getting the WiFi working and ended up having to switch WiFi cards. I don't know how they would have made WiFi network authentication work reliably on a headless device, and it seems like a ton of complicated work vs. just having the customer plug the box into their router. Again, it's a "nice to have" that could have waited until after the initial shipment.