i'd been wondering if/when we'd see something like this- packaged up optimized hardware for local inference at home... 60 Gb vram, i find it attractive but will enough consumers & will the integrations work?