I agree with the OP on most points, but if they're going to build a "proper" video camera, it needs XLR inputs, ND filters, controls on the body for all functions - along with a histogram visible while recording, peaking, and other video-centric features. And put a stinking full size HDMI on it for external recorders.
Anything without 10-bit 4:2:2 internal recording would be a wasted effort. With a log profile, 10-bit solves lots of issues Sony and Panasonic have - in many situations 8-bit log files are just too compressed. I hope Fuji is working hard to minimize those issues with flog out of the XT2 - but since its still 8-bit that probably won't be the case.
Black magic pocket and Micro Cinema cameras record Prores HQ and even raw to SD cards. That should be Fuji's goal.
4k60p and 1080p/120 or 240 fps would be needed to be a compelling alternative to others.
A parfocal cine version of the 16-55 would be nice too, no focus by wire. Fuji has been making broadcast lenses for decades, they know how to build them.
But all the above would be a $5000 camera, and then you're competing with the Canon C100 and Sony FS5 - so it had better be good.