Well Uverse and DTV software running the boxes are different, so unless they decide to get rid of the DTV boxes, those people are safe. Now a few years in the future if they make a change, say year 3 after the close, then those people would be out. They could try to port one of them to the other, but I don't see why because in 3 years they will need to have new boxes out.
If/when the merger goes through, I think it would make sense to get the top Directv and Uverse engineers together to design a new box, using a software tuner. Digitally sample from 54 MHz to 2150 MHz to cover both cable and satellite ranges, including DECA for Directv and MoCA for Uverse, and demodulate in software. It could then tune Uverse QAM channels or Directv DVB-S2 SWM channels. Even ATSC, if they decided to add an OTA port and a couple mixers to shift the frequencies around a bit so everything fits.
The number of channels it could tune at once is basically limited only by computational power. The technology to do this exists in chips on the market now (google "full band capture" for Broadcom's and "full spectrum capture" for Maxlinear's) I think one of them has a chip that can tune 24 channels at once, though the computation required for a 6 MHz QAM256 channel is less than that for a 36 MHz wide DVB-S2 transponder but if it can't do all 9 SWM channels at once the next version would.
That's your "home server", everything else is a client. Maybe using RVU, maybe using something new. If they could get together with Comcast on some standard you'd probably see all new TVs supporting it instead just a fraction like with RVU since between them they'd have over half the TV subscribers in the US. The other cable companies and Dish would probably follow.