As for the comparisons, I plead haste.  (I noticed, too late, that I hadn't dealt with the amount of heat we get from natural gas and coal, directly and indirectly.  That would have to figure somewhere, and I intend to go back and insert a note to that effect.)

Delivering heat to homes isn't necessarily impractical.  If you are willing to put a nuclear plant in tunnels beneath a city (and what better place to put it to eliminate the threat of terrorist attacks?), you could transfer the heat as medium-pressure steam to neighborhood energy recovery turbines [1] and then the exhaust low-pressure steam or hot water for space heat.  The water goes back down to the steam generators by gravity. [2]  I did a writeup on this almost two years ago.

When heat is not required for space heat or to drive absorption A/C, it could be vented through cooling towers.  These might be integrated with office towers or other buildings.

Electric transmission losses, battery losses etc. would probably be on the order of the efficiency gains from electric drivetrains.  This looks close to a wash.

The one thing I didn't consider is higher-grade heat requirements for e.g. industrial process heat.  This could also be supplied by nuclear (which was the original intent of what became the Midland Cogeneration Venture in Midland, MI) but there would be a greater impact on electric output.

[1] Medium-pressure steam is probably better than low-pressure, because the pipes will be much smaller, cheaper and have lower heat losses.

[2] If the reactor is deep enough, gravity could provide a large part of the pressure required for the boiler feedwater.