That was the whole point of my post -- heat pumps Transfer heat from one place to another and in this case they are taking heat generated by the home heating system. Frankly, a tankless coil on the furnace would probably be more efficient because, in any case, in winter, you are heating the water with the home heating system. Nobody said anything about creating cold -- There is no such thing as cold anyway, it is the absence of heat.
A tankless coil is about as
inefficient a method of heating water that's still legal in the US. You break even or pull ever so slightly ahead of a standalone fossil-fired tank during the heating season, but keeping a high-mass boiler hot during the rest of the year just to support a DHW load has truly abyssmal efficiency due to wretched standby loss & cycling loss performance at such a low averge load. Indirect-fired tanks do far better than embedded coils, but even they suffer serious cyling loss inefficiencies in the off season when married to high mass boiler. (Post-burn boiler heat-purge into the indirect can improve performance considerably though.) The bigger the boiler the less efficient it is as a water-heating system, but in most homes an indirect is a net gain in annual efficiency, whereas a tankless coil is a net loss.
You're correct on the efficiency analysis of the heat-pump water heater over some part of the year, but during the cooling season it's a HUGE benefit compared to a standard electric tank, by taking a load off the cooling system. (It's not more efficient that a desuperheater on the central AC or heat pump system though.) In much of the southern US there's a far bigger annual cooling load than heating load.
But that's only half the story. From a net-efficiency of fuel-to-load point of view, electricity is only ~25-35% efficient in the average fossil-fired grid. About 60% of the source fuel energy goes up the flues & cooling systems at the power plant before the juice hits the primary winding of the first transformer. (Combined-cycle gas powerplant do much better, but that's only bringing up the grid average.) If you're reducing your electricity use for heating water by ~50%, by pulling half the heat from a space heated with even an 80% efficiency fossil- furnace or boiler or even a heat pump (COP >>1) you're ahead of the game- it'll take less source fuel and cost less to heat that water. If you heat with electric resistance (electric baseboards/radiant/whatever), it's a wash during the heating season, but a gain during the cooling season. If your heating fuel is propane it'll usually be cheaper to run a heat-pump water heater than a propane-fired water heater, but you'd have to do the math relative to your actual electric & propane rates to be sure.
The heat transfer issue in refrigerators is totally NOT overlooked (by regulators & utilities, anyway.) The efficiency of that power used matters even in heating-dominated climates where that compressor-heat is supporting the heating load. In most places heating via electricity is quite expensive compared to other sources. In cooling dominated climates the compressor's heat is an addtional sensible-load to the cooling system. Getting rid of all of the older less efficient refrigerators & freezers and replacing them with even minimum-efficiency newer unit would take a HUGE base load off the grid overnight. (Many utilities have bounties on them- they will come and pick 'em up and write you a check on the spot, since it's cheaper to buy a few million old refrigerators at $50-100/per than it is to build a power plant to support even half the base load they represent.)