Great Idea, But...
In World were we still waste the heat extrated by refigerators and feezers; let alone the heat in a cooker after the roast dinner, were it could be used to heat up the hot water is clearly showing a legacy issue in design implementation.
WE still can't agree on a standard wall socket for power let alone frequency, worldwide. Heck the closest we have come to that is the USB port,
So we end up with great idea's that everybody know's makes sence but alas the cost to implement can only be accomodated if it is catered for by design and planned out initialy. Sure hindsight os great but look at the UK raileway structure, becasue any improvements have to use the existing infrastructure you are limited to what you can do and in this case, bigger trains, longer trains etc are limited. The only way around that is to majorly upgrade the existing tracks/stations and the cost and impact of doing that is far greater than just doing a new one in conjunction, not like you can get a train into london across london and out the other side without having to change already so it's doable from that aspect.
This leads us back to the whole crux of it that it's viable for new building accomodating this approach more than modifying something already inplace and alas the ability to pump hot water into homes in the UK is somewhat limited. Iceland on the other hand you could probably do that, though only becasue they already do that using already natural thermal resources to heat the water.
But back to microsofts proposed future patent pool. Distrubuted processing aka cloud computings has some uses, but what there proposing is in many ways moving into area's the utility companies are more suited to accomodating and indeed implementing. Heating and cooling would be a logical progression for them, though nightmare to manage and that leads us onto the other area and that is if your cloud cluster is pumping out X amount of heat 24/7 then you need to realy have that heat being not the soul source of heat needed and indeed the miniimium, unless you also control the workload. You have to also look at the variying usage and indeed cost of electricity of the 24 hours of a day and indeed vairations and what impacts that (yes we all know TV does and kettles). This is were you get the human factor and we're not very good at planning, heck why a kettle has to be able to boil water in seconds instead of doing it in a more fuel effecient way and take longer is a great example and something that wont change soon - who would pay for a kettle that took 15 minutes to boil though used half the electricity of one that boiled yoru water in seconds; Whilst that may be a few pence that most will happily pay for the conveniance of having it now, multiplied by all those who take that approach and you see one of the many issues in managing resources.
So for a few niche area's this may have some ground though there are clearly way more wasted heat and other resources that we have already that could be better refocused.
What next, cordless mice that have small balls at each corner that charge micro dynomo's powering the mouse so the more you use it the more power you use the more power you generate; go invent that microsoft and stop stating common sence like it's something nobody has thought of with the angle to patent it.