The assumption, Claudio, is that existing mechanisms in the system will contribute to finding a highly efficient equilibrium alongside our Groove and whatever other mods we have applied to our cars.
one of the main byproducts of combustion is being given away right now - heat. think how much faster a modern car will enter closed loop fuel delivery if some of that heat is applied to the waterjacket. Theoretically that means less fuel burned to get to where the o2 sensors start telling the injectors to trim. also using that heat to warm the fuel before injection will probably mean that it vapourizes more easily, so hopefully that will encourage the computer to trim back pulse durations (the o2 sensors might just see a rich condition more consistently)...and this will theoretically cause the engine to run cooler...so it's not adding significant extra stress to the cooling system or causing premature ageing of an engine or or or. will it burn the oil? no idea. will antifreeze need to be changed more often or the mixture changed? maybe. one step at a time...and small steps at that, right?
BUT -
you have to be prepared for (initially, while the computer HOPEFULLY learns the "new normal") the cooling system to dissipate what could be a LOT of extra heat on the way there., and in my case my electric fans turn on/of and kick into high gear at set points - I'm thinking it would be better for that to happen sooner, and I think that the best way to do THAT is to make my fans PWM controlled - so I don't melt my engine down to a block of aluminum if recycling exhaust heat into the water jacket/cooling system PLUS combustion heat pushes things to a limit I'd rather not face. and all of this means a cooling system that is fully intact - able to withstand pressures and dissipate heat. with electric fans, that also means added strain on the alternator.
so cooling system first, then warm the fuel...if you get a mileage improvement, THEN add the exhaust heat to the water jacket. Does that make sense to you?