I'm a little shaky on the thermodynamics of peltiers, but it seems (and I swear I read something years ago to back this up) that if it extracts power from the equalization of a heat differential, it must also slow down the conduction of heat. Two arguments for why this should be the case:
1. If it were not the case, then assuming ideal peltiers, you should be able to simply stack one peltier after another and get more energy than just one from the same temperature differential. And that's clearly crazy talk. If peltiers do not impede heat conduction, they must have some other limitation that prevents this from being the case.
2. Peltiers are reversible; a thermoelectric generator also acts as a thermoelectric cooler if you put current through it. So when you heat one side of the peltier and cool the other, generating a current, then that current should cause it to simultaneously act as a TEC (analogously to how counter-EMF causes an electromagnetic generator to resist as you turn it, essentially by turning it into a motor trying to turn in the opposite direction). That is, the very current you are generating will cause the peltier to transfer heat from the cold side to the hot side, with the end result that it would act as an insulator.
Edit: However, if I'm right about this, then assuming you can have efficient and flexible peltiers, it should work quite well for winter clothing.
I realized, thinking about this more, that my reasoning must be backwards on point number 2. If it worked as I said, then putting a current through a peltier would actually cause positive feedback resulting in a perpetual heat pump, which is clearly impossible.
So the current running through a peltier acting as a generator should actually transfer heat from the cold side to the hot side, making it act as a better conductor of heat. I don't know how to reconcile this with argument 1, but as far as my knowledge of thermodynamics is concerned, 2 has to trump 1.
So I figure I must be wrong, and peltier generators must conduct heat quite well. In which case I guess it would be good for summer wear - but then, in the summer, the differential between the human body and the surrounding air isn't all that great.
Yeah I was thinking of the approach used in the invention described in the OP, with a pre-chilled slug used to set up a temporary gradient. In that case you could lose too much body heat to warming up the cold mass.
But even without pre-chilling (i.e. in the steady-state case), while you may be right that a TE device necessarily slows down the conduction of heat across a gradient, I don't think it follows that any such 'suit' would result in overheating rather than hypothermia.
Consider 2 people in a 70-degree (F) room. One wearing light clothing, and the other sitting in a tub of water. The water in the tub will equilibrate at somewhere warmer than 70 degrees, and the person will become hypothermic, since the capacity of their 'suit' to conduct heat to the environment is greater than the ability of their body to generate heat to compensate. You could have a TE device that extracted power from the gradient between the warm water and the room without saving the person from hypothermia.
That is to say, I think we're both wrong :): a well-designed suit could keep the amount of heat transferred within the range of the needs and capabilities of the human body.
(Fun trivia: according to Murray Hamlet, a US Army cold weather injury researcher, the leading cause of 'morgue wakeups' in the US is drug users who pass out in the bath, become severely hypothermic, and are wrongly declared dead).
