Comment by perlgeek
> Solar panels only work because there's a temperature difference between the panel and the optical surface of the Sun.
I believe this isn't quite true (even though it's often repeated).
Consider that you could have an LED, at the same temperature as the PV cell, emitting light onto the PC cell, and so the PV cell would "generate" power without a temperature differential.
The Carnot cycle applies (in its initial form, at least), ONLY to generating power from temperature differentials, but PV uses the photoelectric effect, not temperature differentials.
This is not my area of expertise, but I think you can generalize the Carnot efficiency to talk about low-entropy energy sources instead of temperature differentials, but it's not quite as simple as associating a light wavelength with a temperature, because that doesn't work for radiation from something that's not a black body (like a laser or an LED).
I think here there is a misunderstanding on the meaning of 'work'. It's most likely clear to all authors in this chain that if you take an LED and shine it onto an appropriate PV you will get voltage out of it. That much isn't governed by thermodynamic limits.
Likely the word 'work' here is in the context of collecting or generating energy in some manner. Your example here is essentially energy transfer. In analogy what you're describing is the transfer of power from driveshaft to axle, not what is happening in the engine itself.