Earlier quoted context omitted.
Have you noticed how microwave cooking is not warming up the skin any faster than the insides? The wavelength is selected so that it penetrates and warms water molecules. It would cook your insides according to their water content. Surface of the skin has less water than tissue inside, so you would feel pain beneath the skin.
MythBusters tested that, the outsides (of food) heat up first.
Microwaves attenuate as they penetrate a resistive or high-permittivity substance, because they deposit some of their energy as they pass through. In the limit of infinite depth, no radiation survives and there is no heating. Given a homogenous substance, the heating is always strongest at the surface, decaying exponentially (to a good approximation) as you go deeper.
However, heating is not the same thing as temperature. The surface of food in your microwave is exposed to room-temperature air and can therefore cool down by conducting its heat to that air. Food just under the surface can conduct its heat to the surface food, while deeper food cannot. It's easy to set up situations where this results in a reverse thermal gradient penetrating some distance into the food, or even all the way to the center. This is one of the most significant advantages of microwave heating in industrial material processing, because there are many cases where the normal thermal gradient produces cracking and microwave-induced reverse thermal gradients do not.
You can set up a forward thermal gradient with microwave heating by some combination of hot air, strong attenuation, short exposure times, shorter wavelengths, and great depth, although unless the air is actually hotter than the highest temperature reached within the food there will always be a reverse thermal gradient present at the surface, since the heat equation always produces a continuous temperature field at t ≠ 0. You may be able to get the reverse thermal gradient to be entirely outside the solid body if you work hard enough, but a much easier way is to only measure the temperature at intervals large enough that the entire reverse thermal gradient is smaller than the first interval. For example, if the thermal maximum is 8 mm under the skin, you could measure the temperature at the surface, 1 cm, 2 cm, 3 cm, 4 cm, and 5 cm.
The attenuation is dependent on the attenuating medium and on wavelength. As explained in https://en.wikipedia.org/wiki/Active_Denial_System the ADS uses 95-GHz radio waves with 0.4-mm penetration depth in human flesh, while your microwave oven used 2.4-GHz radio waves with 17-mm penetration depth in human flesh.
nabla9 has been downvoted for politely and straightforwardly, if briefly, explaining these perfectly correct and verifiable facts, facts which are central to the discussion. This makes it clear that the voting public at HN has extremely poor judgment and should not be permitted a vote on comments of decent people.