surely there must be some numbers wrong here? How can the total energy supply be only just barely less than the entire amount of energy applied to the planet by solar radiation?

Is it suggesting that covering the entire surface of earth with 100% efficient solar panels would just barely cover 2017's power supply?

Edit: ah I see, world energy is in terawatt hours for an entire year, wheras planetary solar energy is terawatts in a single hour

The bigger question is whether we will heat up the surface top much, given that the energy would be produced by a heat engine.

(I suppose you could build a heat engine by using the temperature differentials of different underground layers, although I've never seen that proposed)

I can be wrong but that number sounds so low that it might mean the energy at the literal surface, in the same vein as how much solar heat energy reaches the planet’s surface.

I assume that the number is vastly greater if you measure the heat 1km below sea level for example.

The only mechanism we would have for significantly increasing the cooling of the Earth's core would involve heading the atmosphere well past the habitable band.

Ultimately if heat isn't radiating from the atmosphere, it isn't leaving the planet.