The general shape of geothermal, where you're up here where it's cool, and you push a fluid down there where it's hot, and you bring the heat back up here to do work... Would a good enough insulator make it thermodynamically feasible to do it from an airship in the Venusian atmosphere?

Like, clearly Earth's crust supports temperature gradients which are steep enough to make the juice worth the squeeze. Presumably you're not going to find gradients so steep in the case of a runaway greenhouse effect, but whatever gradients you do find, are there fundamental limits preventing them from being steep enough?