Not every "tropical curve" (meaning a weighted balanced one-dimensional rational polyhedral complex) is the tropicalization of a curve in the torus, and the tropical inverse problem asks when this is the case. I will discuss variants of this question, and their relevance to birational geometry.