I will talk about the relationship between computability theory, the complexity of sets of reals, and set-theoretical axioms. Using all these, I will prove a ZFC consistency result in fractal geometry, which shows that the Hausdorff dimension of “simple" sets can behave badly under projections.