Design is unscientific in a specific sense: it makes no testable predictions and can accommodate any observation. The multiverse, whatever its problems, is at least a research program with mathematical content.

This is the fairest form of the unscientific charge. Design proponents reply that the design hypothesis does predict something, namely that the universe will look fine-tuned, and that the multiverse, in the versions relevant here, is also untestable in principle since other universes are causally disconnected. The honest conclusion is that neither design nor the multiverse is science in the ordinary sense; both are metaphysical explanations of a scientific fact. Design should not be singled out as the unscientific option, and the accurate statement is that both are outside science, not that both are inside it.