Instead of saying "It doesn't predict anything" and then proceeding to ignore it, a more productive attitude would be to ask questions like: "So can we think of some kind of consequences that different foundational pictures point to?"

Reworded: instead of saying "it doesn't predict anything", [we should] ask "does it predict anything?"

In other words: don't say a theory doesn't predict anything unless you have an actual argument that shows that it doesn't predict anything.