Its a requirement of a physical theory to make predictions. I can't think of a 'hot topic' in theoretical physics which doesn't make predictions. Testable predictions is another matter. By testable I mean having the capability to experimentally test.