I would edit my last line to say: weird models of numbers that don't match how we think "halts in finite steps" usually works.