Source-derived question

How did Cook's proof show that many seemingly unrelated hard computational problems were secretly versions of the same problem?