The Riddle Riddle: Testing Flexible Reasoning in Large Language Models and Humans