Give an example of when you used prototyping or rapid iteration to validate a risky assumption. What was the experiment, how did you measure success, and what did you learn?