In 1918 Cantelli asked whether there exists a non-constant positive measurable function f and two i.i.d. standard normal distrubted random variables X and Y such that the random variable X + f(Y) is also Gaussian. I present a paper of Victor Kleptsyn and Aline Kurtzmann from 2015 where they were able to give an affirmative answer. The core of their argument is the construction of a well-behaved non-constant stopping time T for a Brownian motion B such that the distribution of B at time T is the standard normal distribution. In fact, this stopping time T is a derivative of the classical Root-Solution to the Skorokhod Embedding Problem.