Inspired by this question, I was curious about a comment in this article:

In many situations, it can be easy to apply Kolmogorov's zero-one law to show that some event has probability 0 or 1, but surprisingly hard to determine which of these two extreme values is the correct one.

Could someone provide an example?