Suppose there are 1000 events and 500 will have outcome A and 500 will have outcome B. If you predict a 50% chance of A for every event you'll be perfectly calibrated. On the other hand, if you predict a 90% chance of a certain outcome and you're right for 800 events, you're not perfectly calibrated but you have a lower Brier score (lower is better).
Understanding Time and Space by Steven E Landsburg (subtitle: An Invitation to the Theory of Relativity for Anyone Who is Now, or Has Ever Been, an Inquisitive High School Student)
There was already an infinite amount of noise on the internet from humans. It was called "information overload". But just because it's out there doesn't mean you have to see it.