In the quest to understand the cosmos, we must remain humble, for every answer we find leads to a multitude of new questions.
The concept of probably approximately correct (PAC) learning has been foundational in computational learning theory.