A good learning algorithm should be able to generalize from a limited set of examples to unseen data.
Remaining quiet is what is called wisdom-insight.