The document provides an overview of two approaches to enabling computers to understand the content of the World Wide Web: (1) the structuring solution, which aims to impose structure on web pages similar to databases, and (2) the learning solution, which aims to make computers smarter through machine learning. It then summarizes three examples of researchers pursuing the learning solution, including using naive Bayes and maximum entropy classifiers to categorize web pages. The researchers were able to achieve 60% accuracy for page classification using naive Bayes but saw improved performance from maximum entropy and adding additional heuristics.