Understanding material failure, from macroscopic predictions of structural disintegration to microscopic framework of damage nucleation and propagation, remains a central challenge in a wide range of disciplines. A. A. Griffith, exactly a hundred years ago, first proposed an energy-balance criterion for understanding damage propagation in brittle materials. The field has progressed considerably since then. The current challenges and techniques, spanning from designing self-healing materials to the use of statistical learning in failure predictions, call for new insights from experimental observations and theoretical studies, much of which has remained open questions.