Swarm intelligence (SI) and SI-based algorithms have become popular and useful in almost all areas of sciences and engineering. Significant developments have been made in recent years. This paper provides a short but timely analysis about SI algorithms and their links with self-organization. Emphasis has been on the present developments by analyzing the main characteristics and properties of algorithms, while future directions are pointed out by highlighting key challenges and their implications.