The human flu virus undergoes rapid evolution, which is driven by 
interactions with its host immune system. We describe the evolution 
of viral genomes in terms of a non-equilibrium statistical mechanics 
model. This model successfully predicts the evolution of influenza 
one year into the future. Thus, evolutionary analysis transcends its 
traditional role of reconstructing past history. This has important 
consequences for public health: evolutionary predictions can inform 
the selection of influenza vaccine strains. We discuss the conditions 
of predictability and its links to cooperative phenomena and 
universality. These links highlight the key role of statistical 
mechanics in making evolutionary biology a predictive science.