RNA viruses are frequently tolerant to high levels of mutagenesis. In contrast, DNA viruses are less errorprone and coevolve along with their specific hosts over long time periods. Although both strategies have been successful, the "RNA-strategy" (directly linked to the pathogenic potential of these agents) most often generates novelty (new variants, new strains, and even new viral pathogens). For several decades, intra-host virus evolution has been considered to be a speculative field, far from the main issues of clinical virology. This concept is now changed, due to the evidence that RNA virus evolution is intimately linked to failures in viral disease control and prevention. Antiviral strategies using single and fixed elements (i.e. treatments using one antiviral compound, immunizations using a single recombinant protein) have been unable to control highly dynamic quasispecies, such as human immunodeficiency virus type I (HIV-1) and hepatitis C virus (HCV). The development of combinatorial treatments in HIV-1 infection and the recognition that vaccines should be multivalent are important steps in adapting disease control strategies to the complexity of viral populations. The present report summarizes the strategies adopted to address HIV-1 evolution and its phenotypic consequences, including changes in susceptibility to antiviral compounds, viral fitness, and pathogenic potential. In particular, it is highlighted that sequence-function analyses of the intra-host HIV-I evolution, including studies of viral fitness, have opened up new perspectives not only to studying the pathogenic mechanisms and the virus-host relationships, but also to designing new strategies for monitoring antiviral therapies.