LSD is a new paradigm in search that uses latent semantic distribution to place emphasis on the meanings of words through a distribution and arrangement of their frequency of use to potentially get better search results. By using LSD's latent eigenvector construct to semantically optimize content and consider possible link building, it aims to improve rankings by accounting for both content and links.