The approach toward treatment of Alzheimer's disease has evolved out of a large body of research, resulting in a common trial paradigm directed toward establishing efficacy and safety of putative treatments for dementia. To achieve the ultimate goal of preventing dementia, the future will build on this model but will need to incorporate new paradigms to more efficiently detect meaningful change over time. Among new innovations needed is the capability for assessing trajectories of change over many years. Intrinsic to this success is the ability to capture increasing variability as well as subject heterogeneity inherent in longitudinal aging cohorts. New methods of ubiquitous embedded home sensing and computing afford the opportunity to assess volunteers not only during long periods of time but with greater frequency and convenience. Clinical studies identifying the optimal ways of using these methods are needed.