The fields of VLSI/ASIC manufacturing, computer architecture, and software engineering are in an intertwined state of flux. Fundamental technology trends including power density limits, voltage and frequency scaling, manufacturing variation, resilience to defects, transistor aging effects, and multi-die integration concerns are forcing hardware designers and architects to rethink how they implement computation using silicon. If left unchecked, these limitations threaten an end to the exponential performance and energy scaling afforded by Moore's law. This talk discusses three important general targets for the next 5-10 years of research in computing: energy, reliability, and programmability. We discuss these general targets in the context of a few existing works, as well as long-term research goals.