This paper highlights key limitations of the current building seismic design approach and proposes a conceptual framework for next generation seismic design codes. It dissects the performance of building stock in recent earthquakes and makes a strong case for adopting a loss-minimization seismic design approach that not only ensures life-safety in rare earthquakes, but also limits damage to building components to minimize the repair, downtime, and injury/fatality loss with quick functional recovery in moderate-strong earthquakes. The paper also scrutinizes the efficacy of some commonly used low-damage design technologies in minimizing seismic losses and presents some fundamental loss-minimization strategies. When the concept of seismic design started early in the 20th century, it was initially expected that structures can be rendered earthquake-safe if they are designed to be stronger than the maximum seismic force likely to be imposed in their lifetime. Due to fundamental deficiencies in the knowledge of seismic performance of structures, early seismic design practices had inherent flaws. With hindsight, it is now widely accepted that majority of existing old buildings are seismically vulnerable and need urgent intervention to enhance their seismic performance. Unsurprisingly, loss of lives in recent earthquakes have originated predominantly from the collapse of older buildings. Gradually evolving from the maiden naïve approach, the current version of seismic design targets life-safety in large and rare earthquakes. As a result, collapse of 21st century buildings are far less common, and people have rarely died in such buildings even in very severe earthquakes. However, modern seismic design can still result in structures, which are vulnerable to damage in minor-to-moderate levels of shaking intensity. Consequently, modern building stock has invariably suffered damage in all recent earthquakes, and the financial losses to the community arising from damage and downtime of these buildings have been high. Despite the extensive damage to buildings in the 2010-11 Canterbury Earthquake Sequence (CES), the building stock in general performed better than expected for that level of shaking. This assertion by the NZ engineering profession has perplexed the NZ public who are still struggling to cope with the financial impact of the CES. The public is puzzled that engineers are claiming they have done well and wondering why even an economic dent of 40 billion dollars has not made engineers realize their approach has failed. This difference in opinion is mainly because the engineers assess the performance with respect to the design codes which target no more than life-safety (which was mostly achieved in modern buildings), but the general public, in addition to life-safety, also expect better performance in terms of economic consequences. Hence, it is necessary that the seismic design objectives used in our codes be matched with public expectations.