The author reflects on the evolution of technology over time, noting that breakthroughs often propagate slowly through engineering practice. They highlight the shift from traditional central switching offices to distributed computing and the dominance of RAM in modern systems. The author argues that this has led to a new way of thinking about designing and building applications, where data is stored entirely in RAM and the focus is on memory locality rather than random access. They suggest that this approach simplifies the architecture of distributed databases, but notes that there are still fundamental conflicts between optimizing for reads and writes, and that breakthroughs will continue to be needed to address these challenges.