The design of fusion reactor cores has undergone significant evolution over the past decades, driven by advances in plasma physics understanding, engineering constraints, and the pursuit of reactor-relevant performance. Early core design efforts were primarily guided by empirical scaling laws and simplified assumptions on confinement and stability. However, with the emergence of large-scale tokamak experiments and integrated modeling capabilities, design criteria have progressively shifted toward physics-based and predictive frameworks that incorporate transport, stability, and edge-core coupling. In this work, the historical development of fusion core design is reviewed, including representative DEMO and power plant studies, with a focus on how key design criteria and requirements have evolved in response to advances in plasma physics and fusion technologies. These include, for example, the exploration of alternative plasma shaping concepts such as negative triangularity, as well as the emergence of high-magnetic-field designs enabled by high-temperature superconductors. It is further emphasized that plasma operation scenarios have co-evolved with these developments, transitioning from inductive, pulsed operation toward advanced and steady-state scenarios with increased non-inductive current drive and improved confinement regimes. The influence of these changes on the methodologies used to design plasma scenarios is examined, including the progression from zero-dimensional system codes to integrated, multi-physics simulations. In addition, the emerging role of artificial intelligence and machine learning techniques in reactor core design is addressed. These approaches are increasingly complementing conventional physics-based methods, offering new pathways to efficiently explore high-dimensional design spaces. Furthermore, a new figure of merit is proposed for evaluating reactor-relevant core plasma performance in present-day devices, aimed at bridging the gap between experimental capability and reactor design requirements. Finally, future directions for fusion core design are outlined, emphasizing the need for tighter integration between physics-based modeling, AI-assisted design frameworks, and control-oriented optimization to achieve high-performance, steady-state plasmas.