As model-based engineering expands from technical systems to socio-technical ecosystems, a central challenge emerges: can we model the humans within these systems with the same rigor and utility that we apply to the systems themselves? This talk will address recent progress in representing human cognition, behavior, and interaction, while highlighting the limitations that arise when attempting to capture the full diversity and unpredictability of human experience. Drawing on advances in large language models and multi-agent systems, the talk will explore what it would take to make human representations first-class model elements within the modeling ecosystem.