This talk examines what it would mean to build world models that account for the full range of human sensory experience, not just vision, but touch, smell, sound, and memory. Understanding how people actually experience space, through all their senses and through the lens of prior experience, changes how those environments should be constructed, captured, and modeled. Drawing on recent research at the intersection of spatial AI, XR, and cognitive science, the talk explores how multisensory data can be captured, structured, and used to create virtual environments.