With the appearance of 2D graphical user interfaces in the 1980s, users began to carry out most operations on 2D icon-based interfaces instead of using line-based terminals. With the emergence of smartphones in the 2010s, the notion of portable 2D graphical user interfaces was born, and by today, users accessing digital services are no longer tethered to a single location. All of these advances have led to immense changes in our conceptualization of digital information systems, the effects of which are difficult to overestimate. Recent developments in virtual and augmented reality (VR/AR), as well as in Internet of Things (IoT) and artificial intelligence (AI) are poised to lead to the next major breakthrough in this series of cognitive expansions, bringing to the forefront portable, context-aware spatial interfaces. A consequence of these developments is that users are expecting to be able to access a growing multitude and variety of digital content in ways that are increasingly contextualized and personalized, i.e., relevant to the time, location and topic from the perspective of the users' personal history. In this paper, we propose an adaptive content labeling and storage model that is suitable to these challenges. Our model, called the Graph-Indexed Tensor Store (GITS) has the benefits of being adaptive and personalized, while also allowing for content retrieval to be carried out based on associative search operations. In a preliminary analysis of the model, we address the challenges of what we refer to as syntactic, semantic and pragmatic saturation, and provide ways to quantity and further explore their effects.