This paper proposes an architectural model of subjectivity in which consciousness and will are treated as emergent properties of a system combining a living core with a set of conflicting drives, embodiment, stochasticity and a predictive “Big Model” of the world and self. Instead of searching for a “magic step” from neural or computational processes to qualia, the focus is shifted towards the conditions under which an agent develops an internal field of significance, existential stakes and narrative. The Big Model and its prediction engine are interpreted as an augmented‑reality (AR) interface that the living core “puts on” the world: this interface does not generate drives, but recodes the environment into the coordinates of these drives, highlighting what is edible, dangerous, resourceful or informationally valuable for the agent. The model introduces the notions of a living core, multiple drives with dynamic weights, chaos exposure and right to error, existential risk and “existential equations” of morality, and formulates conditions under which subjectivity can arise and manifest itself in behavior. The proposed framework is then compared to predictive processing and Active Inference, model‑based reinforcement learning, and leading theories of consciousness such as Integrated Information Theory and Global Workspace Theory. Finally, the model is related to Searle’s Chinese Room argument, large language models and Chalmers’ hard problem of consciousness, suggesting a reframing of the hard problem in terms of specific combinations of a living core and interface architecture (Chalmers, 1995) that give rise to a field of experience for the agent.
Roman Mamedov (Fri,) studied this question.