We are presenting Alice Architecture and the ±0 Theory—an attempt at an Autonomous Thinking AGI model. This model integrates the SDE-driven ±0 Theory (Happiness/Unhappiness Model) and a unique HALM (Hierarchical Abstracted Memory). It aims for a truly internally-driven, autonomous AI that pursues homeostasis through self-negation-driven mechanisms. All mathematical formulas and Python code are publicly available in the GitHub repository. We encourage you to review the work. We particularly welcome constructive discussion and validation on the following points: Validation Proposal: Observing changes in the LLM's behavior when existing LLM APIs are integrated with the ±0 Theory / Alice Theory Python code in the same file. Core Technology: The maximization objective of the Total Wellbeing Scalar V(t) and the mathematical/implementation logic of Affective TDL (Affective Temporal Difference Learning). We are currently developing the Python code that connects existing LLM frameworks with the ±0 Theory and Alice Architecture. We look forward to open discussion from everyone.