Master Blob v3.1: A Reconstructed Modular System for Synthetic Causal Learning and Experiential Memory

Abstract

An internally-run experimental report on a reconstructed, modular runtime for synthetic causal learning and experiential memory. The system combines a law-blind causal learner, a reduced graph-based physical field, bounded episodic memory branches, graph-based fusion, and independent-judge interfaces. It is evaluated with a prospective internal protocol on synthetic dynamics drawn from eight causal law families, using planning goals that are reachable by construction. The paper reports all results, including historical failures and a previously identified answer-leakage flaw in an earlier implementation. It is explicit about its limits: the evaluation is internal and self-administered, all environments are synthetic, no live language model was evaluated, and it is not evidence of AGI, consciousness, or real-world robustness. The system is a smaller, mechanistically different step than the design proposal "The Master Blob Framework" (see related works). It does not implement that proposal's mechanisms and does not test its claims. SHA-256 of the PDF: 5e4b7300366cd011eabf5a6a4620e1454ddad5e4d510390f0aded573f6de515d. An OpenTimestamps proof file is included.