I have written a white paper on how stateful language model needs memory governance to enhance durability and retain relevance of the persistent memory. The target architecture is Titans.
The paper is published to ResearchGate - https://www.researchgate.net/publication/400442850_Two_Brains_in_a_Skull_Governance_Architecture_for_Stateful_Language_Models?channel=doi&linkId=69835c8a12f837212a177e33&showFulltext=true