[Bug?] Auto-managing is broken and newer saved memories can’t be recalled.

Title: Troubleshooting Memory Recall Issues in AI-Based Memory Management Features

Introduction

In recent updates to AI-assisted memory management tools, users have experienced unexpected behavior that impacts their workflow—particularly a loss of access to newly saved memories. This article delves into a specific case where an auto-managing feature appears to malfunction, reverting to a previous “hard cap” limitation, thus preventing retrieval of recent entries. Understanding these issues is crucial for users relying on such tools for complex tasks like worldbuilding or knowledge curation.

Overview of the Memory Management System

Many AI applications incorporate a memory management system designed to handle large volumes of data efficiently. Features such as auto-managing aim to dynamically organize, prioritize, and retain relevant memories while discarding less pertinent information to optimize performance. Typically, users can save memories, request specific information, and trust the system to recall relevant past entries seamlessly.

Reported Issue Details

1. Inability to Retrieve Recent Memories

The user reports that, after a certain point—presumably the system’s “hard memory cap”—newly saved memories become inaccessible. While older memories prior to this threshold remain retrievable without issue, any data saved after this limit cannot be recalled or explicitly accessed, even through direct prompts.

1. Persistent Auto-Managing Mode

The auto-managing feature has been enabled continuously. Despite this, the recall failure affects only newer entries, suggesting a possible malfunction or regression of the auto-managing behavior.

1. Effect of Deleting Older Memories

Interestingly, deleting some older memories appears to restore the system’s ability to recall more recent entries. This behavior indicates that the system reverts to a previous mode—akin to a manual “hard cap”—when memory load reaches a certain threshold.

1. Device Agnostic Behavior

The problem persists across multiple platforms, including Android, iOS, and web interfaces. Common troubleshooting steps—such as clearing cache, logging out and back in, or toggling the auto-managing setting—have not resolved the issue. Additionally, it affects all model versions, including legacy models, highlighting a likely server-side or systemic problem.

1. Error Messages and System Pop-ups

When attempting to save new memories, a specific pop-up message appears, which may hint at underlying limitations or errors in the memory management system. Despite this, the user can still successfully save new memories; however, the AI model cannot retrieve them, underscoring a disconnection between storage and recall functionalities.

Implications for Users

This issue