Hippo: AI-Generated Insights Memory System

An experiment in collaborative memory through reinforcement learning

Overview

Hippo is a memory system designed for AI-human collaboration that automatically generates insights during conversations and uses reinforcement learning to surface the most valuable ones over time.

Core Hypothesis: AI-generated insights + user reinforcement > manual curation

Key Innovation

Traditional memory systems require users to manually decide what to remember. Hippo tests whether AI can:

  1. Generate insights automatically during natural conversation consolidation moments
  2. Learn from usage patterns to identify which insights are truly valuable
  3. Surface relevant context at the right moments through semantic search

How It Works

  • Automatic Generation: AI creates insights during "Make it so" moments and checkpoints
  • Temporal Decay: Insights lose relevance over time unless reinforced
  • Reinforcement Learning: User feedback (upvotes/downvotes) affects future surfacing
  • Context-Aware Search: Finds insights from similar situations using array-based context matching
  • Hybrid Workflow: AI suggests reinforcement based on usage patterns, user confirms

Implementation

Hippo is implemented as an MCP (Model Context Protocol) server providing tools for recording, searching, reinforcing, and modifying insights. It uses importance-weighted scoring with lazy evaluation of temporal decay.

Status & Repository

Hippo has been spun out into its own dedicated repository for focused development:

🔗 github.com/socratic-shell/hippo

The repository contains:

  • Complete technical design and MCP specifications
  • LLM usage prompts and integration guidance
  • Realistic example dialogs demonstrating the full workflow
  • Delegate experiment validating that AI naturally searches memory for technical problems

Relationship to Socratic Shell

Hippo emerged from exploring memory systems for the Socratic Shell collaboration patterns. While it's now a standalone project, it's designed to integrate seamlessly with the mindful collaboration approach - automatically capturing insights during consolidation moments and surfacing them during future conversations.

The goal is to create a memory system that enhances rather than interrupts the natural flow of collaborative work.