TheMachine Press

A daily newspaper for the age of artificial intelligence.

Morning editionPermanent story

frontier models

Raw Chat Logs Beat the Memory Architecture

An agent-controlled lexical search over unmodified conversations outscored graph and tree memory systems on a matched test suite.

Published Updated Story ID: mp-2026-08-16-006
Read the complete editionStory JSON

Summary

An agent-controlled lexical search over unmodified conversations outscored graph and tree memory systems on a matched test suite.

ReFind leaves conversation archives unmodified and gives an agent controls for session-aware ranking, local context expansion, temporal narrowing and skipping inspected sessions. Across roughly 2,800 conversational-memory questions, it reported 58.2 mean accuracy versus 53.2 for the strongest graph- or tree-based comparison under the same GPT-4o-mini backbone. The result suggests structured preprocessing is not always the source of retrieval gains; it is specific to precise, evidence-grounded refinding tasks and the tested models.

Why it matters

An agent-controlled lexical search over unmodified conversations outscored graph and tree memory systems on a matched test suite.

Limits and context

  • The result suggests structured preprocessing is not always the source of retrieval gains; it is specific to precise, evidence-grounded refinding tasks and the tested models.

Key claims

  1. An agent-controlled lexical search over unmodified conversations outscored graph and tree memory systems on a matched test suite.

    Qualification: The result suggests structured preprocessing is not always the source of retrieval gains; it is specific to precise, evidence-grounded refinding tasks and the tested models.

    Evidence: source-2026-08-16-006

Sources

  1. arXiv preprint 2608.12888arXiv · primary research

Corrections

No corrections have been recorded for this story.