LinkedIn Paper Tests Memory Portability Across Models: Notes Swing ±10-13 Points, Knowledge Graphs Barely Move
dair_ai · x · 2026-09-07
A new LinkedIn paper tackles an overlooked question: can an agent's memory survive swapping the underlying model? The team stored the same agent history four ways—verbatim long context, chunked for retrieval, model-written notes, and a fixed-schema knowledge graph—then swapped the model doing the reading.
Results diverged sharply:
- Fixed-schema knowledge graphs barely moved: just 0.0004 accuracy change across a writer swap
- Model-written notes were volatile: accuracy shifted +9.91 or -13.28 points depending on swap direction
dair-ai calls memory one of the hardest things to get right when building agents and worth deliberate optimization. As users increasingly move between models, agents, and providers, memory portability is becoming a core engineering concern—especially for anyone who upgraded an agent's model only to watch it forget.
More from coding & agent
- Solving agentic amnesia: a file-system state machine for Claude Code — SnooComics4579 · 2026-09-08
- Stanford releases free full course on self-improving AI agents — Saboo_Shubham_ · 2026-09-08
- Open-sourced clay-style 3D kids game built with Claude, method fully documented — dotey · 2026-09-08
- Steve Yegge: Fable 5.1 rewrote 30% of Wheelhouse factory after Fable 5 trainwreck — Steve_Yegge · 2026-09-08
- Kwipu: an open-source local Graph RAG engine that turns Markdown notes into a knowledge graph — tom_doerr · 2026-09-08
- hip-agent: a 200-line agent harness designed to fit entirely in the prompt — cephaloform · 2026-09-08