AnalysisAI AgentsAugust 10, 2026

Stefania Druga on memory harnesses for long-running research agents

On a literature-review task where every paper fit in the context window, adding a memory harness produced identical accuracy at higher cost — a negative result Druga says marks the boundary of where memory helps, versus long-horizon tasks.

Featured · Stefania Druga

1 source

Daily brief

Get tomorrow's AI brief in your inbox

More stories today

Open the live feed
Stefania Druga on memory harnesses for long-running research agents — AIBriefs