Memory Harnesses for Long-Running Research Agents — Stefania Druga, Sakana.ai
- Source
- AI Engineer
- Author
- AI Engineer
- Date
- context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
- agent harness — The scaffolding around a model that turns it into a working agent — the loop, the tools it can call, and the rules for when to stop.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
It gives a defensible answer to when an memory system earns its cost and which recall policy to reach for first.
“On a literature review task where every paper already fit inside the context window, adding a memory harness changed nothing: the same accuracy, at higher cost.”
AI Engineer
“Move to a long horizon task where the answer sits at step 124 and the question arrives at step 500, far outside the window, and the harness becomes the entire game.”
AI Engineer
“Her framing is that memory is a write, manage, read control loop wrapped around the model, not a database you attach to it.”
AI Engineer
“The oracle pointedly does not reach the ceiling, because giving a model the right memory does not make it use the right memory.”
AI Engineer
“Ranked recall was also cheaper, which is the line worth keeping: bad memory is expensive, since it burns tokens and sends the agent the wrong way.”
AI Engineer
videoAn AI Research Agent That Runs Your Experiments — Tim Sweeney, Weights & Biases
videoBeating RL With Reflection: GEPA and Optimize Anything — Lakshya A. Agrawal, GEPA
videoHow We Built an Agent That Improves Itself — Zubin Aysola, Weights & Biases
videoAutoresearch Made Our Models 3x Faster — Tejas Bhakta, Morph
Checking sign-in…
Loading comments…