Google paper: Self-Improvement Memory boosts continual recursive self-improvement in agent harnesses
omarsar0 · x · 2026-11-17
Google researchers propose Continual Recursive Self-Improvement (CRSI) and Self-Improvement Memory (SIM), which groups past harness edits by problem type and stores rationale, predicted/observed outcomes, and evaluation context.
- With SIM, 9 of 10 continued coding runs improved, averaging +2.36 points over harness-only continuation (p=0.003).
- Cross-model transfer is fragile: an Opus-evolved harness started at 68.41 on Gemini, below the plain seed harness's 73.89 — SIM helps bridge the gap.
- The paper underscores the importance of traces and a functional, adaptive, structured persistent memory in agent harnesses.
More from coding & agent
- Codex lacks visual indicator for multi-day agent runs, user resorts to renaming tasks — chrisalbon · 2026-11-17
- Poolside's 'everything collapses into the model' thesis shapes new agent runtime Rex — sull · 2026-11-17
- FrontierCS 2 benchmark challenges AI agents to beat published CS research blind — arena · 2026-11-17
- VPS MCP Server lets AI models manage Docker containers via Portainer API — modelcontextprotocol · 2026-11-17
- Qualcomm demos an on-device decision model routing AI requests in 8.6ms on NPU — carrycooldude · 2026-11-17
- User's open letter to OpenAI: Dot agent degraded after a week, switching to Claude Duet — Danny7Guns · 2026-11-17