{"slug": "scaling-laws-for-agent-harnesses-via-effective-feedback-compute", "title": "Scaling Laws for Agent Harnesses via Effective Feedback Compute", "summary": "Researchers introduced Effective Feedback Compute (EFC), a new scaling coordinate that measures only informative, non-redundant feedback in AI agent harnesses, outperforming raw-compute metrics in predicting failure rates across synthetic and real benchmark tasks. In controlled experiments, EFC-based coordinates achieved R² values up to 0.99, while matched-budget interventions improved success rates from 0.27 to 0.90 without increasing raw cost or tool calls. The findings indicate that scaling agent performance depends more on converting raw computation into durable, task-sufficient feedback than on total computational expenditure.", "body_md": "# Computer Science > Computation and Language\n\n[Submitted on 28 May 2026]\n\n# Title:Scaling Laws for Agent Harnesses via Effective Feedback Compute\n\n[View PDF](/pdf/2605.29682)\n\n[HTML (experimental)](https://arxiv.org/html/2605.29682v1)\n\nAbstract:Agent harnesses increasingly determine the performance of language-model systems by deciding how models call tools, receive feedback, verify intermediate states, store memory, and revise solutions. Yet current test-time scaling analyses often parameterize this process by raw expenditure -- tokens, tool calls, operations, wall time, or cost -- which does not distinguish useful feedback from redundant or unstable interaction. We introduce \\emph{Effective Feedback Compute} (EFC), a trace-level scaling coordinate that credits feedback only when it is informative, valid, non-redundant, and retained for subsequent decisions, and we normalize it by task demand when comparing tasks with different feedback requirements. Across synthetic controllable tasks, executable code tasks, real benchmark traces, held-out splits, and a prospective validation batch, EFC-based coordinates consistently predict failure rates better than raw-compute baselines and a strong multivariate SAS baseline. In controlled scaling, raw tokens and tool calls explain limited variation ($R^2=0.33$ and $0.42$), SAS reaches $0.88$, while Oracle-EFC and Estimated-EFC reach $0.94$ and Oracle-EFC/$D_{\\mathrm{task}}$ reaches $0.99$. Matched-budget interventions show that improving feedback quality raises success from $0.27$ to $0.90$ while raw cost and tool calls are fixed. On mixed real traces, NRS-EFC/$D_{\\mathrm{task}}$ reaches $R^2=0.92$ while raw compute has near-zero or negative fit, and it remains the best predictor in a prospective holdout ($R^2=0.85$). These results suggest that harness scaling is governed less by how much computation is spent than by how efficiently raw budget is converted into durable, task-sufficient feedback.\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer\n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers\n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps\n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations\n\n*(*[What are Smart Citations?](https://www.scite.ai/))# Code, Data and Media Associated with this Article\n\nalphaXiv\n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers\n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub\n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub\n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face\n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast\n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower\n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender\n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\n\nHave an idea for a project that will add value for arXiv's community? [ Learn more about arXivLabs](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/scaling-laws-for-agent-harnesses-via-effective-feedback-compute", "canonical_source": "https://arxiv.org/abs/2605.29682", "published_at": "2026-05-30 03:34:45+00:00", "updated_at": "2026-05-30 03:45:41.259950+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "large-language-models", "ai-agents", "ai-research"], "entities": ["Effective Feedback Compute", "EFC", "SAS"], "alternates": {"html": "https://wpnews.pro/news/scaling-laws-for-agent-harnesses-via-effective-feedback-compute", "markdown": "https://wpnews.pro/news/scaling-laws-for-agent-harnesses-via-effective-feedback-compute.md", "text": "https://wpnews.pro/news/scaling-laws-for-agent-harnesses-via-effective-feedback-compute.txt", "jsonld": "https://wpnews.pro/news/scaling-laws-for-agent-harnesses-via-effective-feedback-compute.jsonld"}}