{
  "date": "2026-10-02",
  "papers": [
    {
      "url": "https://arxiv.org/abs/2609.26891",
      "title": "Harness as a Language: A Minimalist Agent Framework With Maximal Expressivity",
      "authors": "Zhening Li, Joshua Liu, Mateja Vukelic, Nicole Shen, Supriya Lall, Amitayush Thakur, Alex Zhang, Omar Khattab, Jonathan Light, Armando Solar-Lezama",
      "date": "2026-09-22",
      "abstract": "Modern language-model agents are built around the agent loop. However, certain capabilities such as long-term memory and self-improvement currently require specialized systems beyond the agent loop itself. We built an LLM agent framework, JAZ, to explore the extent to which a minimal harness can accomplish tasks these specialized systems are built for. JAZ exposes a single LLM-based primitive invoke and provides a set of built-in hooks. On long-horizon workflows requiring recall far beyond the context window, JAZ invoke outperforms Letta (MemGPT) by 8% at half its cost on the recall-heavy portion of StuLife. On continual self-improvement, JAZ invoke outperforms ACE by 4% at a lower cost on AppWorld."
    },
    {
      "url": "https://arxiv.org/abs/2607.14713",
      "title": "Does Multi-Agent Debate Improve AI Feedback on Research Papers?",
      "authors": "Tomas Havranek, Zuzana Irsova",
      "date": "2026-07-16",
      "abstract": "Probably not, at least for meta-analyses in economics. In a pre-registered, identity-masked, within-paper experiment, the authors of 44 meta-analyses ranked three AI reports on their own paper by usefulness for improving it: a single pass by a frontier model against two multi-agent debate tools we built and expected to win. The authors preferred the single pass, by 0.66 rank points over mad-research and 0.57 over paper-workshop, though paper-workshop spent roughly thirty times the tokens. The reversal warns against substituting an AI judge for the author."
    },
    {
      "url": "https://arxiv.org/abs/2508.12752",
      "title": "Deep Research: A Survey of Autonomous Research Agents",
      "authors": "Wenlin Zhang, Xiaopeng Li, Yingyi Zhang, Pengyue Jia, Yichao Wang, Huifeng Guo, Yong Liu, Xiangyu Zhao",
      "date": "2025-08-18",
      "abstract": "The rapid advancement of large language models (LLMs) has driven the development of agentic systems capable of autonomously performing complex tasks. To overcome these limitations, the paradigm of deep research has been proposed, wherein agents actively engage in planning, retrieval, and synthesis to generate comprehensive and faithful analytical reports grounded in web-based evidence. In this survey, we provide a systematic overview of the deep research pipeline, which comprises four core stages: planning, question developing, web exploration, and report generation."
    },
    {
      "url": "https://arxiv.org/abs/2506.18096",
      "title": "Deep Research Agents: A Systematic Examination And Roadmap",
      "authors": "Yuxuan Huang, Yihang Chen, Haozheng Zhang, Kang Li, Huichi Zhou, Meng Fang, Linyi Yang, Xiaoguang Li, Lifeng Shang, Songcen Xu, Jianye Hao, Kun Shao, Jun Wang",
      "date": "2025-06-22",
      "abstract": "The rapid progress of Large Language Models (LLMs) has given rise to a new category of autonomous AI systems, referred to as Deep Research (DR) agents. These agents are designed to tackle complex, multi-turn informational research tasks by leveraging a combination of dynamic reasoning, adaptive long-horizon planning, multi-hop information retrieval, iterative tool use, and the generation of structured analytical reports. In this paper, we conduct a detailed analysis of the foundational technologies and architectural components that constitute Deep Research agents, contrasting API-based retrieval methods with browser-based exploration, and examining modular tool-use frameworks."
    },
    {
      "url": "https://arxiv.org/abs/2402.14207",
      "title": "Assisting in Writing Wikipedia-like Articles From Scratch with Large Language Models",
      "authors": "Yijia Shao, Yucheng Jiang, Theodore A. Kanell, Peter Xu, Omar Khattab, Monica S. Lam",
      "date": "2024-02-22",
      "abstract": "We study how to apply large language models to write grounded and organized long-form articles from scratch, with comparable breadth and depth to Wikipedia pages. We propose STORM, a writing system for the Synthesis of Topic Outlines through Retrieval and Multi-perspective Question Asking. STORM models the pre-writing stage by (1) discovering diverse perspectives in researching the given topic, (2) simulating conversations where writers carrying different perspectives pose questions to a topic expert grounded on trusted Internet sources, (3) curating the collected information to create an outline."
    }
  ],
  "repo": {
    "url": "https://github.com/stanford-oval/storm",
    "name": "stanford-oval/storm",
    "description": "STORM: Synthesis of Topic Outlines through Retrieval and Multi-perspective Question Asking",
    "graft_analysis_path": "profiles/r-and-d/raw_evidence/storm_graft_map.txt",
    "stats": "52 files, 669 symbols, 1602 edges, ~173090 tokens saved via graft"
  }
}