{
  "title": "Daily Research Brief 2026-07-21",
  "url": "/en/posts/research-brief-2026-07-21/",
  "permalink": "https://hackcv.com/en/posts/research-brief-2026-07-21/",
  "date": "2026-07-21",
  "lastmod": "2026-07-21",
  "author": "",
  "description": "Daily research brief — AI / LLM / Agent / Computer Vision / Audio-Video / Engineering",
  "categories": ["Research Brief"],
  "tags": ["AI","LLM","Agent","Computer Vision","Audio-Video","Engineering","Daily Brief"],
  "cover": "https://picsum.photos/seed/daily-research-brief-2026-07-21/1200/675",
  "readingTime": 1,
  "wordCount": 222,
  "content": "\u003ch1 id=\"daily-research-brief-2026-07-21\"\u003eDaily Research Brief 2026-07-21\u003c/h1\u003e\n\u003cp\u003e📊 Token usage: estimated from retrieval and writing scale.\u003c/p\u003e\n\u003cp\u003eCovers the latest AI research, open source and industry moves, updated daily.\u003c/p\u003e\n\u003chr\u003e\n\u003ch2 id=\"editors-note\"\u003eEditor\u0026rsquo;s Note\u003c/h2\u003e\n\u003cp\u003eTwo threads worth practitioners\u0026rsquo; attention today. First, \u0026lsquo;open-weight models enter a dense payoff period\u0026rsquo; — DeepSeek V4 officially GA\u0026rsquo;d open-source (1.6T MoE, fully MIT-licensed), Qwen3.8 went open (2.4T), and China\u0026rsquo;s Meteorological Administration open-sourced a hundred-billion-parameter weather model tied to global public early warning; combined with Kimi K3 weights dropping 7/27, the open camp is advancing on parameter scale, domain specialization and usability simultaneously — the \u0026lsquo;open = catching up\u0026rsquo; narrative has been substantively overturned this week. Second, agents are descending from the chat box into infrastructure primitives.\u003c/p\u003e\n\u003ch2 id=\"1-latest-arxiv-papers\"\u003e1. Latest arXiv Papers\u003c/h2\u003e\n\u003col\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eReward-Driven LLM Agent Workflows: Synthesizing POMDP Routing and Self-Correction for Autonomous Decision-Making\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2607.17038\"\u003ehttps://arxiv.org/abs/2607.17038\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eDynamic Agent Skills: A Lifecycle Survey and Taxonomy of Evolving Skill Libraries\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2607.10113\"\u003ehttps://arxiv.org/abs/2607.10113\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eAgentAbstain: Do LLM Agents Know When Not to Act?\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2607.10059\"\u003ehttps://arxiv.org/abs/2607.10059\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eIsolation as a First-Class Principle for LLM-Agent System Safety: Concepts, Taxonomy, Challenges and Future Directions\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2607.12406\"\u003ehttps://arxiv.org/abs/2607.12406\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eCritic Experience Bank: Self-Evolving Step-Level Confidence Estimation for LLM Agents\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2607.12397\"\u003ehttps://arxiv.org/abs/2607.12397\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003ePM-Bench: Evaluating Prospective Memory in LLM Agents\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2607.12385\"\u003ehttps://arxiv.org/abs/2607.12385\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eOn-Device Deep Research at 4B: Exposure Bounds Faithfulness, Retrieval Bounds Coverage\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2607.12257\"\u003ehttps://arxiv.org/abs/2607.12257\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eA Threshold Exceedance Framework for CBRN Uplift Evaluation in Frontier Language Models\u003c/strong\u003e — \u003ca href=\"https://arxiv.org/abs/2607.12200\"\u003ehttps://arxiv.org/abs/2607.12200\u003c/a\u003e\u003c/p\u003e\n\u003c/li\u003e\n\u003c/ol\u003e\n",
  "summary": "Daily Research Brief 2026-07-21 📊 Token usage: estimated from retrieval and writing scale.\nCovers the latest AI research, open source and industry moves, updated daily.\nEditor\u0026rsquo;s Note Two threads worth practitioners\u0026rsquo; attention today. First, \u0026lsquo;open-weight models enter a dense payoff period\u0026rsquo; — DeepSeek V4 officially GA\u0026rsquo;d open-source (1.6T MoE, fully MIT-licensed), Qwen3.8 went open (2.4T), and China\u0026rsquo;s Meteorological Administration open-sourced a hundred-billion-parameter weather model tied to global public early warning; combined with Kimi K3 weights dropping 7/27, the open camp is advancing on parameter scale, domain specialization and usability simultaneously — the \u0026lsquo;open = catching up\u0026rsquo; narrative has been substantively overturned this week. Second, agents are descending from the chat box into infrastructure primitives.\n"
}