{
  "id": 14385237,
  "title": "SEAR: Self-Evolving Agentic Robotics",
  "url": "https://urgent.news/2026/10/13/sear-self-evolving-agentic-robotics",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-13T17:40:20.000Z",
  "source": {
    "name": "arXiv cs.AI",
    "slug": "arxiv-cs-ai",
    "url": "https://arxiv.org/abs/2610.18792v1"
  },
  "original_language": "en",
  "account": null,
  "summary": "Frontier language-model agents can now improve robots on their own: they write control code, test it, and keep what works as skills and memories for the next task. We introduce SEAR (Self-Evolving Agentic Robotics), a framework and benchmark built around two questions. Is the agent really learning, or just trying more often? And does what it learns still help on new tasks, in a new simulator, and…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}