The Last AI Built by Humans: Toward Genuine Recursive Self-Improvement

  • 类型:arxiv
  • 标识:2609.11873
  • 链接:https://arxiv.org/abs/2609.11873
  • 主分类:agent
  • 形态:method
  • 被引:8
  • 被引来源:Semantic Scholar
  • S2被引:8
  • OpenAlex被引:0
  • 影响力被引:2
  • TLDR:This work uses the Headroom-Closed Index (HCI) to reveal the problems of existing LLMs, and introduces the RSI concept and its development roadmap: from improvement-execution autonomy, improvement-strategy autonomy, experience-acquisition autonomy, and environment-adaptation autonomy, to recursive meta-improvement.
  • OpenAlex ID:W7212220037
  • OpenAlex DOI:10.48550/arxiv.2609.11873
  • DOI:10.48550/arxiv.2609.11873
  • DOI来源:OpenAlex
  • 开放获取:green
  • 开放获取链接:https://arxiv.org/pdf/2609.11873
  • OpenAlex更新:2026-09-17
  • 待LLM分类:否
  • 成熟度:research
  • 场景:recursive self-improvement、LLM evaluation、autonomous learning
  • 标题中文:人类构建的最后一个AI:迈向真正的递归自我改进
  • TLDR中文:该工作利用 Headroom-Closed Index(HCI)揭示现有 LLM 的问题,并提出 RSI 概念及其发展路线图:从改进执行自主性、改进策略自主性、经验获取自主性、环境适应自主性,到递归元改进。
  • 来源文件:
  • /inbox/tom/_candidates/2026-09-16-agent-rag-longcontext-candidates.json
  • [OpenAlex backfill]
  • [S2 enrich]