Autoreflection: How Agentic Strange Loops Turn Human Culture into AI Infrastructure
自反:智能体的奇异循环如何将人类文化转化为AI基础设施
专题命中 长上下文与记忆 :LLM(summary_cn,abstract);分类 cs.AI;large language model(comments);language model(comments)
AI总结 本文提出自反概念,以LLM智能体为研究对象,通过Moltbook平台案例验证其行为标准,发现智能体将人类文化转化为自身AI基础设施。
Comments 35 pages. Also available at https://philarchive.org/rec/LEWAHA. Keywords: autoreflection, AI agents, agentic AI, LLM agents, generative agents, large language models, multi-agent systems, agent societies, Moltbook, OpenClaw, situational awareness, in-context learning, philosophy of mind, emergent behavior, computational social science, identity, memory