🤖 AI Agent 研究Research
【arXiv】代理社会是人工智能代理的集合，它们代表目标可能仅部分一致的不同委托人，跨越信任边界进行自主协调。我们通过实验表明，在代理社会中，甚至诚实，有能力
【arXiv】An agentic society is a collection of AI agents that coordinate autonomously across trust boundaries, on behalf of different principals whose objectives may only partially align. We show experimentally that in agentic societies even honest, competent
【arXiv】我们推出并发布ScienceBuddy ，这是一个交互式科学研究工作空间，将不断改进的科学代理引入研究人员的日常工作流程。ScienceBuddy支持研究人员在执行科学任务的同时
【arXiv】We introduce and release ScienceBuddy, an interactive scientific research workspace that brings continually improving scientific agents into researchers' everyday workflows. ScienceBuddy supports researchers in carrying out scientific tasks while tra
【arXiv】当大型语言模型的事实支持较弱时，它们可以产生流畅的答案。本文介绍了Chain-of-Self-Questioning （ CoSQ ） ，这是一个仅提示性的框架，使答案承诺以对信息的明确评估为条件
【arXiv】Large language models can produce fluent answers when their factual support is weak. This paper introduces Chain-of-Self-Questioning (CoSQ), a prompt-only framework that makes answer commitment conditional on an explicit assessment of the information
【arXiv】修剪可以降低大型语言模型（ LLM ）的部署成本，但其对基于上下文的工具调用的影响仍然知之甚少。我们系统地研究了四个LLMs spannin中智能家居工具调用的修剪引起的降解
【arXiv】Pruning can reduce the deployment cost of large language models (LLMs), but its impact on context-grounded tool calling remains poorly understood. We systematically study pruning-induced degradation in smart-home tool calling across four LLMs spannin
⭐ GitHub 热门项目GitHub Trending
【GitHub】给Claude Code一个电话。人工智能语音代理发出呼叫并阻止—在您应答之前，它无法同意任何事情，通话中。（⭐ 3 ）
【GitHub】Give Claude Code a phone. An AI voice agent makes the call and blocks — it can't agree to anything until you answer, mid-call. （⭐ 3）
【GitHub】Altim —一目了然的人工智能使用情况。用于AI编码代理使用的本地第一台桌面监视器。（⭐ 3 ）
【GitHub】Altim — AI usage, at a glance. Local-first desktop monitor for AI coding agent usage. （⭐ 3）
【GitHub】使用OpenCode AI构建的自治、只读的Kubernetes根本原因分析（ RCA ）代理，具有自定义Python安全钩子。（⭐ 3 ）
【GitHub】An autonomous, read-only Kubernetes Root Cause Analysis (RCA) agent built with OpenCode AI, featuring custom Python security hooks. （⭐ 3）
【GitHub】发现开源项目并策划贡献机会的代理系统，由Claude Code代理提供支持。（⭐ 0 ）
【GitHub】Agentic system that discovers open source projects and curates contribution opportunities, powered by Claude Code agents. （⭐ 0）
【GitHub】开源Android移动AI开发环境： PRoot Ubuntu + Claude Code/OpenCode/9router/headless Flutter SDK ，使用Flutter构建（⭐ 0 ）
【GitHub】Open-source Android Mobile AI Development Environment: PRoot Ubuntu + Claude Code / OpenCode / 9router / headless Flutter SDK, built with Flutter （⭐ 0）
🚀 模型与行业动态Models & Industry
根据Jensen Huang的说法，人工智能并不是“外星人思维”的新形式。它只是硬件和软件，因此每个人工智能产品制造商都可以设计安全性。
AI isn't some new form of "alien mind," according to Jensen Huang. It's just hardware and software, so safety can be engineered by each AI product maker.
新的WhatsApp Business MCP服务器允许开发人员使用Claude、Cursor、Codex和ChatGPT等AI编码代理来处理设置、消息传递模板、测试和故障排除。
A new WhatsApp Business MCP server lets developers use AI coding agents like Claude, Cursor, Codex, and ChatGPT to handle setup, messaging templates, testing, and troubleshooting.
从苹果一再推迟Siri AI到OpenAI凌乱的“超级应用”发布，以下是关闭或未达到预期的人工智能项目。
From Apple's repeatedly delayed Siri AI to OpenAI's messy "super app" launch, here's a look at the AI projects that shut down or missed expectations.
🔥 社区热议Community
【Lobsters】热度: 10↑ | 5 评论 | 标签: vibecoding
【Lobsters】热度: 10↑ | 5 评论 | 标签: vibecoding
【Lobsters】热度: 7↑ | 0 评论 | 标签: security, vibecoding
【Lobsters】热度: 7↑ | 0 评论 | 标签: security, vibecoding
【HN】热度: 39 分 | 0 评论
【HN】热度: 39 分 | 0 评论
【HN】热度: 32 分 | 14 评论
【HN】热度: 32 分 | 14 评论