Meta Superintelligence Labs 发布重组后首个模型 Muse Spark，已上线 Meta AI 应用及网站，未来数周将覆盖 WhatsApp、Instagram、Facebook、Messenger 和智能眼镜

Anthropic 推出专为安全漏洞发现设计的新模型 Claude Mythos Preview，与六大科技巨头联合启动 Project Glasswing 网络安全项目，几乎无需人工干预即可标记系统漏洞

仅 26 人的美国初创公司 Arcee 打造出高性能开源大模型，在 OpenClaw 用户群体中快速走红，证明小团队也能在开源模型领域突围

Google DeepMind 发布 Gemma 4 系列开源模型，涵盖 1B/4B/12B/27B 四个参数规模，支持原生多模态输入输出，在同等参数量下达到前沿性能，27B 版本支持设备端运行。

微软推出三款自研 MAI 系列基础模型，减少对 OpenAI 的依赖。

OpenAI宣布关闭Sora产品，下载量暴跌65%，曾让好莱坞颤抖的AI视频生成工具最终沦为弃子。

开源 AI 协作工具，带记忆功能，今日新增 1187 star

开源 Agent 管理平台，可将编码 Agent 转化为真正队友，今日新增 1680 star

Astropad 推出 Workbench，将远程桌面重新定义为 AI Agent 监控工具，支持从 iPhone/iPad 低延迟操控 Mac Mini 上的 Agent

提出编译式 AI 范式，LLM 在编译阶段生成可执行代码，之后工作流确定性地运行无需再调用模型，适用于医疗等高可靠性场景

提出可审计 Agent 架构，为自主 AI 系统的每一步决策提供可追溯、可验证的证据链，解决 Agent 部署中的信任和合规问题

提出混合视图方法处理机器数据（日志、指标等），让 LLM 在结构化查询和自然语言理解之间灵活切换

arXiv:2602.11236v2 Announce Type: replace-cross Abstract: Building general-purpose embodied agents across diverse hardware remains a central challenge in robotics, often framed as the ''one-brain, man

arXiv:2509.19695v3 Announce Type: replace Abstract: Task oriented dialog systems often rely on static exploration strategies that do not adapt to dynamic dialog contexts, leading to inefficient explor

arXiv:2601.14004v4 Announce Type: replace Abstract: Mechanistic Interpretability (MI) has emerged as a vital approach to demystify the opaque decision-making of Large Language Models (LLMs). However,

arXiv:2502.11271v2 Announce Type: replace-cross Abstract: Solving complex reasoning tasks may involve visual understanding, domain knowledge retrieval, numerical calculation, and multi-step reasoning.

arXiv:2603.10652v2 Announce Type: replace-cross Abstract: In real-world deployment, vision-language models often encounter disturbances such as weather, occlusion, and camera motion. Under such condit

arXiv:2604.12686v1 Announce Type: new Abstract: Recent advances in deep learning underscore the need for systems that can not only acquire new knowledge through Continual Learning (CL) but also remove

MIT Technology Review探讨AI时代以隐私为主导的用户体验设计如何建立用户信任