01政策相关
2Anthropic 据报道签约高达 5170 亿美元算力协议，锁定至少 14.8 GW 算力
3 天前据 The Information 报道，Anthropic 在十一个月内签署了价值高达 5170 亿美元的算力合同，自 2025 年 10 月以来锁定至少 14.8 GW 算力，并计划自建数据中心。
最高法发布涉人工智能纠纷案件审理意见，明确 AI 换脸拟声等裁判规则
4 天前最高人民法院发布《关于依法审理涉人工智能纠纷案件的意见》，共 5 部分 24 条，明确 AI 换脸拟声、AI 复活逝者、大数据杀熟、仿冒名人带货、网络开盒及自动驾驶事故等情形的裁判规则。文件规定未经同意生成可识别的虚拟数字形象或合成人声构成侵害人格和声音权益，仿冒名人带货构成欺诈的消费者可主张惩罚性赔偿，车辆缺陷与驾驶人过错结合致损时可同时请求驾驶人及生产者、销售者担责。
04论文研究
8EyeMakeYou: Identity-, Task-, and Subjective-State-Conditioned Diffusion for High-Frequency Gaze Synthesis
4 天前arXiv:2609.04501v1 Announce Type: new Abstract: Eye movement biometrics (EMB) is an emerging behavioral modality for user authentication, particularly in virtual- and augmented-reality systems, where gaze dynamics contai…
Matched Starts, Divergent Objects: How Human-AI Collaboration Forms What It Explains
4 天前arXiv:2609.04542v1 Announce Type: new Abstract: Scholarly knowledge is typically encountered in stabilized form, while the process histories through which research objects, claims, and contributions acquire form remain l…
Beyond Prompt-to-App: Accountable Translation in Teacher-Facing Agentic Authoring
4 天前arXiv:2609.04679v1 Announce Type: new Abstract: Natural-language app builders let domain experts create software, but their pipelines transform professional intent across compilation, generation, checking, and approval. …
MIVAIS: A Study Environment for Multi-Agent Mixed-Initiative Visual Analytics Applications
4 天前arXiv:2609.04983v1 Announce Type: new Abstract: Mixed-initiative Visual Analytics (VA) systems empower human users by interleaving human intuition with software agents and their machine intelligence. However, the develop…
How a Chatbot's Response Style Shapes a Classroom: A Multi-Agent Simulation of Students Consulting AI
4 天前arXiv:2609.05018v1 Announce Type: new Abstract: LLM-based chatbots are increasingly used as everyday confidants. Because they are designed to maximize user satisfaction, they can respond with excessive empathy and affirm…
Scales, Reflections, and Conversations: A Multi-Modal Approach to Emotion Annotation
4 天前arXiv:2609.05046v1 Announce Type: new Abstract: Mental health concerns are increasing worldwide, highlighting the need for interventions that support everyday emotional well being. Prior work has demonstrated the potenti…
Beyond Bias: Participatory and Reflective Approaches to Cultural AI
4 天前arXiv:2609.05102v1 Announce Type: new Abstract: Generative AI systems increasingly shape cultural production, yet creative intentions, cultural meanings, and interpretive practices often can't be articulated through comp…
TherMosaic: Accelerating Perceived Thermal Transitions Through Spatiotemporal Thermal Feedback
4 天前arXiv:2609.05347v1 Announce Type: new Abstract: Thermal feedback can enrich immersive interaction, but thermoelectric devices often change temperature too slowly to match interactive timing. We present TherMosaic, a spat…
05技巧与观点
2Anthropic 发布 Claude Platform 降本指南并更新 claude-api 技能
4 天前Anthropic 介绍用 Claude Platform 降低成本、保持性能的三种方法：提高提示词缓存命中率、升级到前沿模型时清除提示词反模式、按任务校准 effort。
GPT-6 Astra爆火后，卡兹克谈执行能力贬值与判断力断层
4 天前GPT-6 Astra上线后，大量用户用它自主操控Blender、Houdini、Unity、Aseprite等专业软件，做出游戏Demo、3D复刻旧金山艺术宫等作品，其中旧金山艺术宫案例里Astra自己搜索几百张参考图、翻到美国国会图书馆的老扫描文件找柱子尺寸，多数工作在夜间自主完成。