HTML · ★ 10,355 · 🍴 673 · 📈 2,855 stars today
29 editorial diagram types for Claude Code. Self-contained HTML + SVG. No shadows, no Mermaid-slop.
中文介绍 为Claude Code提供29种出版物级图表模板，每个都是自包含的HTML+SVG文件，无外部依赖。与常见的Mermaid图表相比，风格更干净现代，无阴影等花哨效果。适合开发者在AI编程流程中快速生成架构图、示意图等高质量文档配图。
Rust · ★ 1,785 · 🍴 222 · 📈 227 stars today
Macro is a unified workspace for teams: email, chat, docs, tasks, agents, calls, and CRM — @-linked together with shared AI memory.
中文介绍 Macro是一个团队级统一工作空间，把邮件、聊天、文档、任务、智能体、通话和CRM整合在一起，用@链接和共享AI记忆打通信息孤岛。解决了多工具切换和上下文丢失的问题，适合注重协作效率的团队。技术层面属于AI驱动的集成平台。
Python · ★ 5,710 · 🍴 624 · 📈 845 stars today
Graph-Native Infrastructure for Context and Accountable AI Systems
中文介绍 基于图原生的基础设施，专为上下文感知和可问责AI系统设计。它利用图结构组织和管理AI运行的上下文，使系统行为可追踪、可审计。典型用户是AI平台开发者，用于构建金融、医疗等需要严格合规和可解释性的智能应用。
TypeScript · ★ 43,865 · 🍴 3,056 · 📈 1,235 stars today
Orca is the ADE for working with a fleet of parallel agents. Run any coding agent with your own subscription. Available on desktop, mobile and VPS.
中文介绍 Orca是一个用于管理并行代理集群的Agent开发环境（ADE）。用户可使用自有订阅运行任意编码代理，并在桌面、移动端和VPS上统一调度。适合需要同时推进多个自动化任务的开发者，提高编码代理的利用率和协同效率。
Shell · ★ 144,557 · 🍴 23,418 · 📈 1,873 stars today
A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.
中文介绍 这个项目打包了一个“AI代理公司”，包含前端设计、社区运营、创意生成、事实核查等各具特色的专家代理，每个代理都有独立的个性和工作流程。用户可以直接调用这些代理完成对应任务，相当于拥有一个虚拟数字团队。技术上是多代理协作框架。
Python · ★ 36,938 · 🍴 6,150 · 📈 266 stars today
Kronos: A Foundation Model for the Language of Financial Markets
中文介绍 Kronos是一个面向金融市场的语言基础模型，专门学习价格、交易量等市场数据的“语言”模式。它可应用于量化分析、趋势预测等场景，为金融从业者提供数据驱动的洞察。基于大规模金融语料训练，是专业投资者和量化研究员的辅助工具。
Python · ★ 61,964 · 🍴 12,122 · 📈 215 stars today
小红书笔记 | 评论爬虫、抖音视频 | 评论爬虫、快手视频 | 评论爬虫、B 站视频 ｜ 评论爬虫、微博帖子 ｜ 评论爬虫、百度贴吧帖子 ｜ 百度贴吧评论回复爬虫 | 知乎问答文章｜评论爬虫
中文介绍 MediaCrawler是一个支持多个中文社交平台的内容爬虫，可抓取小红书、抖音、快手、B站、微博、百度贴吧、知乎等平台的帖子、视频和评论数据。适用于数据分析师、运营人员和研究者进行舆情分析、竞品调研和用户行为研究。基于Python实现。
Python · ★ 45,556 · 🍴 3,713 · 📈 476 stars today
AI turns documents or topics into real, native PowerPoint decks—with native shapes, transitions and animations, data-backed charts and tables on demand, audio narration from speaker notes, and support for your own .pptx templates. · by Hugo He
中文介绍 ppt-master利用AI将文档或主题直接转换为原生PowerPoint演示文稿，生成的是真正的原生形状、切换动画和数据图表，还可从演讲者备注合成音频旁白。省去了手动排版的麻烦，适合商务人士和教育者批量制作专业课件。技术基于生成式AI与Office格式处理。
Go · ★ 87,543 · 🍴 10,309 · 📈 139 stars today
RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs
中文介绍 RAGFlow是一款开源RAG引擎，将检索增强生成与Agent能力融合，为LLM提供高质量上下文层。它解决了大模型事实不准、知识过时的问题，适合企业内部知识库问答、智能客服和文档检索等场景，是构建企业级AI应用的常用基础设施。
TypeScript · ★ 77,724 · 🍴 14,299 · 📈 571 stars today
The open-source app everyone uses to manage agents at work
中文介绍 Paperclip是一个开源应用，用于管理工作场景中使用到的各类AI代理。它提供统一的界面来配置、监控和协调多个代理，帮助企业有序地应用AI自动化。适合已经部署了多个AI助手的团队，提升治理和协作效率。
Rust · ★ 819 · 🍴 85 · 📈 421 stars today
中文介绍 NVIDIA NeMo项目下的开源仓库，目前没有官方描述。考虑到NeMo主要面向AI模型开发，它可能与模型训练或推理相关，但具体功能需查阅源码和文档。适合关注NVIDIA AI工具链的开发者留意。
TypeScript · ★ 36,063 · 🍴 5,027 · 📈 86 stars today
人人都能用英语
中文介绍 一个降低英语使用门槛的开源项目，主张“人人都能用英语”。可能包含学习方法、实用工具或内容资源，重点服务零基础和怕开口的普通学习者。项目具体的技术实现尚不明确，需要查看仓库来了解。
Python · ★ 20,345 · 🍴 3,298 · 📈 74 stars today
SpiderFoot automates OSINT for threat intelligence and mapping your attack surface.
中文介绍 SpiderFoot是一个自动化OSINT工具，用于威胁情报收集和攻击面映射。它可以自动从大量公开资源中汇总、关联数据，帮助安全团队识别暴露面、发现潜在风险。适合渗透测试、红队和蓝队使用，也是安全运营中的信息收集利器。
Dart · ★ 87,806 · 🍴 4,870 · 📈 213 stars today
An open-source cross-platform alternative to AirDrop
中文介绍 LocalSend是开源的跨平台AirDrop替代方案，利用Wi-Fi局域网直接在设备间传输文件，无需互联网或外部服务器。支持各大桌面和移动系统，解决了不同生态之间快捷传输的痛点，同时保护隐私，适合个人日常文件和团队内网传输。
Python · ★ 8,707 · 🍴 1,397 · 📈 65 stars today
Official Python inference and LoRA trainer package for the LTX-2 audio–video generative model.
中文介绍 LTX-2是音频-视频生成模型的官方Python工具包，提供推理和LoRA训练能力。开发者可以加载模型进行音视频生成，并通过LoRA对其进行低成本微调。适合AIGC研究者、内容创作者以及需要在本地部署生成模型的工程团队。
Kotlin · ★ 4,220 · 🍴 411 · 📈 40 stars today
Agent framework for the JVM. Pronounced Em-BAY-bel /ɛmˈbeɪbəl/
中文介绍 embabel-agent是一个面向JVM的Agent框架，让Java、Kotlin等开发者可以用熟悉的技术栈构建AI智能体。它可能提供了代理调度、工具调用等核心机制，简化集成复杂度。适合需要将AI代理能力嵌入企业级Java服务的团队。
Python · ★ 4,220 · 🍴 303 · 📈 315 stars today
14MB foundation model for tiny devices; phones, wearables, smart home, and robots.
中文介绍 Needle是一个仅14MB的基础模型，专为手机、穿戴设备、智能家居和机器人等微型终端设计。它在极低存储和算力条件下实现端侧AI推理，降低了对云端的依赖并保护隐私。适合嵌入式开发者和物联网产品团队落地轻量AI功能。
👍 553
We introduce BDH-CQ, a reasoning model that combines in-context learning with recurrent latent reasoning. Inputs presented at inference time continuously update the model's recurrent memory; the model then solves a query through iterative computation in a high-dimensional latent space, without verba
中文介绍 提出BDH-CQ推理模型，将上下文学习与循环潜在推理结合。推理时输入持续更新循环记忆，模型在高维潜在空间迭代计算求解查询，无需显式语言输出。核心贡献在于用潜在推理增强上下文学习能力。
👍 185
On-policy (Self-)Distillation (OPD / OPSD) has shown strong potential for post-training large language models (LLMs). However, existing methods still rely heavily on external supervision, including ground-truth signals, environmental feedback, or guidance from larger models, and therefore fall short
中文介绍 针对现有策略自蒸馏依赖外部监督的问题，提出无需任何监督的在线策略自蒸馏方法。利用模型自身输出作为训练信号，在策略分布上进行自我蒸馏，摆脱对真值标签、环境反馈或大模型指导的依赖，提升LLM后训练效果。
👍 173
After an older adult misses a medication dose, a software agent can send another reminder and an embodied agent can bring the medication. Yet neither explains whether the person forgot, is confused, has side effects, or deliberately refused, nor what support is appropriate. This reveals a structural
中文介绍 指出当前软件与具身智能体只能执行提醒、送药等动作，无法理解用户遗忘、困惑或拒绝的深层原因。提出以人为中心的ComBodied Agents范式，强调智能体应结构化理解用户状态与需求，实现真正贴合人的辅助。
👍 114
Agentic systems are increasingly expected to improve after deployment, yet single-entity self-evolution is often bounded by a static learning context, such as fixed tasks and feedback. This survey focuses on co-evolution in agentic systems, a multi-component form of self-evolution in which multiple
中文介绍 综述智能体系统中的共同进化（co-evolution）机制。指出单智能体自我进化受限于固定任务与反馈，而多组件协同进化可实现部署后的持续自我改进，超越人类预设设计。系统梳理了相关方法、挑战与未来方向。
👍 86
Leading large language model providers now conceal their models' step-by-step reasoning, or chain-of-thought, to protect intellectual property and limit information leakage. Rather than storing these traces server-side, providers return them to the client as blocks of encrypted text, which the clien
中文介绍 研究从专有LLM API窃取思维链推理痕迹的方法。供应商将逐步推理加密后返回客户端，攻击者可利用客户端可访问的加密文本块，通过侧信道或解密手段还原推理内容，揭示当前保护机制存在安全漏洞。
👍 46
Memory systems have shown promise for improving agent performance, but their potential remains largely unexplored for small language models, which struggle to generate sufficient successful trajectories on their own. We propose Agent Memory Distillation (AMD), a training-free framework that transfer
中文介绍 提出Agent Memory Distillation（AMD），一种无需训练的框架，将大模型教师的分层记忆迁移至小型LLM智能体。通过层级教师记忆传递成功轨迹经验，帮助小模型在没有足够自生成轨迹的情况下提升任务表现，拓展记忆系统在小模型上的应用。
👍 27
Conversational assistants increasingly recommend follow-up edits to help users continue a task. Existing systems primarily target text-only interactions, leaving image-creation conversations underexplored. In image-creation tasks, useful follow-up edit suggestions must reflect user preferences, offe
中文介绍 面向对话式图像编辑场景，研究如何生成视觉对齐的后续编辑建议。现有系统仅支持文本交互，而图像创作中建议需符合用户偏好并具视觉一致性。提出新方法生成与当前图像内容匹配的后续编辑推荐，填补该方向空白。
👍 24
Large language model (LLM) inference serving is increasingly constrained by memory rather than compute. As long-context and long-form reasoning workloads become more prevalent, the key-value (KV) cache dominates both memory footprint and memory traffic during LLM token generation, i.e., decode. In p
中文介绍 提出OasisKV，用前瞻稀疏预取（lookahead sparse prefetching）将解码期KV缓存扩展到HBM之外。针对长上下文与长推理场景中KV缓存主导内存占用和流量的问题，通过预测性稀疏预取降低内存瓶颈，提升LLM推理吞吐。
👍 22
Self-improving coding agents that iteratively rewrite their own source code have demonstrated impressive performance on coding tasks. However, existing solutions generally derive self-modification from a single failure trajectory at a time, overlooking rich comparative signals available in the agent
中文介绍 针对自改进代码智能体仅从单条失败轨迹学习、忽视对比信号的问题，提出Mendel Gödel Machine。通过比较进化机制，从多条轨迹中提取对照差异，递归重写自身源码，实现更稳健的自我改进，提升编码任务性能。
👍 16
Running a business is a challenging form of intelligent work. Operators must infer opportunities from partial signals, commit capital under uncertainty, adapt to delayed outcomes in a changing market, and satisfy regulatory obligations before trading legally. Frontier LLM agents can increasingly com
中文介绍 提出Business Arena基准，在逼真市场环境中评测LLM智能体经营企业的能力。任务要求智能体从不完整信号中推断机会、在不确定条件下配置资本、适应延迟的市场反馈并满足监管要求，全面评估其决策与适应能力。
👍 15
Large language model (LLM) agents are increasingly deployed as personal assistants. Existing evaluations, however, mostly use short, self-contained requests in static environments. Everyday life assistance is different. A task runs for weeks rather than minutes. The world keeps changing while the ag
中文介绍 提出VibeLifeBench基准，评测智能体在动态生活世界中的主动性与持久性。现有评测多基于静态环境中的短时任务，而日常任务持续数周且环境不断变化。该基准要求智能体长期跟踪目标、主动决策并适应变化，填补生活助手评估空白。
👍 12
Large language model evaluations typically focus on performance under nominal conditions, creating an illusion of capability where models comfortably walk a narrow, highly optimized generation corridor. In real-world deployments, however, complex system prompts, safety guardrails, and structural con
中文介绍 针对LLM评测仅关注标称条件、造成能力错觉的问题，提出解码层禁忌（Decoding-Level Taboo）诊断压力测试。在复杂系统提示、安全护栏等真实部署条件下，通过禁忌式解码扰动检验模型鲁棒性，暴露其在非优化生成通道上的脆弱点。
👍 11
Long-horizon research agents solve open-ended tasks through iterative retrieval, aggregation, and synthesis, but context grows rapidly while the marginal value of additional evidence often declines. This leads to unnecessary token cost, higher latency, and noisier inputs for final report generation.
中文介绍 针对长程研究智能体上下文快速增长而新增证据边际价值递减的问题，提出边际价值估计方法。在迭代检索与综合过程中动态评估额外token的收益，提前终止低价值检索，降低延迟与token成本，提升最终报告生成质量。
👍 11
Self-evolving agents accumulate reusable skills by appending successful procedures and failure fixes. Over time, the same requirement is often restated in several branches, examples, and warnings, while common action sequences are copied rather than reused. The resulting skill becomes expensive to i
中文介绍 提出SkillZip，无需评估的技能压缩方法，用于自进化智能体。通过发现技能库中可复用的结构化模式，消除重复分支、示例与警告，合并相同操作序列，压缩技能表示，在降低存储与推理开销的同时保留关键能力。
👍 11
General-purpose reward models are increasingly the bottleneck for scaling robot learning, yet the recipe for learning value-related capabilities from large-scale heterogeneous corpora remains underexplored. Existing approaches tie supervision to task-internal anchors such as preferences or normalize
中文介绍 提出RynnValue，利用时间距离（temporal distance）扩展机器人价值基础模型。现有方法将监督绑定到任务内部锚点，难以利用大规模异构语料。RynnValue以时间距离为通用监督信号，学习跨任务价值函数，缓解奖励模型瓶颈，提升机器人学习扩展性。
👍 10
Long-document understanding often requires reasoning over many visually rich pages, making inference costly and prone to context rot. In this work, we propose InSight-doc, an agentic visual perception framework that treats visual resolution as an adaptive reasoning-time resource. InSight-doc starts
中文介绍 提出InSight-doc，将视觉分辨率视为可自适应分配的推理时资源，用于长文档理解。框架以智能体方式动态选择页面区域与分辨率，避免对每页高分辨率处理，降低推理成本并缓解上下文腐化，提升多视觉页面文档的理解效率。
👍 10
Visual document retrieval (VDR) is dominated by multi-billion-parameter models that are slow to index at full corpus scale and expensive to serve. Prior compression routes either train a smaller multi-vector encoder from scratch or distil only the query side; neither yields a compact single-vector r
中文介绍 提出DistilVDR，通过双学生蒸馏（dual-student distillation）训练紧凑的端到端视觉文档检索器。针对现有数十亿参数模型索引慢、服务贵的问题，在压缩时同时蒸馏查询端与文档端，得到轻量单向量检索器，兼顾检索性能与部署效率。
👍 10
We study reference-free post-training for multilingual machine translation with open large language models. Starting from the supervised-finetuned MiLMMT-46-v0.1 models, we apply Group Relative Policy Optimization (GRPO) with a reward that averages two reference-free quality estimation models and is
中文介绍 研究开放大语言模型的无参考多语机器翻译后训练。在监督微调模型MiLMMT-46-v0.1基础上，采用GRPO强化学习，以两个无参考质量估计模型的平均分数为奖励，不依赖参考译文进行优化，提升多语翻译质量，扩展了后训练范式。
👍 10
Large language models (LLMs) are increasingly deployed as mobile assistants, where a key challenge is leveraging personal information scattered across multiple applications (apps) to complete user instructions. However, due to the lack of dedicated benchmarks, their capabilities remain poorly unders
中文介绍 提出SPIEval基准，评估LLM作为移动助手时利用跨应用分散个人信息的能力。现有评测缺乏专门基准，该工作构建多应用场景下的任务，检验模型聚合、推理并执行涉及散落信息指令的表现，填补移动助手评估空白。
👍 10
With the rapid advancement of large language models (LLMs), harnesses have become essential infrastructure for deploying agents across a wide range of domains. The fast-evolving harness ecosystem has also made rigorous capability evaluation increasingly important. However, efficiently building an en
中文介绍 提出A²E，一个端到端智能体审计引擎。针对LLM harness生态系统快速演进带来的能力评估难题，A²E提供高效的自动化审计流程，支持对智能体能力进行系统化构建与验证，降低评估成本，提升部署可靠性。
@thejessezhang · 85.7K 粉丝 · 270.4K 阅 · 510 赞 · 30 转
Two-thirds of our deployment work is now done autonomously by our own product (via Duet). This post is about why that is, our philosophy on building a product + deployment motion, and how that's
中文介绍 作者分享自家产品通过 Duet 已自动完成三分之二部署工作的实践，阐述「产品+部署」协同构建的哲学与动机，探讨 AI 自主完成工程任务的可能性，属于少见的团队自用案例。
@addyosmani · 408.3K 粉丝 · 175.5K 阅 · 509 赞 · 65 转
For much of human history, we've evaluated code quality via code review: someone reads what you wrote and makes sure it's clean, thoughtful, fast, understandable, and tests well. For agents, that
中文介绍 Addy Osmani 探讨 agentic 代码质量：人类历史上代码质量靠人工审查保证，但面对编码 agent，如何定义「好」成为核心工程问题，并提出面向代理的评估新思路，是 AI 工程领域的重要观点。
@VibeMarketer_ · 30.1K 粉丝 · 107.1K 阅 · 528 赞 · 29 转
I am going to show you how to turn the models, agents, skills, and automations living across your team's tools and workflows into one coordinated system. What most teams have today looks very
中文介绍 VibeMarketer_ 分享如何将团队分散在各工具和工作流中的模型、agents、skills 和自动化整合为一个协调系统，解决当前团队 AI 工具碎片化问题，强调从现状出发构建统一 harness 而非推倒重来。
@nvidia · 2.6M 粉丝 · 59.5K 阅 · 548 赞 · 66 转
The new lightweight open model and routing library delivers greater control over AI, data and workflows across edge devices, PCs, workstations, data centers and the cloud. Source: NVIDIA Blog By Kari
中文介绍 NVIDIA 发布轻量级开放模型 Nemotron 3.5 Lightning 和路由库 NeMo Switchyard，为边缘设备、PC、工作站、数据中心及云端带来对 AI、数据和工作流更强的控制力，主打更快更高效的 agentic AI。
@GoogleCloudTech · 1.3M 粉丝 · 46.8K 阅 · 505 赞 · 63 转
By @Saboo_Shubham_ and @secchi_elia When coding agents write the code, defining what's good is the real engineering job. Coding agents can now build and improve other agents. Write the instructions,
中文介绍 Google Cloud Tech 分享每个 AI 工程师应懂的 7 条自改进 agent 循环规则：编码 agent 写代码时，定义「好」才是真正的工程；agent 已能构建并改进其他 agent，核心在于编写指令。
@dexhorthy · 30.0K 粉丝 · 42.9K 阅 · 635 赞 · 31 转
tl;dr make your agent converse visually instead of in walls of prose. Lighter and faster than HTML, good enough for most dev-work shaped problems. Coding agents are pretty much unreadable The
中文介绍 dexhorthy 提出 /show-me 方法：让编码 agent 用紧凑可视化表达代替长篇文字对话，比 HTML 更轻量快速，足以应对多数开发工作，解决 agent 输出「不可读」问题。
@mattyp · 44.5K 粉丝 · 33.5K 阅 · 651 赞 · 48 转
I’m always chasing better tools. Notes apps, workflows, optimizations, and now, personal agents. It’s an easy trap to fall into - always building the “custom” thing and sacrificing the work as a
中文介绍 mattyp 介绍自建个人代理 Grok Bot，并反思追逐更好工具的习惯：总是构建「定制」方案反而可能牺牲实际工作产出，提醒在优化工具与做事之间保持平衡。
@ericzakariasson · 80.4K 粉丝 · 32.2K 阅 · 719 赞 · 52 转
Grok 4.6 is out! I've used it for a few weeks as my daily driver across the normal mix of coding and knowledge work, and built a few projects with it specifically to push on where it holds up. It's
中文介绍 ericzakariasson 发布 Grok 4.6 使用指南：数周内将其作为日常主力，覆盖编码与知识工作，并专门构建项目测试其能力边界，给出实际表现评估。
Business and technology leaders need no convincing that the time of agentic AI is here. Organizations are rapidly adopting agents, and few executives doubt the technology’s potential to transform work. But many organizations find that realizing the desired return on investment (ROI) from AI hinges o
中文介绍 商业与科技领袖已确信智能体AI时代到来，企业正加快采用AI智能体，但许多组织发现难以实现预期投资回报；文章强调需要以可信数据支撑智能体规模化落地。
中文介绍 Hugging Face 发布 OlmoEarth 嵌入（Embeddings）功能，用户可在 OlmoEarth Studio 中自定义导出嵌入向量，供下游分析使用。
Introducing sign-language-to-text (SL2T), our breakthrough model powering new sign language features for Deaf and hard of hearing users.
中文介绍 DeepMind 推出手语转文字模型 SL2T（Sign-Language-to-Text），为失聪和听障用户提供新的手语功能。
中文介绍 Hugging Face 博客介绍 Liquid AI 的 LFM2.5-VL-3B 模型，宣称其为边缘设备提供更好、更快的视觉能力。
Speculative Decoding by any other name would distil as sweet
中文介绍 文章围绕“如何窃取推理轨迹”展开，提到投机解码（Speculative Decoding）概念，并探讨其与蒸馏的关系。
OpenAI research reveals how enterprises are adopting agentic AI, using ChatGPT and Codex, and how frontier firms are pulling ahead in AI adoption.
中文介绍 OpenAI 研究显示，企业正采用智能体AI，通过 ChatGPT 和 Codex 推动工作；前沿企业在 AI 采用方面处于领先。
See how RingCentral uses ChatGPT Work and Codex to accelerate AI product development and centralize operational intelligence across engineering and operations.
中文介绍 RingCentral 借助 ChatGPT Work 和 Codex 加速 AI 产品开发，并在工程与运营环节集中整合运营智能。
中文介绍 本期 AI 新闻摘要涵盖：Cursor 推出 GitHub 竞品、OpenAI 首席运营官离职，以及 Gemini 达成 10 亿用户里程碑。
Pharma is suddenly paying for Bio × AI tools, and Chai is leading the pack with four deals closed this summer. Cofounder Matt McPartlon and Product leader Neil Patil explain why.
中文介绍 制药行业开始为“生物×AI”工具付费，Chai Discovery 今夏已达成四笔交易；联合创始人 Matt McPartlon 与产品负责人 Neil Patil 解释原因。
中文介绍 IBM Research 在 Hugging Face 博客发布文章，称其方法可用比 ACE 更少的 token 完成同样任务，涉及 ALTK-Evolve 与 SLDD。
OpenAI begins testing ads in ChatGPT to support free access, with clear labeling, answer independence, strong privacy protections, and user control.
中文介绍 OpenAI 开始在 ChatGPT 中测试广告，以支持免费访问；广告将明确标注，保证答案独立性，并采取严格隐私保护与用户控制。
OpenAI and AWS are making Daybreak cybersecurity capabilities available through Amazon Bedrock to support enterprise security workflows.
中文介绍 OpenAI 与 AWS 合作，通过 Amazon Bedrock 提供 Daybreak 网络安全模型，以支持企业安全工作流。
**xAI's Grok 4.6** advances frontier pricing and performance, scoring **61 on the Intelligence Index** and showing strong agentic results, with **Grok 4.7** already in training. **Alibaba's Qwen3.8-Max** open weights release features a **2.4T parameter model with 95B active MoE**, notable for day-0
中文介绍 xAI 的 Grok 4.6 在智能指数（Intelligence Index）得 61 分，前沿定价与性能提升，智能体表现强劲；Grok 4.7 已在训练中。阿里巴巴开源的 Qwen3.8-Max 为 2.4T 参数、95B 激活参数的 MoE 模型。
a small win for american open models - Glimmer runs on a fits on a single RTX 3090!
中文介绍 美国开放权重模型取得小胜：Muse Glimmer 可在单张 RTX 3090 上运行；Muse Glimmer 和 Spark 的开放权重被视为实现个人超级智能的承诺。
中文介绍 本期 AI 新闻摘要涵盖 Muse Glimmer 发布、OpenAI 网络安全（Cyber）进展，以及 Claude 挑战黎曼猜想。
本次21个标的技术面呈「股票商品偏强、加密偏弱、中概中性」的分化格局。美股内部有支撑但分化加剧：SPY距52周高点仅-0.56%且多头排列，NVDA上涨3.03%且MACD在零轴上方；但META跌破全部均线并延续MACD死叉，MSFT的RSI 70.8进入超买区，显示动量集中于部分标的。加密板块整体受压：BTC-USD距52周高点-49.73%，MACD死叉（macd -16.33 vs 信号55.14），恐慌贪婪指数29处于恐慌区、加密总市值2.26T维持缩量；ETH-USD RSI 50.4中性，SOL-USD在SMA20附近整理，三者均无趋势性突破。中概股缺乏一致性：BABA/PDD/JD均位于SMA50上方、RSI在54-58之间，MACD多空交错；腾讯控股5日跌6.22%且MACD死叉，相对偏弱。商品外汇端强弱分明：黄金期货RSI 72.2超买但5日涨6.01%，WTI原油MACD金叉且5日+9.48%；美元/人民币RSI 25.4超卖且空头排列，美元指数则在SMA200附近区间运行。综合来看，美股与商品动能占优，加密与社交股技术面偏弱，中概多数处于中性整理。
现价224.09高于SMA20/50/200（208.38/206.26/194.56），呈多头排列；MACD 4.32高于信号线2.17，RSI 62.4未进入超买区，单日上涨3.03%，距52周高点仅-5.26%。短期动量结构偏强，技术setup明确偏上行。
现价4500.9高于SMA20（4141.94）、SMA50（4163.49）及SMA200（4483.61）全部均线，MACD 65.98远高于信号线14.15，5日涨幅6.01%；但RSI 72.2已进入超买区，上方存在技术性回归压力，整体偏上行但需留意超买回调。
现价578.85低于SMA20（600.96）、SMA50（597.85）与SMA200（629.16），均线空头排列；RSI 43.9运行于50中轴下方，MACD -6.44低于信号线-5.66，死叉结构延续；距52周高点-27.3%，技术setup明显偏下行。
现价6.74低于SMA20（6.76）、SMA50（6.77）与SMA200（6.89），空头排列；RSI 25.4处于超卖区，MACD -0.0097低于信号线-0.008，且接近52周低点。下行结构完整，超卖或有技术性停顿，但态势仍偏下行。
现价1876.57贴近SMA20（1890.57），高于SMA50（1812.46）但低于SMA200（2033.95）；RSI 50.4恰好落在中轴，MACD 12.95略低于信号线19.11，方向信号相互抵消。加密恐慌贪婪指数29处于恐慌区，ETH技术读数偏中性。
VIX 恐慌指数
10Y 美债收益率 (%)
美元指数 DXY
S&P 500 ETF
Nasdaq 100 ETF
Apple
Microsoft
Nvidia
Alphabet
Tesla
Meta
Bitcoin
Ethereum
Solana
阿里巴巴 (BABA)
拼多多 (PDD)
京东 (JD)
腾讯控股 (0700.HK)
黄金期货
WTI 原油期货
美元 / 人民币
本报告仅为技术指标状态描述，不构成投资建议；过去走势不代表未来表现，指标读数可能随行情快速变化，仅供技术指标解读参考。
Vicki Brady still paid close to $7m despite 20% of her bonus being docked by the company. Follow today’s news live Get our breaking news email, free app or daily news podcast Legislation forcing tech companies to pay for Australian journalism to be introduced The government will today introduce its
中文摘要 澳大利亚电信公司Telstra在经历全国性网络中断后，仍向首席执行官维姬·布雷迪支付约680万美元年薪，仅扣减其20%奖金。此外，维多利亚州州长告诫抗议保利娜·汉森的学生“留在学校”。
The vessel was crowded far beyond capacity and was hit by strong waves, the police said. While dozens have been rescued, the search continues for an unknown number of missing people.
中文摘要 赞比亚卡里巴湖一艘渡轮因严重超载并遭遇强浪而倾覆，已造成逾40人死亡。数十人获救，仍有不明数量人员失踪，搜救工作持续进行。
From Iceland to Spain, people flocked to viewing spots to see the moon completely block out the sun.
中文摘要 欧洲迎来数十年来首次日全食，从冰岛到西班牙，大量民众聚集在观测点，观赏月亮完全遮住太阳、白昼转黑夜的罕见天象。
Erling Haaland’s name and image turned up on hundreds of cocaine bricks seized by Ecuadorian police.
中文摘要 厄瓜多尔警方查获半吨可卡因，毒品砖上印有挪威足球明星埃尔林·哈兰德的名字和肖像，目前案件正在进一步调查中。
President Trump again insists US has 'total control' of strait, while Tehran resolute that the US must meet its demands.
中文摘要 围绕霍尔木兹海峡控制权，美国总统特朗普坚称美国拥有“完全控制”，伊朗方面则回应称美国必须满足其要求，双方各执一词。
Two families have been trapped by settlers for days, casting a spotlight on the army’s struggle to curtail settler violence and protect Palestinians living in the occupied West Bank.
中文摘要 约旦河西岸两个巴勒斯坦家庭遭以色列定居者围困多日，以军短暂现身随后撤离。事件凸显以军难以遏制定居者暴力、保护巴勒斯坦人。
Trump's poll numbers appear to be down, but in this town, his support is as strong as ever.
中文摘要 美国一个以制造导弹为业的小镇对特朗普支持依然坚定，当地口号称“不保守就别来”。尽管特朗普民调下滑，该镇民众热情不减。
Heidi Beirich, an expert on far-right groups, has been charged over alleged misuse of donor funds to pay informants.
中文摘要 特朗普政府逮捕南方贫困法律中心专家海迪·贝里奇，指控其涉嫌滥用捐赠资金支付线人。该中心以研究极右组织闻名，案件引发关注。
Lawsuit claims Trump profits unfairly by charging up to $100,000 monthly for exclusive access to presidential posts.
中文摘要 多家媒体集团起诉特朗普，指控其通过收取最高每月10万美元费用，向媒体提供总统社交媒体帖子的独家访问权，属不公平获利。
Trump boarded a different plane in Turkey during a Nato summit for security reasons involving Iran.
中文摘要 北约峰会期间，特朗普因伊朗相关安全原因在土耳其换乘另一架飞机秘密离开，其内阁成员则留在诱饵飞机上以转移视线。
Some see them as sinister or hunt them for ritual use but these birds play a vital role in nature's cycle.
中文摘要 尼日利亚秃鹫数量持续减少，有“秃鹫之王”称号的环保人士正努力扭转趋势。秃鹫虽被部分人视为不祥或猎杀用于仪式，但对自然生态循环至关重要。
In the comments, Kelce called Swift his "wife" for first time and said it was a "childhood dream" to wed at Madison Square Garden.
中文摘要 特拉维斯·凯尔斯首度公开与泰勒·斯威夫特的婚礼细节，称在麦迪逊广场花园完婚是“童年梦想”，并首次称斯威夫特为“妻子”，形容婚礼是“一生中最美好的夜晚”。
Japanese risk aversion and conservatism blamed for slow AI take-up by the country's business sector.
中文摘要 日本企业在应用人工智能方面进展缓慢，外界普遍归咎于该国企业的风险规避心态和保守主义，导致新技术采用率偏低。
Leavitt will depart from her role to spend time with her young children and family, the president says.
中文摘要 美国总统特朗普宣布，白宫新闻秘书卡罗琳·莱维特将离职，以便有更多时间陪伴年幼子女和家人。目前继任人选尚未公布。
An online petition accusing Pete Hoekstra of diplomatic interference in Canadian domestic affairs has garnered more than 60,000 signatures.
中文摘要 数千名加拿大人签署在线请愿，要求驱逐美国大使皮特·霍克斯特拉，指控其干预加拿大内政。请愿已获得超过6万份签名。
Prices for Japan’s corporate goods continued to rise at an elevated pace in July, keeping high cost pressure on companies, as central bank officials continue to consider whether to proceed with additional interest rate hikes to contain inflation.
Before robots can take over our jobs, they first need to learn how humans move. With cameras strapped to their foreheads, blue-collar workers in India are training humanoid robots by filming how they use their hands in what might be AI's latest gold rush. (Source: Bloomberg)
Brazil’s Congress approved measures to curb spending growth in a move that aims to signal greater budget discipline backed by President Luiz Inacio Lula da Silva ahead of his reelection bid later this year.
South Korean stocks climbed, putting the benchmark index on track to enter a technical bull market, as a global rebound in the AI trade fueled a rapid reversal from last month’s historic rout.
Traders will be watching whether Indonesian President Prabowo Subianto’s state budget speech on Friday can convince investors that fiscal discipline remains a priority, helping restore confidence in the country’s battered assets.
Treasury Wine Estates Ltd.’s full-year earnings fell by more than a third as the company battled a global downturn in alcohol consumption and supply concerns in China and the US.
CapitaLand Investment Ltd.’s profit rose in the first half, driven by higher contributions from private funds and various real estate investment trusts, and the property asset manager forecast earnings growth.
ASX Ltd.’s full-year profit missed estimates and costs grew ahead of incoming chief executive officer Anthony Attia’s arrival at the Australian exchange next month.
Origin Energy Ltd.’s full-year profit beat analyst estimates, as the uptake of electric vehicles and home batteries in Australia accelerated.
Gold steadied near $4,400 an ounce after a subdued US inflation report relieved near-term pressure on the Federal Reserve to hike interest rates.
NSTA has fined one company more than £16mn this year and vows to enforce deadlines
Japanese risk aversion and conservatism blamed for slow AI take-up by the country's business sector.
10 回复 · 程序员 节点
32 回复 · 程序员 节点
16 回复 · Linux 节点
10 回复 · 程序员 节点
28 回复 · 程序员 节点
10 回复 · 程序员 节点
9 回复 · Apple 节点
16 回复 · Apple 节点
39 回复 · Apple 节点
11 回复 · Apple 节点
该源今日无内容。