TypeScript · ★ 6,358 · 🍴 285 · 📈 1,725 stars today
Next generation e2e testing framework for web and mobile apps.
中文介绍 面向 Web 与移动应用的下一代端到端（E2E）测试框架，旨在解决传统 E2E 工具在跨平台场景下配置复杂、稳定性差的问题。测试工程师和 QA 团队可用它编写覆盖 Web 与移动端的自动化测试，简化测试流程并提升回归测试的可靠性。
Shell · ★ 278,171 · 🍴 23,291 · 📈 889 stars today
Skills for Real Engineers. Straight from my .agents directory.
中文介绍 由开发者 Matt Pocock 整理的工程技能集合，内容直接取自其个人 .agents 目录。为使用 AI 编码代理的工程师提供可复用的技能与配置，帮助代理在真实开发任务中具备更实用的能力，适合希望增强 AI 助手的开发者参考。
Python · ★ 17,991 · 🍴 1,805 · 📈 619 stars today
Give your agent CAD superpowers.
中文介绍 为 AI 代理赋予 CAD 建模能力的项目，让 agent 可通过自然语言文本生成或操作 CAD 模型。它把文本指令与计算机辅助设计工具打通，适合需要自动化建模、快速生成设计草图的工程师与设计团队使用。
C++ · ★ 6,592 · 🍴 489 · 📈 949 stars today
Tool for automatic PS5 executables porting to Linux and Windows
中文介绍 自动将 PS5 可执行文件移植到 Linux 和 Windows 平台的工具。针对 PlayStation 5 游戏或程序难以在其他系统运行的问题，通过自动化转译流程降低移植门槛，适合游戏移植爱好者、模拟器开发者及逆向工程研究者使用。
JavaScript · ★ 77,708 · 🍴 4,629 · 📈 616 stars today
The design language that makes your AI harness better at design.
中文介绍 一套设计语言，目标是让 AI 工具在界面设计上表现更出色。通过规范化的设计指令与风格约束，弥补 AI 生成设计时审美不足的问题，适合使用 AI 辅助设计的产品设计师和前端开发者，用于提升生成结果的视觉质量。
TypeScript · ★ 97,193 · 🍴 8,564 · 📈 534 stars today
Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
中文介绍 为 AI 代理提供跨会话的持久上下文记忆。它捕获代理在会话中的所有操作，用 AI 压缩后把相关上下文注入未来会话，解决重启后失忆的问题。支持 Claude 等工具，适合需要长期连续任务和个性化记忆的开发场景。
Python · ★ 54,424 · 🍴 3,125 · 📈 326 stars today
A skill to stop your coding agent from burying the answer. ADHD-friendly output.
中文介绍 一个让编码 AI 代理不再把答案埋在长篇输出里的技能，面向希望快速看到结论的 ADHD 用户。它约束代理输出更简洁直接、结论先行，避免冗余解释，适合需要高效获取答案、减少阅读负担的开发者。
TypeScript · ★ 9,527 · 🍴 1,060 · 📈 2,956 stars today
Reverse engineer anything with agents, from app behavior down to native binaries.
中文介绍 利用 AI 代理进行逆向工程的项目，覆盖从应用行为分析到底层原生二进制文件的层面。它把逆向流程交由代理自动化执行，降低分析门槛，适合安全研究员、逆向工程师和需要理解闭源程序行为的技术人员使用。
Cuda · ★ 8,712 · 🍴 1,372 · 📈 199 stars today
DeepGEMM: clean and efficient BLAS kernel library on GPU
中文介绍 DeepSeek 开源的高性能 GPU BLAS 内核库，聚焦 GPU 上干净高效的矩阵乘法实现。它为深度学习训练与推理提供底层算力支撑，采用 CUDA 编写并针对新架构优化，适合模型训练和推理框架开发者及高性能计算研究人员使用。
Shell · ★ 157,837 · 🍴 25,448 · 📈 623 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 一个开箱即用的 AI 代理团队集合，涵盖前端、社区运营、创意灵感、现实检查等多种角色，每个代理都有专门技能、个性和工作流程。用户可按需调用不同专家代理，适合希望快速搭建多角色 AI 协作流程的团队和个人。
JavaScript · ★ 5,719 · 🍴 783 · 📈 1,419 stars today
Self-hosted gym & body-weight tracker — plan routines, log workouts (supersets, warm-ups, cardio), see which muscles are trained, fatigued or detrained, import from FitNotes/Strong/Hevy, passkey login. Your data, your server.
中文介绍 可自托管的健身与自重训练追踪工具，支持规划训练计划、记录超级组、热身与有氧运动，并展示各肌群的训练、疲劳与退化状态。可从 FitNotes、Strong、Hevy 导入数据，支持 passkey 登录，适合注重隐私的健身爱好者。
HTML · ★ 44,048 · 🍴 2,842 · 📈 228 stars today
Editorial diagram design for Claude Code, Codex, GitHub Copilot, Factory Droid, and Pi. 42 diagram types. Self-contained HTML + SVG. No shadows. No Mermaid slop.
中文介绍 面向 Claude Code、Codex、GitHub Copilot 等 AI 编程工具的编辑级图表设计资源，内置 42 种图表类型，输出自包含的 HTML 与 SVG，无阴影、不依赖 Mermaid。适合需要美观、可定制图表而非默认 Mermaid 风格的开发者与文档作者。
👍 24
LLM-based agents are increasingly capable of generating complex 3D structures, with the potential to reshape how objects are designed and realized in the physical world. Yet, producing elegant geometry is fundamentally different from producing objects that can be built and perform their intended fun
中文介绍 针对 LLM Agent 生成 3D 结构「好看但不可造」的问题，LMBuild 提出面向可建造性与功能性的评测基准，考察生成结构在物理约束下能否被实际构建并完成预期功能，从而将几何美学与工程可实现性区分开来。
👍 22
Proactive LLM agents can turn idle compute into useful support before users ask. Yet even correct work can misread user context, impose review costs, or undermine trust. This work proposes foundations for designing, realizing, and evaluating proactive LLM agents around three joint principles (3T): T
中文介绍 主动式 LLM Agent 可在用户提问前提供帮助，但错误预判会引发审视成本与信任损耗。本文以 3T 等联合原则为基础，提出主动式 Agent 的设计、实现与评估框架，并给出分层技术体系与 Proactivity-Gym 评测环境。
👍 17
Recent advances have enabled unified omni-modal models in understanding audio, vision, and language. However, existing benchmarks, training data, and learning methods largely treat the modalities independently, leaving the capability of audio-visual joint reasoning poorly evaluated and insufficientl
中文介绍 现有全模态模型虽能分别理解音频、视觉与语言，但基准、训练数据与方法多将模态割裂，音视频联合推理能力未被充分评测。OmniReasoning 面向该缺口构建评测与训练方案，推动音视频联合推理的能力边界。
👍 17
Geometry optimization is a major cost in many quantum-chemical workflows: each optimization step requires one force evaluation, and at the density-functional level that evaluation dominates the wall time. Research in this area has produced a broad range of optimization methods, and we ask whether a
中文介绍 几何优化中每步都需一次力评估，DFT 级别下评估耗时主导整体成本。本文让语言模型自动发现更快的分子弛豫优化器，以替代人工设计：由 LLM 搜索优化策略，减少达到收敛所需的力评估步数，从而加速量子化学工作流。
👍 14
Search agents repeatedly make short decisions about relevance, evidence sufficiency, and search actions. Using generative language models for these decisions introduces latency and unreliable confidence. We present SearchJev, a fast and calibrated System-1 model that separates search decisions from
中文介绍 搜索 Agent 需反复判断相关性、证据充分性与下一步动作，用生成式 LLM 决策会带来延迟且置信度不可靠。SearchJev 将搜索决策与文本生成解耦，训练快速且校准良好的 System-1 模型，直接输出分类决策，兼顾速度与置信可靠性。
👍 9
Interactive world models are increasingly capable of generating environments and acting within them, yet deliberately editing an existing executable world remains underexplored. We formulate world editing as intervening on an existing world while preserving properties that should remain unchanged, a
中文介绍 交互式世界模型已能生成环境并行动，但有意识地编辑既有可执行世界仍缺乏研究。本文将世界编辑形式化为对已有世界施加干预并保持应不变属性，提出随深度递增的干预方法，使编辑在可执行世界中被系统性度量与实现。
👍 8
Does conversational memory need LLM-extracted facts, or is selecting the right raw turns enough? Published results disagree. Extraction-based systems report gains from distilled facts. Recent studies find raw history with good ranking does as well, but disagree about whether ranking matters. We ran
中文介绍 对话记忆究竟需要 LLM 抽取事实，还是选对原始轮次即可，已有结论相互矛盾。本文以类型化决策模型开展预注册实验，系统比较抽取式与选择式记忆在何种条件下等价或优劣，厘清排序质量等因素的作用。
👍 8
Manipulation behaviors vary widely across objects and scenes, but they share a small set of reusable skills, and planning with these skills helps embodied agents generalize to new tasks. Yet an agent can only plan with skills it knows. Recovering skills from observed experience, the inverse of plann
中文介绍 操作行为随物体场景各异，却共享少量可复用技能，用技能规划有助于泛化，但 Agent 只能规划已知技能。Video2Skill 从流式观察经验中反向恢复可复用具身技能，形成技能库供规划调用，提升新任务上的泛化能力。
👍 8
Embodied agents now take on ever longer tasks. For long tasks, knowing only whether a task finally succeeds or fails says little; the steps along the way matter. Progress Reward Models (PRMs) score how far a task has come at every step, and serve as dense rewards, verifiers and monitors. Yet in long
中文介绍 长时程任务中仅看最终成败信息量不足，过程奖励模型 PRM 需逐步评估进展。本文指出具身 PRM 若缺乏恰当上下文便会失准，分析何种上下文信息对进展打分至关重要，并据此改进长任务中的密集奖励、验证与监控。
👍 8
Tool-using AI agents are increasingly deployed across enterprise software systems, yet widely used benchmarks primarily evaluate nominal task completion, conflating baseline planning competence with operational fault recovery. We introduce UndoBench, a benchmark spanning 36 base workflows and 36 fau
中文介绍 现有工具 Agent 基准多只考察任务完成，混淆了基础规划能力与故障恢复能力。UndoBench 构建 36 个基础工作流与 36 类故障，将任务胜任度与撤销、回滚等恢复能力分离评测，揭示 Agent 在操作纠错上的真实短板。
👍 7
Computer-use agents need to reliably ground action targets in complex desktop scenes, where multiple applications, overlapping windows, and visually similar controls compete for attention. Existing training data rarely pair such scenes with dense annotations or vary them in a controlled way. We intr
中文介绍 计算机操作 Agent 需在应用繁多、窗口重叠、控件相似的桌面场景中准确落点，而现有数据缺少密集标注。DeskForge 从桌面环境构造可控变化的大规模密集监督数据，为 GUI 定位与操作训练提供更可靠的标注来源。
👍 6
Latent reasoning lets a large language model (LLM) think in a continuous space and verbalize only the answer. We argue that an effective latent thought must meet five requirements: it should be useful, helping produce the correct answer rather than merely changing it, diverse, so that resampling yie
中文介绍 潜空间推理让 LLM 在连续空间思考、只输出答案。本文用 flow matching 研究其关键要素，提出有效潜思维应满足五个条件，如有用（助益正确答案而非仅改变输出）、多样（重采样可得不同解）等，并据此分析训练与采样设计。
👍 4
Real-world time series are frequently driven by exogenous events and structural shifts, rendering conventional forecasting based solely on historical numerical observations insufficient. While language models can retrieve external news, standard retrieval-augmented approaches struggle with high nois
中文介绍 真实时间序列常受外部事件与结构突变驱动，仅凭历史数值难以预测。SEER 面向时序预测提出自演化的事件推理与检索机制，缓解标准 RAG 在检索外部新闻时的高噪声问题，动态筛选有效事件信息以提升预测精度。
👍 4
Reinforcement learning (RL) has greatly advanced the capabilities of large language models (LLMs), but its memory demands remain a barrier to broader adoption. We introduce LoGRA, an approach to RL post-training that reduces memory by retaining useful learning signals in low-rank gradient sketches.
中文介绍 RL 后训练显著提升 LLM 能力，但显存开销制约推广。LoGRA 以低秩梯度草图保留有用学习信号，在训练中压缩存储优化器状态，从而降低 RL 后训练的内存占用，使更大规模的 LLM 强化学习更易落地。
👍 4
Modern information systems, including many agentic workflows, use dense retrieval to explore large amounts of unstructured data. However, dense retrieval relies on surface-level semantic similarity, which is insufficient for increasingly complex search applications. Here, we investigate agentic retr
中文介绍 密集检索依赖表层语义相似度，难以应对日益复杂的搜索任务。本文系统考察面向复杂任务的 Agent 式检索，比较其相对语义检索的性能与成本，分析多步检索在何时带来收益、何时开销不划算，给出选型依据。
👍 3
Jev-style decision models return categorical probability distributions over predefined options without generating free-form text, enabling software systems to act on their outputs directly. In this work, we investigate the extent to which general-purpose LLMs already possess this capability out of t
中文介绍 本文考察通用 LLM 在无需生成自由文本的情况下，直接输出预定义选项上类别概率分布的能力，即 Jev 式决策模型。研究其开箱即用的程度，并给出何时以及如何微调，使 LLM 输出可直接被软件系统使用。
👍 3
Face Image Quality Assessment determines the suitability of captured face images for automated face recognition (FR), a critical capability for reliable biometric systems. Existing state-of-the-art FR-integrated FIQA methods suffer from temporal instability: as the feature space evolves during train
中文介绍 人脸图像质量评估中，与识别模型集成的方法在训练时特征空间不断演化，导致评分时间不稳定。本文提出累积相对点间隔分数，通过学习相对排序式目标缓解该漂移，使 FIQA 评分在训练过程中保持稳定。
👍 2
Memory self-evolution uses task feedback to iteratively improve executable memory programs that store and retrieve information from past interactions. Existing approaches typically adopt holistic evolution, deriving revision directions from mixed feedback and judging progress by overall performance.
中文介绍 记忆自演化利用任务反馈迭代改进可执行记忆程序。现有做法多采用整体演化，混合反馈难辨改进方向。本文提出以能力为驱动的自演化方法，将反馈与具体能力单元对应，实现更精细的定位与修订，提升记忆程序效果。
👍 2
Real-world users often exhibit highly heterogeneous preferences over multiple objectives for LLM responses. A lightweight aligner can tailor these responses to individual preferences, but scarce user-specific feedback makes personalized training difficult. Learning shared initializations across user
中文介绍 真实用户偏好高度异质，个性化对齐又面临反馈稀缺。本文提出跨用户协同的个性化偏好对齐方法：在用户间学习共享初始化，再以少量用户专属反馈微调轻量对齐器，从而在数据不足时仍能拟合个体偏好。
👍 2
Autoregressive transformers remain comparatively weak for protein sequence and structure generation. We study the role of target representation: amino acid tokens encode residue identities without explicit contextual semantics, while backbone coordinates require a discrete representation in our fram
中文介绍 自回归 Transformer 在蛋白质序列与结构生成上相对薄弱。本文研究目标表示的作用：氨基酸 token 仅编码残基身份而缺少显式上下文语义，主链坐标在本框架中需离散化，据此学习潜在蛋白质语言以改进自回归生成。
@BrianRoemmele · 489.9K 粉丝 · 1.1M 阅 · 533 赞 · 85 转
You do not build a Grok Bot. You hire one. The difference is the whole article. A chatbot answers a question and forgets the room. A Grok Bot has a name, a job, a conversation that persists, and a
中文介绍 作者提出「雇佣 Grok Bot」而非「构建」的概念，来自其所谓首家零人类公司的实操经验。核心区别在于 chatbot 问答即忘，而 Grok Bot 拥有名字、岗位和持久对话记忆。角度偏概念梳理，适合关注 AI agent 长期记忆设计的读者。
@stevenmarkryan · 458.8K 粉丝 · 124.9K 阅 · 541 赞 · 84 转
Incredible things are happening in AI. What’s more incredible is that it’s still EXTREMELY early. As in, Day 0.000. Things are only going to get crazier from here. In case you’ve been living under a
中文介绍 播主 STEVEN MARK RYAN 的观察：AI 进展惊人但只是「第 0.000 天」，未来会更疯狂。核心论点是掌控算力者掌控一切。属于宏观趋势评论，无具体工具或数据，适合作为行业情绪参考。
@TryLiveAvatar · 1.1K 粉丝 · 123.4K 阅 · 692 赞 · 188 转
The best way to learn a language was always a patient tutor. For the first time, everyone can have one. Here's what that changes for learners, teachers and schools. The hardest part is speaking Ask
中文介绍 TryLiveAvatar 借 AI 数字人谈语言学习：人人可拥有耐心的一对一外教。指出口语是最难环节，并分析这对学习者、教师和学校带来的变化。属于 AI 教育场景的产品视角短文，无具体数字。
@mustafasuleyman · 1.2M 粉丝 · 99.2K 阅 · 587 赞 · 104 转
This is the prediction Nobel laureate Daron Acemoglu makes in the first issue of The Humanist Review, our new magazine exploring the future of AI, published by MAI. He argues we need to stop
中文介绍 Mustafa Suleyman 引用诺奖得主 Daron Acemoglu 的预测：未来 10 年 AI 只替代人类 5% 的工作，呼吁停止恐慌。来源是其新办杂志 The Humanist Review 首期，探讨 AI 未来。属于反焦虑的宏观观点分享。
@gregisenberg · 724.2K 粉丝 · 44.1K 阅 · 569 赞 · 53 转
For the FIRST TIME in HISTORY, you can describe a physical product in one sentence and, a week later, hold it in your hands, cut from metal and built to your exact measurements. That's VIBE
中文介绍 Greg Isenberg 提出「氛围制造」（Vibe Manufacturing）：用一句话描述实体产品，约一周后即可拿到按精确尺寸切割的金属成品。强调这是历史上首次实现，属于概念造词型内容，背后指向 AI 驱动的按需定制硬件供应链。
@onlyonealexia · 1.6K 粉丝 · 26.5K 阅 · 509 赞 · 53 转
A List of Web3, AI & Web2 Hackathons Compiled by @onlyonealexia There are a lot of hackathon lists that look impressive until you actually open the links. I compiled a list of ongoing and upcoming
中文介绍 博主整理了 10 月的 Web3、AI 与 Web2 黑客松清单，强调市面上多数列表点开链接才发现水分，这份是筛选过的进行中和即将开始的活动。适合想参赛练手或找项目的开发者按需取用。
@Hesamation · 90.6K 粉丝 · 21.4K 阅 · 618 赞 · 51 转
if you have an ADHD brain, you probably have six different goals, you get overwhelmed, and you miss days without working on any one of them. that's me. I love agents, harnesses, RL, LLM architecture,
中文介绍 针对兴趣过多、目标分散的「ADHD 大脑」，博主分享一套一小时内可搭好的 AI 系统，用来管理六个不同方向和容易断更的目标。作者自称热爱 agents、harness 和 LLM 架构，属于个人工作流 + agent 实践分享。
中文介绍 DeepMind 发布 EmbeddingGemma 2，这是一款开放、轻量的多模态嵌入模型，用于将文本与图像等数据映射至统一向量空间。
Atlassian and OpenAI are expanding their partnership to connect frontier models with enterprise knowledge and help teams plan, build, and deliver work.
中文介绍 Atlassian 与 OpenAI 扩大合作，将前沿模型接入企业知识库，帮助团队规划、构建并交付工作。
Jump Trading uses OpenAI to expand quantitative research. See how longer-running AI workflows combine multiple data sources with human review.
中文介绍 Jump Trading 使用 OpenAI 扩展量化研究，通过长时间运行的 AI 工作流整合多源数据并加入人工审核。
OpenAI publishes new results on open problems in mathematics from an internal frontier model and shares Lean proof formalizations and research details on GitHub.
中文介绍 OpenAI 公布内部前沿模型在数学开放问题上的新结果，并在 GitHub 分享 Lean 证明形式化及研究细节。
Learn how OpenAI and Ironclad are training and evaluating AI agents on complex contracting workflows to advance computer use for professional work.
中文介绍 OpenAI 与 Ironclad 合作，在复杂合同工作流上训练和评估 AI 智能体，以推进专业工作的计算机使用能力。
中文介绍 Hugging Face 博客介绍 Falcon-Emirati，展示大语言模型如何学习阿联酋方言、文化及语言细微差别。
A small win for US open source
中文介绍 Latent Space 报道 Reflection Beam 模型，参数规模为 501B-A23B，被视为美国开源模型的一次小胜利。
中文介绍 TLDR AI 汇总当日动态：Instinct 群聊功能、Reflection 的 501B 模型，以及 OpenAI 文本水印。
For all the data that AI systems continually amass and analyze, enterprise AI agents often suffer from a curious shortcoming: a lack of knowledge. More than data, knowledge is the understanding of what the data means in the context of individual organizations. AI agents need this understanding to re
中文介绍 MIT Tech Review 指出，企业 AI 智能体常因缺乏组织语境知识而受限，需将智能体与企业知识相连接。
How OpenAI is approaching text watermarking under EU rules. Learn where watermarks apply, how detection works, and why access starts with researchers.
中文介绍 OpenAI 说明其应对欧盟文本来源规则的方案，介绍文本水印的适用范围、检测方式及为何先向研究人员开放。
In 2026, the question for enterprise AI is no longer whether predictive models can outperform statistical forecasts—that argument is settled. The big question now is how to enable predictive systems to act on their own conclusions without drifting from business intent. The frontier has moved from pr
中文介绍 MIT Tech Review 探讨代理式 AI 时代的预测分析，核心问题转为如何让预测系统自主行动而不偏离业务意图。
OpenAI introduces a new visual ad format in ChatGPT and expands measurement tools, attribution partnerships, and brand suitability for advertisers.
中文介绍 OpenAI 在 ChatGPT 中推出新视觉广告格式，并扩展测量工具、归因合作及品牌适配性，面向广告主。
Over the summer I talked to the CEO of Springboards, a startup building an LLM that’s designed to come up with a wider variety of responses than its mainstream rivals do. At the start of the call, he said something that’s been stuck in my head since: “We often say that we’re a self-loathing AI…
中文介绍 MIT Tech Review 探讨公众对 AI 既反感又难舍的矛盾现象，并引用初创公司 Springboards CEO 的观察。
Yossi Matias, Vice President & Head of Google Research, explores how AI is beginning to reshape biology, infrastructure, manufacturing, and science, and why its greatest impact may come when it intersects with other fields. Step inside the newsroom with our MIT Technology Review editors for sharp an
中文介绍 MIT Tech Review 的 EmTech Future 2026 活动上，Google 研究负责人 Yossi Matias 探讨 AI 如何重塑生物、基础设施、制造与科学。
中文介绍 TLDR AI 汇总当日动态：Prime Inference、智能体数量增长，以及 Claude 学院。
美股四大指数 ETF 技术面整体偏多：SPY 现价 779.09，站上 SMA20/50/200（765.72/765.22/721.49），MACD 2.2981 高于信号线 1.2671 并触发金叉，距 52 周高仅 -0.32%；QQQ 759.66 同样多头排列，RSI14 69.9 逼近超买；NVDA、MSFT 均现 MACD 金叉，TSLA 5 日 +7.89% 但位于 SMA200 下方，META 则现死叉，分化初现。加密方面，恐慌贪婪指数 71 处于「贪婪」，总市值 2.91T（24h -3.45%），BTC 主导率 58.7%、ETH 11.2%；BTC 85327.02 与 ETH 2686.34、SOL 120.05 仍处 SMA 多头排列，但三者 1 日均为负，MACD 均低于信号线，BTC 距 52 周高 -31.84%、ETH -43.51%，反弹力度弱于美股。中概股技术面最弱：BABA、PDD、JD 虽因超跌触发 MACD 金叉，但仍为空头排列，0700.HK 现死叉且 5 日 -0.88%，四者距 52 周高在 -26% 至 -44% 之间。商品外汇端，DXY 101.97 逼近 52 周高、RSI14 68；10Y 收益率 5.27 现超买（RSI 70.5）；黄金 4179.5 空头排列、RSI 36.7 偏弱；WTI 原油 90.56 站上 SMA50/200；VIX 15.01 走低，波动率环境整体温和。
现价 5.27，高于 SMA20 的 5.10、SMA50 的 4.86 与 SMA200 的 4.46，呈多头排列，距 52 周高仅 -1.5%；RSI14 达 70.5 已进入超买区，MACD 0.1337 高于信号线 0.123，红柱延续。收益率强势与 VIX 回落至 15.01 同步，构成宏观端最显著的动量读数。
现价 15.01，低于 SMA20 的 15.89、SMA50 的 15.66 与 SMA200 的 18.05，呈空头排列；MACD -0.0692 已下穿信号线 -0.0335，形成死叉，1 日 -3.29%、5 日 -6.42%，逼近 52 周低（仅高 12.18%）。波动率持续压缩，说明当前市场未定价明显避险需求。
现价 738.88 虽高于 SMA20 的 702.17 与 SMA200 的 628.81，但 MACD 33.674 已低于信号线 34.801，触发死叉，是本组中唯一出现该信号的超大盘股；RSI14 为 64 处于中性偏热，距 52 周高 -5.25%。趋势字段标为中性，与 5 日 +3.25% 的反弹形成背离。
现价 4179.5 同时低于 SMA20 的 4305.36、SMA50 的 4370.04 与 SMA200 的 4551.52，呈空头排列；MACD -67.64 深于信号线 -49.01，绿柱扩张；RSI14 仅 36.7，逼近超卖。5 日涨跌 0%，距 52 周高 -25.18%。在美元指数 101.97 走强、10Y 收益率 5.27 上行的组合下，金价动量读数明显偏弱。
现价 78.4 贴近 SMA20 的 78.26，但低于 SMA50 的 83.68 与 SMA200 的 94.35，空头排列未改；MACD -1.7362 略高于信号线 -1.8197，刚现金叉，RSI14 43.8 位于中性偏弱区。1 日 +4.01% 与 5 日 -0.11% 相抵，属于方向未明的技术修复状态。
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
以上均为基于公开行情计算的技术指标读数与形态复述，过去走势不代表未来表现，指标信号亦可能快速反转。本内容仅供技术指标解读参考，不构成任何投资建议或买卖依据。
Ruling on Mount Pleasant coalmine shows ‘we cannot continue to dig up coal … and pretend the consequences have nothing to do with us’, group says A Hunter Valley community group has won Australia’s first high court case to consider climate change, in a ruling advocates say sets a binding national pr
中文摘要 澳大利亚猎人谷社区团体在高等法院关于Mount Pleasant煤矿的气候变化诉讼中胜诉，这是澳首例高等法院审理气候变化案件。该团体称裁决表明「不能继续挖煤却假装后果与我们无关」。
Experts and lawyers called the apparent recovery unprecedented, noting she is believed to be the only known person to survive a lethal injection.
中文摘要 美国死囚Christa Pike在注射死刑执行失败后苏醒并能说话，律师称其恢复情况前所未有。专家称她被认为是已知唯一在注射死刑中幸存的人。
The mountainous district is key to control of Mocha, the Red Sea port the Houthis seized last month.
中文摘要 也门部队在塔伊兹省发起攻势，试图夺回胡塞武装控制的山区al-Waziiya区。该区是控制红海港口摩卡的关键，胡塞武装上月占领了摩卡。
Australian adults again able to access the world’s most popular porn site after parent company introduces age checking via Apple devices Get our breaking news email, free app or daily news podcast Australian adults will be able to access Pornhub once again after the site brought in age-checking on A
中文摘要 Pornhub在因年龄验证规定封禁澳大利亚访问后恢复服务。其母公司通过苹果设备引入年龄检查，澳大利亚成年人可再次访问该网站。
At least 10 civilians were killed in West Kordofan and 5,500 displaced in Blue Nile as clashes widened on Tuesday.
中文摘要 苏丹布尔汉拒绝谈判，誓言从快速支援部队（RSF）手中夺回全部领土。周二西科尔多凡至少10名平民死亡，青尼罗河州5500人流离失所，冲突扩大。
More than one million people could lose life-saving aid by October as funding shrinks and arrivals from Sudan continue.
中文摘要 南苏丹援助危机加剧，粮食库存告急。因资金减少及来自苏丹的持续流入，到10月可能有超过100万人失去救生援助。
Saint-Denis Mayor Bally Bagayoko was tear-gassed after shielding student protesters from riot police.
中文摘要 法国圣但尼市长Bally Bagayoko在保护学生抗议者免受防暴警察冲击时遭催泪瓦斯喷射。
Yemeni forces claim they have tightened control over the highest peak in the Jabal Habashi district of Taiz.
中文摘要 也门部队声称已控制塔伊兹省Jabal Habashi区的最高山峰，加强了对该战略制高点的掌控。
Commissioner said Shenzhen Qingcheng, the app maker behind the HeyCyan app in the smartglasses, failed to respond to her inquiries Get our breaking news email, free app or daily news podcast The Australian privacy regulator has opened an investigation into the China-based software company behind the
中文摘要 澳大利亚隐私监管机构对中国深圳公司Shenzhen Qingcheng展开调查，该公司是Kmart智能眼镜中HeyCyan应用的开发商，未回应监管机构的问询。
Turkiye’s first domestically produced electric SUV, the Togg T10X burst into flames at a charging station.
中文摘要 土耳其首款国产电动SUV Togg T10X在充电站起火燃烧。
World Health Organization detects no sign of further outbreak since Darya Shipilova died last week The World Health Organization has said the risk of an epidemic in Russia is low after the sudden death of a lab technician who worked at a plague research institute in Siberia. Darya Shipilova, who was
中文摘要 世界卫生组织表示，西伯利亚瘟疫研究机构实验室技术员Darya Shipilova上周突然死亡后，俄罗斯爆发瘟疫疫情的风险较低，未发现进一步疫情迹象。
Government report finds the country could be hit with more frequent and intense heatwaves, as well as damaging storms and cyclones New Zealand could be hit with more frequent and intense heatwaves and catastrophic glacier melts in the near future, as greenhouse gas levels hit record highs, a new rep
中文摘要 新西兰政府报告显示，温室气体水平创历史新高，可能导致更频繁强烈的热浪、破坏性风暴和飓风，以及灾难性冰川融化。
The Bolsonaro family emerged triumphant in Brazil’s elections, mounting a surprising political resurrection in Latin America’s largest nation.
中文摘要 博索纳罗家族在巴西选举中获胜，在拉美最大国家实现令人意外的政治复活。
A sharp increase in human activity is affecting the way the Himalayan monal lives and communicates, studies show.
中文摘要 研究表明，人类活动的急剧增加正影响喜马拉雅虹雉的生活和交流方式，这种美丽的喜马拉雅鸟类正在改变其鸣叫声。
Frequent outages have fuelled protests against the government of interim Venezuelan President Delcy Rodriguez.
中文摘要 委内瑞拉频繁停电引发针对临时总统Delcy Rodriguez政府的抗议，公众愤怒日益加剧。
Taiwan is emerging as the stronger bet for equity investors due to its deeper linkages across the AI supply chain and a more upbeat earnings outlook. Bloomberg's Winnie Hsu breaks down the news. (Source: Bloomberg)
中文摘要 彭博报道，台湾因在人工智能供应链中联系更深、企业盈利前景更乐观，正取代韩国成为全球股市中更受投资者青睐的市场。彭博的Winnie Hsu对此进行解读。
Former Canadian Minister of International Trade and Milken Institute Senior Fellow Mary Ng outlines Canada's expanding trade push across Asia and the Indo-Pacific, emphasizing that deeper market access in energy, technology, and agriculture could reach 3 billion consumers and drive export growth. (S
中文摘要 加拿大前国际贸易部长、米尔肯研究所高级研究员Mary Ng表示，加拿大正扩大在亚洲及印太地区的贸易布局，若能深化能源、技术和农业领域的市场准入，可触达约30亿消费者。
As equity investors enter the final stretch of a pivotal year that’s propelled South Korea and Taiwan to the forefront of the global AI trade, the latter is emerging as the stronger bet.
中文摘要 彭博报道，在人工智能交易推动韩国和台湾股市年内领跑全球后，随着AI交易范围扩大，台湾正被视为比韩国更强的投资标的。
Two years after JLR relaunched Jaguar into a blizzard of controversy, it has unveiled its new electric car.
中文摘要 捷豹路虎两年前将捷豹品牌重塑引发争议后，如今正式发布新款电动汽车，标志着捷豹以纯电豪华品牌定位重新出发。
Asset Management One Co. plans to boost spending on talent over the next three years as rising Japanese interest rates broaden investment choices and push pension funds, university endowments and individuals to rethink how they allocate their money.
中文摘要 日本Asset Management One计划未来三年将人才相关支出提高30%。日本利率上升拓宽了投资选择，促使养老金、大学捐赠基金和个人重新考虑资产配置。
British carmaker unveils £130,000 electric car following radical rebrand
中文摘要 英国汽车制造商捷豹在品牌重塑后推出售价13万英镑的电动汽车，正式进军豪华电动车市场。
Gold held gains as increasing oil supplies from the Middle East and a decline in bond yields eased pressure on the US Federal Reserve to hike interest rates this month.
中文摘要 金价持稳，中东石油供应增加及债券收益率下降，减轻了美联储本月加息的压力，市场对加息的押注有所降温。
BHP Group will sell its Kambalda nickel concentrator plant and associated land in Western Australia to Gold Fields Ltd. for an undisclosed price.
中文摘要 必和必拓（BHP）将把位于西澳大利亚、已停产的Kambalda镍精矿厂及相关土地出售给Gold Fields Ltd.，交易价格未披露。
Samsung Electronics Co.’s earnings will test whether it can convince investors of its long-term outlook, an increasingly critical task as record profits have failed to reinvigorate the stock.
中文摘要 三星电子即将公布的业绩将考验其能否说服投资者相信长期前景。尽管利润创纪录，三星股价仍未回升，投资者希望看到盈利繁荣可持续的证据。
Ken Leech, the former co-chief investment officer at Western Asset Management Co., agreed to pay $3 million to settle a US Securities and Exchange Commission lawsuit claiming he’d cherry picked winning trades.
中文摘要 Western Asset Management前联席首席投资官Ken Leech同意支付300万美元，与美国证券交易委员会（SEC）就涉嫌挑选优质交易（cherry-picking）的诉讼达成和解。
Frozen yoghurt, which was a huge craze in the 2000s and 2010s, has made its return with multiple chains popping up across the wider UK.
中文摘要 冷冻酸奶在2000年代和2010年代曾风靡一时，如今在英国卷土重来，多家连锁店涌现，但每桶12英镑的价格能否支撑其持续热度仍有待观察。
Order of the Sinking Star is due out on Thursday and has roughly 1,500 puzzles for players to try
中文摘要 《时空幻境》（Braid）创作者Jonathan Blow推出新作《Order of the Sinking Star》，将于周四发售，内含约1500道谜题，他称这是「史上最大的解谜游戏」。
13 回复 · Apple 节点
24 回复 · Python 节点
89 回复 · 程序员 节点
9 回复 · 程序员 节点
6 回复 · 程序员 节点
5 回复 · 程序员 节点
59 回复 · 程序员 节点
11 回复 · 程序员 节点
10 回复 · 程序员 节点
17 回复 · Apple 节点
该源今日无内容。