BuilderIO/agent-native
TypeScript · ★ 6,258 · 🍴 564 · 📈 607 stars today
A framework for building agentic apps
中文介绍 BuilderIO 推出的 Agent-native 框架,专为构建以 AI Agent 为核心的应用程序而设计。它提供编排、工具调用与状态管理等基础设施,帮助开发者快速搭建多智能体协作系统。适用于需要集成 LLM 工作流的 SaaS 应用、自动化流程以及对话式平台。
TypeScript · ★ 6,258 · 🍴 564 · 📈 607 stars today
A framework for building agentic apps
中文介绍 BuilderIO 推出的 Agent-native 框架,专为构建以 AI Agent 为核心的应用程序而设计。它提供编排、工具调用与状态管理等基础设施,帮助开发者快速搭建多智能体协作系统。适用于需要集成 LLM 工作流的 SaaS 应用、自动化流程以及对话式平台。
HTML · ★ 25,895 · 🍴 1,781 · 📈 609 stars today
Scale computer-use 2.0 with open-source drivers, cross-OS fleets, and benchmarks for training, evaluation, and data generation.
中文介绍 Cua 是一个开源的 Computer Use 2.0 基础设施,支持跨操作系统设备集群管理,并提供训练、评估和数据生成的基准测试。面向构建自主桌面操作 Agent 的团队,用于在真实 GUI 环境中训练和测试视觉-动作模型。
TypeScript · ★ 18,179 · 🍴 2,237 · 📈 844 stars today
OpenStock is an open-source alternative to expensive market platforms. Track real-time prices, set personalized alerts, and explore detailed company insights — built openly, for everyone, forever free.
中文介绍 OpenStock 是一款开源的股票市场平台替代品,提供实时价格追踪、个性化价格提醒和公司详情查看等功能。采用完全开源模式,免费向公众开放,适合开发者、投资者和科技爱好者搭建自有行情工具。
Rust · ★ 7,955 · 🍴 533 · 📈 167 stars today
Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors
中文介绍 ai-memory 为 Agent 编码 CLI 提供长期记忆解决方案,支持在不同 Agent 厂商之间无缝交接上下文。解决多智能体协作中记忆丢失和信息断裂问题,适用于使用多种编码助手的开发团队。
Go · ★ 16,566 · 🍴 1,574 · 📈 460 stars today
Secure environments for developers and their agents
中文介绍 Coder 是一个为开发者及其 AI Agent 提供安全隔离工作环境的平台,支持通过 API 和 CLI 快速创建开发沙箱。适用于需要为 AI 编码助手提供受控执行环境的团队,保障数据安全与操作可追溯性。
Python · ★ 36,007 · 🍴 5,281 · 📈 424 stars today
中文介绍 Anthropic 官方面向金融服务业的解决方案项目,提供基于大语言模型的行业应用范例和最佳实践。涵盖风险控制、合规审查、客户服务等场景,适用于金融机构探索 AI 落地路径。
Rust · ★ 12,467 · 🍴 1,140 · 📈 32 stars today
🥧 Savoury implementation of the QUIC transport protocol and HTTP/3
中文介绍 Quiche 是 Cloudflare 出品的 QUIC 传输协议和 HTTP/3 实现,用于高性能网络连接。提供可靠的数据传输和拥塞控制,适用于构建低延迟的实时通信、CDN 加速和下一代 Web 基础设施。
Python · ★ 13,765 · 🍴 1,336 · 📈 169 stars today
MVT (Mobile Verification Toolkit) helps with conducting forensics of mobile devices in order to find signs of a potential compromise.
中文介绍 MVT(Mobile Verification Toolkit)是用于移动设备取证分析的开源工具集,帮助检测设备是否被恶意软件入侵或间谍工具感染。适用于安全研究人员、执法机构和移动安全团队进行威胁调查。
Python · ★ 8,612 · 🍴 1,596 · 📈 250 stars today
AutoClip : AI-powered video clipping and highlight generation · 一款智能高光提取与剪辑的二创工具
中文介绍 AutoClip 是一款 AI 驱动的视频高光提取与剪辑工具,可自动识别长视频中的精彩片段并生成短视频。适用于内容创作者快速制作二创视频、赛事精彩集锦及社交媒体短内容。
★ 104,395 · 🍴 4,464 · 📈 182 stars today
科技爱好者周刊,每周五发布
中文介绍 科技爱好者周刊由阮一峰编写,每周五发布,精选前沿技术资讯、工具和行业洞察。面向技术人员和科技爱好者,帮助快速了解全球技术趋势和新项目。
TypeScript · ★ 38,088 · 🍴 3,780 · 📈 394 stars today
Project NOMAD is an offline-first knowledge and education server. Wikipedia, thousands of books, courses, maps, and optional local AI, all running on hardware you own with no internet required.
中文介绍 Project NOMAD 是一款离线优先的知识与教育服务器,本地运行维基百科、数千本图书、课程、地图及可选本地 AI,完全无需互联网。适用于教育机构和需要数据主权的场景,保障隐私与可用性。
Rust · ★ 3,835 · 🍴 469 · 📈 50 stars today
OpenAI Codex 桌面端/CLI 的可视化管理工具,具有Provider/API 切换、会话同步、提示词注入、Skills/MCP 管理、TOML 配置可视化的跨平台工具。
中文介绍 Codex-X 是 OpenAI Codex 桌面端与 CLI 的可视化管理工具,支持多 Provider/API 切换、会话同步、提示词注入、Skills/MCP 管理以及 TOML 配置可视化。面向日常使用 Codex 的开发者,跨平台可用。
👍 126
Data agents aim to fulfill natural-language instructions over heterogeneous data, including tables, files, and databases. However, data agents face a challenging agent-data gap: heterogeneous data resides outside the agent, while the agent can access it (e.g., column names and file paths) only throu
中文介绍 EvoOntology 提出面向数据 Agent 的自演化本体层,用于弥合异构数据与 Agent 感知之间的 agent-data gap。方法让 Agent 在与数据交互中持续抽取并维护统一本体(表、文件、数据库的语义与关系),再据此执行自然语言指令,从而减少仅靠列名或路径检索的脆弱性。贡献是提供可增量更新、可共享的数据本体表示,为跨数据源的智能 Agent 提供可复用语义层。
👍 103
Training capable coding agents via reinforcement learning (RL) requires diverse tasks with reliable verifiers. Open-source codebases offer a rich source of such tasks, while existing methods typically rely on development artifacts such as issues and commits, limiting the range of tasks that can be e
中文介绍 CodeMidas 从开源代码库本身大规模构造可验证 RL 训练环境,用于训练编程 Agent。方法将仓库代码直接转化为多样化任务并生成可靠验证器,减少对 issue、commit 等开发工件的依赖,从而扩大任务范围与数据规模。贡献是为 agentic coding RL 提供可扩展、任务多样、验证可控的环境生成框架,提升编程 Agent 的训练泛化。
👍 92
Reusable skills give agents transferable procedural knowledge, making scalable acquisition essential for extending agents beyond prior experience. Existing methods face two limitations: trajectory-based synthesis requires interactions with specific environments, while document-derived skills may lac
中文介绍 该工作提出从大规模代码中 grounded 合成可复用技能,以扩展 Agent 的程序性知识。方法绕开依赖特定环境轨迹或文档的技能生成方式,直接从代码抽取可执行、可迁移的原子技能并规模化组织,解决现有 skill synthesis 覆盖窄、泛化差的问题。贡献是提供面向 agentic intelligence 的可转移技能库构建流程,使 Agent 在新任务上具备更强的先验能力。
👍 80
Humans can seamlessly adapt to both physical and digital worlds, suggesting that while a digital-to-real gap exists in embodiment, environment and task, human intelligence itself may transfer across this gap. This naturally raises a fundamental question: can the intelligence of vision-language model
中文介绍 论文研究能否将 VLM 的通用智能迁移到机器人控制。方法在数字到实体的具身鸿沟下,利用视觉语言模型的高层语义与推理能力作为规划或决策基础,桥接仿真/数字环境与真实机器人任务。贡献是提出从 VLM 到机器人控制的知识迁移框架,探索跨 embodiment 场景下视觉语言智能的可复用性,推动通用机器人控制。
👍 66
An LLM agent's capability is largely magnified by its harness, namely the prompts, control flow, tooling, memory, and context management surrounding the frozen backbone model. Recent methods increasingly automate this process by iteratively proposing and selecting component-wise edits of an agent ha
中文介绍 RRSI 针对 LLM Agent 的 harness(提示、控制流、工具、记忆与上下文管理)提出正则化递归自改进。方法迭代提出并筛选 harness 组件编辑,同时引入正则化约束避免过拟合局部实例,使 frozen backbone 在外部系统层面获得稳定增益。贡献是提供可自动化演进且鲁棒的 Agent harness 自优化框架,提升 Agent 能力上限。
👍 63
Computer-use agents (CUAs) have advanced along two separate lines: graphical interaction and software development through code and the command line. Real digital work requires both, interleaved rather than stacked end to end. We study hybrid CUAs that autonomously decide when to explore an interface
中文介绍 RecreationWorld 构建可规模化、可验证的混合电脑使用 Agent(C UA)环境。方法在同一任务中交替进行图形界面交互与代码/命令行软件开发,要求 Agent 自主判断何时使用 UI 何时使用编程接口。贡献是提供跨两种操作范式的训练与评测基准,推动 C UA 从单一模式向更接近真实数字工作的混合能力发展。
👍 36
Retrieval-Augmented Generation (RAG) systems over enterprise knowledge bases must ingest heterogeneous document formats -- PDFs, Word documents, presentations, and scans -- whose content is locked inside complex visual layouts, multi-column pages, and dense tables. Rule-based extraction and OCR dest
中文介绍 D-RAC 提出面向企业文档的检索感知分块方法,先对 PDF 做规范化再转换为多模态 Markdown,以保留复杂版式、多栏与密集表格结构。方法减少 RAG 系统中纯规则或 OCR 抽取造成的结构丢失,使分块与后续检索更一致。贡献是提供通用文档摄取流程,提升企业知识库 RAG 在异构格式下的检索质量与可维护性。
👍 31
We define OmniVChat (Omni Video Chat) as the task of native audio-visual dialogue between a user and an omni model. In OmniVChat, omni models directly and simultaneously receive audio and video from a user and return text. The user's query is embedded in the audio and video, without a separate text
中文介绍 OmniVChat 定义原生音频-视觉对话任务:omni 模型同时接收音频与视频输入并返回文本,用户查询嵌入在多模态流中而非独立文本。论文构建该任务的合成数据、评测基准与训练方案,推动端到端视听对话能力。贡献是补齐 omni 模型在多模态交互范式上的评测与训练缺口,促进实时、自然的视听理解与响应。
👍 31
Professional design requires any-color control: the ability to specify an object's target color with any 24-bit hex value for image generation and editing. Prior work has explored color generation, editing, and colorization, but often relies on dedicated color representations or specialized inferenc
中文介绍 Paint-Anything 面向图像生成与编辑的统一任意颜色控制,支持以任意 24 位 hex 值指定目标颜色。方法避免依赖专门颜色表示或特殊推理流程,将颜色控制统一进生成/编辑框架,提升专业设计中的色彩精确性。贡献是提供单一模型同时覆盖颜色生成、编辑与着色的通用能力,简化工作流并增强可控性。
👍 25
Educational foundation models must solve problems, understand curriculum structure, diagnose learner difficulties, and provide appropriate instructional support. Existing educational language models often focus on either problem solving or tutoring, with training mixtures organized by source or task
中文介绍 OmniEdu 提出面向学习与教学的综合教育基础模型,要求同时具备解题、理解课程结构、诊断学习者困难与提供教学支持的能力。方法采用统一的训练混合而非按来源或任务割裂组织,弥合现有教育模型仅覆盖部分功能的问题。贡献是提供覆盖教学全流程的开放基础模型,推动教育场景中的个性化指导与诊断。
👍 24
Professional graphic design is a long-horizon agentic task in which structured, editable artifacts emerge from many interdependent actions, yet outcomes admit no reliable programmatic oracle. We introduce a continual adaptation framework in which a frozen frontier model operates professional design
中文介绍 Designer-RSI 将专业平面设计建模为长程 agentic 任务,采用持续适应框架,让冻结的前沿模型从用户流量演化过程性记忆,逐步积累可编辑、结构化产物的生成策略。方法应对设计成果难以编程验证的挑战,通过交互反馈更新内部经验。贡献是面向 agentic graphic design 的在线进化方案,提升复杂设计任务的自适应能力。
👍 22
Recent advances in video generative models have enabled high-fidelity, temporally coherent video generation. However, these models often struggle to satisfy prompts requiring specialized knowledge, specific identities, physical consistency, or ordered events. In this paper, we present VideoGen-Agent
中文介绍 VideoGen-Agent 通过强化学习增强视频生成 Agent 的提示遵循能力,重点改善需要专业知识、特定身份、物理一致性与事件顺序的生成场景。方法将生成质量与指令约束纳入奖励信号,训练 Agent 更好地编排视频生成行为。贡献是提升高保真、时序一致视频在复杂语义约束下的可控性,推动 agentic 视频生成落地。
👍 20
We present onPanda, an interactive tool for efficiently annotating LLM alignment data and agent trajectories. onPanda adopts token-level correction as its core interaction: while reading a model response, the annotator locates the first inappropriate token and either picks a substitute from the mode
中文介绍 onPanda 是用于高效标注 on-policy 对齐数据与 Agent 轨迹的交互式工具,核心交互是 token 级修正:标注者在阅读模型输出时定位首个不当 token,并从候选中选择替换。方法降低对齐数据标注成本,提升修正粒度与效率。贡献是面向 LLM 与 Agent 训练的高效标注工作流,帮助更快构建高质量对齐数据集。
👍 18
Manipulation policies must know which objects matter and where they are, yet the pretrained backbones that current robot foundation models build on, from language in vision-language-action models (VLAs) to video generation in world-action models (WAMs), do not directly require this metric grounding,
中文介绍 Grounded Action Model 将 3D grounding 作为机器人基础模型的关键环节,强调操作策略必须明确识别相关对象及其位置。方法补足当前 VLA 与 WAM 预训练骨干缺乏直接度量级 grounded 能力的短板,强化空间感知与对象定位。贡献是提出以 3D grounding 为核心的机器人基础模型框架,提升真实操作任务的可靠性。
👍 14
Repository-level software engineering (SWE) comprises heterogeneous task categories, whose progress under pooled agentic reinforcement learning can be uneven: gains in some categories coincide with regressions in others, while aggregate resolution obscures these changes. Motivated by this category s
中文介绍 论文研究仓库级软件工程(SWE)任务中的类别感知迭代专家训练。方法关注池化 agentic RL 下不同任务类别进展不均、总指标掩盖回归的问题,提出按类别迭代优化专家策略。贡献是为 agentic SWE 提供更稳定的训练机制,缓解任务间冲突,提升各类别的一致改进与整体可靠性。
👍 14
Business intelligence (BI) is a cornerstone of enterprise decision-making and is widely used by enterprise users in software such as Power BI and Tableau. In traditional BI workflows, users need to prepare data by (1) identifying relevant tables, (2) performing data transformations, and (3) building
中文介绍 BI-Agent 与 BI-Bench 面向端到端商业智能自动化。方法覆盖传统 BI 工作流中的关键步骤:识别相关数据表、执行数据变换、构建分析视图,并推动 Agent 自主完成这些流程。贡献是提供自动化 BI 的 Agent 框架与评测基准,降低 Power BI、Tableau 等场景中的数据准备与分析门槛,提升企业决策效率。
👍 14
We present MintAct, a family of vision-language models that unifies UI grounding, multi-step navigation across mobile, desktop, and web, and visual tool use, trained at 2B, 4B, and 8B scales. Through careful design of our environments, data, and training recipes, MintAct models match the performance
中文介绍 MintAct 是统一数字环境视觉 Agent 的视觉语言模型族,规模覆盖 2B、4B、8B。方法在同一框架中整合 UI grounding、跨移动/桌面/网页的多步导航与视觉工具使用,通过环境、数据与训练配方设计提升泛化。贡献是提供轻量化、多平台统一的视觉 Agent 方案,在保持小模型成本的同时匹配更大模型性能。
👍 11
Existing approaches to persona simulation with Large Language Models (LLMs) mostly rely on shallow character descriptions that fail to sustain coherent character behavior across extended interactions. We introduce Deep Persona, a psychologically grounded, three-layered architecture that organizes pe
中文介绍 Deep Persona 提出基于心理学原理的三层角色架构与评测框架,用于提升 LLM 角色扮演的长期一致性。方法从浅层人设描述转向结构化人格建模,使角色在扩展交互中维持稳定行为。贡献是提供心理可解释的角色模拟架构与评价体系,增强 LLM 在长期角色扮演与模拟场景中的可信度。
👍 11
Agent harnesses, the external systems that mediate model-environment interaction, can substantially improve agent performance, but their gains remain tied to the harness at deployment. Because the best harness varies across domains, instances, and models, a general-purpose agent must either settle f
中文介绍 Harness-Zero 提出以 Agent 作为 Harness 进行蒸馏,把外部 harness 带来的能力增益内化到模型自身。方法针对不同领域、实例和模型的最佳 harness 差异问题,减少对部署时外部系统的依赖。贡献是提供通用 Agent 的 harness 蒸馏方案,使模型在保持性能的同时获得更强的自主性。
👍 10
Vision-Language-Action (VLA) policies achieve strong performance in robotic manipulation but remain brittle once execution deviates from nominal trajectories. We propose CARE (Corrective Atomic Robotic Execution), a framework that improves recovery by learning from failures encountered during execut
中文介绍 CARE 提出经验引导的原子纠正执行框架,用于提升 VLA 策略在执行偏离标称轨迹时的恢复能力。方法从执行失败中学习可复用的纠正动作,将失败经验转化为原子级修复策略。贡献是增强机器人在真实操作中的鲁棒性,减少因偏离而导致的任务失败,推动 VLA 策略从执行到自恢复的闭环提升。
@obie · 13.1K 粉丝 · 157.2K 阅 · 738 赞 · 98 转
I’m excited about a future that I suspect will be very hard on many long-time friends and certain people that I love. The same technology that is allowing me to be more successful than ever is
中文介绍 obie 从个人角度反思 AI 对人际关系的冲击,认为当前技术虽让自身更成功,但未来可能对长期朋友和特定人群造成巨大压力,提出「肉代」式的生存焦虑与时代判断。
@DhravyaShah · 62.1K 粉丝 · 120.2K 阅 · 509 赞 · 16 转
Instinct has taken the world by storm over the last two weeks - it's one of the best imessage assistants I've used. as with every product, I decided to reverse-engineer Instinct's memory to find out
中文介绍 Dhravya Shah 通过逆向工程拆解 iMessage 助手 Instinct 的 memory 机制,剖析其近期爆火背后的设计细节,揭示其数据存储与召回逻辑,为同类产品提供参考。
@twostraws · 120.7K 粉丝 · 77.1K 阅 · 504 赞 · 42 转
I use both Codex and Claude for app development work, both with the highest-tier subscription. Here's how I've found the two compare, plus some tips for success… TL;DR I use both agents extensively,
中文介绍 twostraws 对比 Codex 与 Claude 在 App 开发中的表现,基于自身双高级订阅的实战经验,总结两款 agent 的优势差异与使用建议,给出可落地的开发工作流优化技巧。
@omarsar0 · 320.6K 粉丝 · 73.5K 阅 · 523 赞 · 60 转
As you may have seen recently, Jev (a generalist System One model) took Twitter by storm. Many misleading examples are everywhere, but they don't really show Jev's real capabilities. Hence, we
中文介绍 omarsar0 面向新手撰写 Jev 入门指南,指出近期 Twitter 上大量示例存在误导,未能展现 Jev 作为通用 System One 模型的真实能力,系统介绍其用法与适用场景。
@trycua · 19.0K 粉丝 · 59.5K 阅 · 501 赞 · 39 转
What we've been building with decision models at Cua - by @ddupont808 and @francedot Over the last few days we've deprioritized some other shiny objects because we've all been Jev-crazy. TypeSafe
中文介绍 Cua 团队分享其在 decision model 领域的技术积累,解释 Jev 与 System One 模型对 computer use 未来的启示,强调 TypeSafe 等工程方向,展现团队技术路线。
@RohOnChain · 72.5K 粉丝 · 57.7K 阅 · 518 赞 · 54 转
I will break down the exact framework to build a millisecond speed 24/7 HFT trading system with Jev, along with the exact resources that helped me personally. Let's get straight to it. Bookmark This -
中文介绍 RohOnChain 详解用 Jev 构建 24/7 毫秒级高频交易系统的完整框架,涵盖资源配置与技术路径,分享其本人实际使用的工具与资料,定位实战向教程。
@matthewcanham · 10.7K 粉丝 · 57.1K 阅 · 516 赞 · 45 转
If you’ve been hanging out on the internet the past few days, you’ve probably come across Jev, a new AI model that’s taken the AI community by storm. Jev is not a traditional LLM. It’s built for
中文介绍 matthewcanham 以通俗方式向普通用户解释 Jev,强调其并非传统 LLM,而是面向快速决策的新 AI 类型,帮助非技术群体理解其定位与近期热度来源。
@LangChain · 265.9K 粉丝 · 46.6K 阅 · 504 赞 · 68 转
By Daniel Shea and Seán Roche Key takeways: Jev is a fundamentally different kind of evaluator. It returns typed answers directly instead of generating text like an LLM judge. Jev was dramatically
中文介绍 LangChain 提出 Jev-as-a-Judge 用于 agent 评估,指出 Jev 直接返回带类型答案而非生成文本,评估效率显著优于传统 LLM judge,为评测体系提供新方案。
@manish_fp · 846 粉丝 · 44.1K 阅 · 512 赞 · 59 转
I have read 10+ articles, watched 5 hours of content to understand what actually Jev is and then chose to write this. It is not a chatbot, It’s a whole new kind of AI, more like a classifier which
中文介绍 manish_fp 整理 Jev 101,说明它不是聊天机器人而更接近分类器类 AI,梳理阅读 10+ 篇文章、观看 5 小时内容后的核心认知与可构建方向,适合入门者。
@_avichawla · 77.5K 粉丝 · 37.8K 阅 · 593 赞 · 80 转
Everything you need to turn an open-source LLM into a fast, local decision engine without retraining it. It covers next-token scoring, fixed choices with probability distributions, SGLang, and a
中文介绍 _avichawla 分享将开源 LLM 改造为本地快速决策引擎的方法,无需重训,涉及 next-token scoring、固定选项概率分布与 SGLang,给出可复现的本地化方案。
@murmur_arc · 983 粉丝 · 11.8K 阅 · 553 赞 · 137 转
Prologue — No LLMs, Only Neurons This project is called murmur, and it is alive right now on Arc mainnet (Chain ID 5042), settling in real USDC. To date the flies have completed more than 23,000
中文介绍 murmur 项目在 Arc 主网(Chain ID 5042)运行,采用真实 USDC 结算的果蝇群体链上文明模拟,无 LLM 仅用神经元,已完成超 23,000 次行为记录。
crowning a new Chinese frontier lab
中文介绍 小米发布开源权重模型MiMo-V2.6-Pro,参数规模1T(激活42B),训练成本约300万美元,被视为当前领先的开源前沿模型,标志着中国实验室在开源大模型领域取得突破。
中文介绍 MLX工具omlx的创建者与维护者Jun Kim加入Hugging Face,将负责支持MLX社区相关工作。MLX是面向Apple Silicon的机器学习框架,此次合作有望强化其在Hugging Face生态中的支持。
The definitive Jev podcast with its lead creator.
中文介绍 Latent Space播客发布与TypeSafe AI首席执行官Diogo Almeida的对谈,主题为「Jev」系列中面向生产环境而非「全能」的System One模型设计思路,是该播客核心创作者的深度访谈。
中文介绍 Hugging Face博客发布研究文章,提出借鉴物理学中伊辛优化问题的方法对大语言模型进行块级剪枝,将模块移除建模为伊辛模型问题,探索更高效的LLM压缩路径。
OpenAI is working with an independent Advisory Group on Mathematics and Artificial Intelligence to guide the review and communication of emerging AI results.
中文介绍 OpenAI成立数学与人工智能独立顾问小组,负责评估并沟通新兴AI研究成果,以推动该领域研究的透明度和可信度。
With GPT-6 Astra, Higgsfield AI makes video ad creation easier for small businesses and brings new creative tools to market faster.
中文介绍 Higgsfield AI借助GPT-6 Astra模型,在一日内推出多项新视频功能,使中小企业的视频广告制作更便捷,并加速创意工具的市场落地。
Our 15-month investigation into death and surveillance along the US-Mexico border began with a simple question: Why did so many people die near government surveillance towers meant to help track and apprehend them? This story is part of Dying on Camera, a collaboration between MIT Technology Review
中文介绍 MIT科技评论与媒体合作公布历时15个月的调查,制作了美国边境「虚拟墙」沿线的死亡综合地图,揭示监控塔周边发生大量死亡事件的异常情况。
MIT Technology Review today published our investigation into how many people have died near the “virtual wall” of surveillance towers that the US government has installed along the US-Mexico border. We found cases of people who walked undetected through areas surveilled by advanced, AI-enabled tower
中文介绍 MIT科技评论在调查美墨边境「虚拟墙」监控塔失效问题后提出四项政策建议,指出有人类从先进监控系统覆盖区域未被察觉地通过,呼吁改进监控与救援机制。
When José Morales Bernal crossed the border into the United States on April 8, 2024, the day before his 32nd birthday, it should have triggered a chain of technological alerts and human responses. As he walked through the desert in southern New Mexico that morning, he was within range of three surve
中文介绍 MIT科技评论调查指出,尽管美国在边境监控上投入数十亿美元,仍存在大量死亡未被及时阻止的案例,例如2024年4月一名男子在新墨西哥州南部沙漠中被监控塔覆盖区域穿行后遇难。
She had only walked for a couple of hours, and already she was lost. It was early afternoon on Sept. 14, 2025, when 30-year-old Graciela Gómez Hernández crossed the border from the eastern edge of Tijuana into Southern California, sending voice messages to her mother and sister as she walked. This s
中文介绍 MIT科技评论报道,2025年9月一名30岁女性Graciela Gómez Hernández从蒂华纳东端徒步进入南加州后死亡,事发时不远处即设有政府监控摄像头,暴露边境监控系统严重失灵。
OpenAI outlines a path to shared global AI standards, calling for coordinated evaluation, reporting, and governance to improve safety.
中文介绍 OpenAI发布关于构建下一阶段AI标准的框架建议,呼吁建立全球协同的评估、报告与治理机制,以提升AI安全性。
Explore new OpenAI Academy learning paths for employees, developers, leaders, educators, and students to build and demonstrate practical AI skills.
中文介绍 OpenAI Academy新增面向员工、开发者、领导者、教育者和学生的多条学习路径,帮助各类用户构建并展示实用AI技能。
Using GPT-5.6, V7 turns scattered company files into context agents can use to complete complex, source-linked work.
中文介绍 OpenAI发布V7版本,基于GPT-5.6将公司分散的文件转化为可供AI代理使用的上下文,实现具备来源溯源能力的复杂任务处理,为AI代理提供机构记忆。
中文介绍 Hugging Face博客发布tokenizers v1版本更新,重点介绍编码、解码及扩展性的实测表现与优化。
中文介绍 TLDR AI简报汇总当日要闻:Muse推出连接器、Meta发布SAM 3.1模型、Gemini相关工具遭公司滥用等问题。
美股方面,SPY、QQQ、AAPL、MSFT、NVDA、GOOGL 呈多头排列且多数接近 52 周高点,QQQ 近 1 日 +2.77%、近 5 日 +4.55%,MACD 金叉,动量偏强;VIX 近 5 日 -13.9%、RSI 45.2,恐慌度收敛但 MACD 仍死叉,避险情绪温和。加密方面,恐慌贪婪指数 78 处于极度贪婪区间,总市值 2.92 万亿美元(24h -0.5%)、BTC 主导率 58.9%、ETH 11.4%;BTC RSI 71.6 超买且 MACD 金叉,ETH、SOL 同样金叉但 RSI 67-69 仍偏强,短线动量向上但短线过热。中概股 BABA、PDD、JD、0700.HK 均处空头排列,BABA 近 5 日 +5.97%、MACD 金叉但趋势仍 bearish,PDD RSI 40.6 偏弱,板块整体在修复但均线压制明显。商品外汇端,黄金 RSI 46.1 中性、MACD 绿柱;WTI 原油近 5 日 -15.46%、MACD 死叉动能转弱;USD/CNY 6.70、RSI 24 超卖且空头排列;美元指数与 10Y 美债收益率均接近 52 周高,宏观端整体呈现风险偏好回升叠加避险资产分化的格局。
近 1 日 +11.34%、近 5 日 +11.36%,现价 741.25 已显著高于 SMA20 624.18 与 SMA50 608.82,MACD 29.82 远高于信号线 18.95,动量极强;但 RSI 77.8 进入超买,52 周高距 -3.81% 且 SMA200 624.63 被价格反超,短线技术上偏上行但过热。
现价 86022.72,近 5 日 +12.59%,SMA20 79350.77 < SMA50 74122.34 < SMA200 70672.55 多头排列,MACD 2310.27 高于信号 1719.41 金叉;RSI 71.6 超买,距 52 周高 -31.84%,叠加加密恐慌贪婪指数 78 极度贪婪,短线动量仍偏上行但需警惕过热。
近 1 日 -6.59%、近 5 日 -15.46%,价格 89.47 跌破 SMA20 93.5,MACD 3.5585 低于信号 4.2188 形成死叉,绿柱扩大;尽管仍高于 SMA50 87.11 与 SMA200 81.02 呈多头排列,但短线动能明显走弱,动量向下概率较高。
现价 79.3,SMA20 81.57、SMA50 85、SMA200 96.18 三线自上而下空头排列,MACD -2.16 与信号 -2.037 仍为负且绿柱存在,RSI 40.6 偏弱,52 周高距 -43.12%,近 5 日 -0.11% 几乎横盘,技术结构持续承压。
现价 6.70,SMA20 6.71、SMA50 6.74、SMA200 6.84 呈空头排列,RSI 24 进入超卖,MACD -0.0102 几乎贴近信号 -0.0103;距 52 周低仅 0.22%,近 5 日 -0.14%,技术上空头趋势中但超卖信号显示下行动能已接近阶段性衰竭,方向中性。
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
本报告基于公开技术指标读数与近期行情数据撰写,仅供技术指标解读参考,不构成任何投资建议;过去走势不代表未来表现,市场存在不可预测风险,请独立判断。
Co-ordinated suicide attacks killed 270 people across several churches and hotels in Colombo in 2019.
中文摘要 斯里兰卡法院将宣判2019年复活节连环自杀爆炸案,该袭击致270人死亡。
Australian scientists found unequal impact of the climate crisis would not be reversed even if implicit goal of the 2015 Paris agreement is achieved Sign up for climate and environment editor Adam Morton’s free Clear Air newsletter here The unequal impact of the climate crisis on the world’s poorest
中文摘要 澳大利亚科学家发现,即使实现《巴黎协定》净零目标,热带地区仍将是气候危机不均等影响最后的受益者。
Disappearance of jet components sparks concern some of the program’s technology secrets could have been exposed Follow our Australia news live blog for latest updates Get our breaking news email, free app or daily news podcast The Department of Defence is examining how F-35 fighter jet parts bound f
中文摘要 澳大利亚调查绑定美国的F-35战斗机零部件在香港失踪事件,担心技术机密可能已外泄。
Without an agreement sanctions on nearly 3,000 entities accused of supporting the war on Ukraine will lapse at the end of the day Meanwhile, the Kremlin said it considered all western sanctions on Russian individuals to be illegal and that they should all be revoked. Kremlin spokesperson Dmitry
中文摘要 欧盟在对俄制裁修改上陷入僵局,涉及近3000个实体的制裁将于当日到期;克里姆林宫称所有西方对俄制裁非法应予撤销。
More than 74,000 people remained under evacuation orders in the Tokyo region as authorities warned of landslides and floods.
中文摘要 台风杜荣给日本带来破纪录降雨,致5人死亡;东京地区逾7.4万人仍在避难指令下。
The US president said in an interview ahead of the United Nations meeting that he was considering ‘wiping Iran out’ The UK has agreed to provide defensive air support to Saudi Arabia to help counter Houthi fighters and restore stability in the Middle East while driving down the cost of living at hom
中文摘要 伊朗革命卫队警告若美国升级冲突将「改变战争地理」;英国同意为沙特提供防御性空中支援以应对胡塞武装。
The UK Foreign Office says it is supporting the man's family and working with local authorities.
中文摘要 一名英国男子在科莫湖为救青少年跳水后失踪,英国外交部称正协助其家属并与当地当局合作。
Footage appears to show a male in the area carrying a firearm, while officials say the attacker has been captured.
中文摘要 土耳其一所高中附近发生枪击事件,11人受伤,袭击者已被警方捕获。
An initiative to show that LFW is "bigger than London" launched this year - but will it make a difference?
中文摘要 伦敦时装周今年发起「超越伦敦」倡议,试图淡化地域标签,但其效果仍存疑问。
特习会还未开始,台湾问题已成焦点。据知情人士向路透社透露,中国领导人习近平本周访问华盛顿期间,预计将敦促美国总统特朗普根据1982年的一份中美联合公报停止对台军售,并指出习近平可能会在参观美国国家档案馆时提出这一要求。
中文摘要 据知情人士向路透社透露,习近平本周访美预计将当面敦促特朗普依据1982年中美联合公报停止对台军售,可能在参观国家档案馆时提出。
Sri Lanka were bowled out for 69 runs, 10 less than Mandhana's belligerent match-winning innings as India retained gold.
中文摘要 板球亚洲杯决赛中,印度以147分优势击败斯里兰卡,后者仅得69分,印度保住金牌。
Here’s a look at what the 193-member United Nations does and how it works as leaders convene in New York for the annual General Assembly.
中文摘要 美国《纽约时报》介绍联合国及其193个成员国的运作机制,正值各国领导人在纽约出席年度联大。
Police arrest suspect in shooting near high school as injured are rushed to local hospitals for treatment.
中文摘要 土耳其西部一所高中附近发生枪击,8人受伤并送医,嫌疑人已被警方逮捕。
After a sprawling world tour, Praxis, a New York-based, crypto-backed “digital nation,” has signed a deal to build a physical city in Uruguay.
中文摘要 纽约加密支持的「数字国家」Praxis在环球巡演后与乌拉圭达成协议,计划在当地建设实体城市。
Ms. Dion, the French Canadian diva, has embarked on her first solo tour of France since the pandemic. The French are swooning.
中文摘要 法国加拿大歌星席琳·迪翁展开疫情以来首次法国个人巡演,首场在巴黎引发歌迷热烈反响。
Also in today’s newsletter: North Korea tests new hypersonic weapon and the risks of the yen carry trade
中文摘要 Financial Times报道,一桩据称由克里姆林宫支持的伪造方案成功蒙蔽了全球多家银行。同日刊新闻还提及朝鲜试射新型超音速武器,以及日元套息交易相关风险。
Market expectations for earnings at European luxury-goods makers are overly optimistic given softening demand conditions and buying trends, according to analysts at RBC.
中文摘要 RBC分析师认为,市场对未来欧洲奢侈品制造商的盈利预期过于乐观,理由是需求明显趋弱、消费者购买力下降,现有预测或难以实现。
Donald Trump Jr., Chase Herro and a group of their business partners descended on Manhattan’s Jazz at Lincoln Center concert hall last year to promote an expansive vision for World Liberty Financial, the Trump family’s cryptocurrency venture.
中文摘要 特朗普家族的加密货币项目World Liberty Financial在曼哈顿Jazz at Lincoln Center推介扩张蓝图。尽管整体计划不及预期,项目仍通过Donald Trump Jr.、Chase Herro等合作伙伴获取可观收益。
Brent fell toward $99 a barrel, after tumbling over the previous four sessions. Oil erased earlier gains on a report that Iran could reopen the Strait of Hormuz if the US lifted a blockade of its ports. US President Donald Trump is set to address the United Nations General Assembly in New York later
中文摘要 布伦特原油价格回落至每桶约99美元,此前连续四个交易日下跌。消息显示伊朗若美方解除对其港口封锁,或重新开放霍尔木兹海峡;特朗普即将出席联合国大会。
Nvidia Corp.’s sinking stock valuation is sending a warning signal about the chipmaker’s prospects for maintaining its booming profit growth.
中文摘要 英伟达股价估值持续下滑,被认为是一个警示信号,意味着这家芯片制造商可能难以维持此前的高利润增速。
"Bloomberg: The Asia Trade" brings you everything you need to know to get ahead as the trading day begins in Asia. Bloomberg TV is live from Tokyo and Sydney with Shery Ahn and Haidi Stroud-Watts, getting insight and analysis from newsmakers and industry leaders on the biggest stories shaping global
中文摘要 Bloomberg Asia Trade节目指出,中美继续在纽约进行谈判,Meta发布的Muse Agent消息推动芯片板块上涨。节目由东京和悉尼演播室连线报道。
National Stock Exchange of India Ltd. paid about 1.86 billion rupees ($19.4 million) in fees to the 20 banks that worked on its $2.4 billion initial public offering, according to an exchange filing.
中文摘要 印度国家证券交易所(NSE)为24亿美元IPO项目向参与承销的20家银行支付约18.6亿卢比(约1940万美元)费用,相关数据来自交易所文件披露。
A summer of wildfires and heatwaves underlines the threats to the country's economy.
中文摘要 BBC报道,西班牙在多个领域应对气候变化,近期夏季持续的森林大火和热浪凸显了该国经济面临的气候威胁。
Germany is leading European drive to reduce dependence on America for space-based intelligence, says US group’s CEO
中文摘要 美国卫星制造商Planet Labs在德国柏林开设工厂,CEO表示德国正引领欧洲减少对美依赖、发展自主太空情报能力,军工需求激增。
Traditional banks want online-only operators to fund hubs required to mitigate branch closures
中文摘要 FT报道,传统银行与金融科技公司围绕英国高街支付中心(high street hubs)资金来源展开争夺,银行希望仅在线上运营的公司参与为网点关闭配套项目出资。
Borrowing was higher than expected in August as inflation piles pressure on the government ahead of the Budget.
中文摘要 英国财政部数据显示,8月政府借贷规模超出预期,在通胀压力下,财政部长在预算案发布前面临更大的财政平衡压力。
Grupa Kety SA may refinance bridge debt for its biggest-ever takeover with its first conventional bond issue, as the Polish aluminum-products maker accelerates its push beyond eastern Europe.
中文摘要 波兰铝制品企业Grupa Kety SA计划通过首次常规债券发行,为其迄今最大并购(约4.5亿欧元Metra交易)再融资,并加速向东欧以外扩张。
9 回复 · 程序员 节点
4 回复 · 程序员 节点
4 回复 · Apple 节点
6 回复 · 程序员 节点
16 回复 · Apple 节点
4 回复 · 程序员 节点
9 回复 · 程序员 节点
6 回复 · Apple 节点
8 回复 · Python 节点
26 回复 · Linux 节点
该源今日无内容。