Go · ★ 31,845 · 🍴 2,258 · 📈 3,231 stars today
Fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in multi-language ruleset (NPE, thread-safety, XSS, SQL injection), OpenAI & Anthropic compatible.
中文介绍 阿里巴巴开源的代码审查工具，采用确定性流水线与 LLM Agent 混合架构，支持行级精准评论和多语言规则集，针对 NPE、线程安全等常见缺陷进行检查。已在阿里大规模生产环境中验证，适合工程团队集成到 CI/CD 流程中，提升代码质量和审查效率。
JavaScript · ★ 7,226 · 🍴 421 · 📈 927 stars today
A coding-agent skill for multi-phase security audits with independently verified, machine-readable findings
中文介绍 Cloudflare 推出的编码代理安全审计技能，支持多阶段安全审计流程，产出可独立验证的机器可读审计结果。主要面向使用 AI 编码代理的团队，用于自动化安全扫描、漏洞发现与合规检查，降低人工审计成本。
C · ★ 35,031 · 🍴 3,679 · 📈 1,546 stars today
Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦
中文介绍 纯 C 语言实现的推理引擎，零外部依赖，可在现有硬件上运行前沿 MoE 模型，专家层从磁盘流式加载。体积小巧、无需 GPU 集群，适合个人开发者或边缘设备本地部署大模型推理，兼顾性能与硬件兼容性。
Swift · ★ 5,590 · 🍴 267 · 📈 1,179 stars today
Tinycast — a tiny, fully native macOS launcher, hotkeys, and clipboard history.
中文介绍 面向 macOS 的轻量级原生应用，整合启动器、全局快捷键和剪贴板历史功能。完全原生实现，无额外依赖，适合追求效率和简洁体验的 macOS 用户，替代或补充 Raycast、Alfred 等工具的核心功能。
TypeScript · ★ 54,380 · 🍴 6,790 · 📈 417 stars today
The open-source AI voice studio. Clone, dictate, create.
中文介绍 开源 AI 语音工作室，支持语音克隆、语音听写和内容创作。通过 AI 模型生成或转换语音，适合播客创作者、有声内容制作人和开发者，用于快速生成个性化语音素材或自动化语音处理任务。
Swift · ★ 13,340 · 🍴 1,587 · 📈 547 stars today
中文介绍 命令行工具，用于 iPhone 相关设备管理或配置操作。适合开发者和高级用户通过终端批量或自动化管理 iOS 设备，具体功能需参考仓库文档。
Python · ★ 24,282 · 🍴 2,922 · 📈 110 stars today
Open source repository of plugins primarily intended for knowledge workers to use in Claude Cowork
中文介绍 Anthropic 开源的插件集合，专为 Claude Cowork 设计，面向知识工作者场景。提供文档处理、数据分析、协作编辑等增强功能，适合企业用户在办公环境中通过 Claude 提升知识管理效率。
TypeScript · ★ 7,309 · 🍴 1,087 · 📈 778 stars today
Ever® Gauzy™ - Open Business Management Platform (ERP/CRM/HRM/ATS/PM) - https://gauzy.co
中文介绍 Ever Gauzy 是开源商业管理平台，整合 ERP、CRM、HRM、ATS 和项目管理功能。适合中小企业统一处理财务、客户、人力和项目流程，减少多系统切换，支持数据集成与自动化工作流。
Rust · ★ 30,857 · 🍴 3,209 · 📈 58 stars today
Anki is a smart spaced repetition flashcard program
中文介绍 Anki 是经典的间隔重复闪卡程序，基于艾宾浩斯遗忘曲线优化记忆复习节奏。支持多端同步和自定义卡组，适合语言学习者、医学学生和备考人群进行长期记忆训练。
Java · ★ 77,794 · 🍴 8,606 · 📈 1,059 stars today
Ghidra is a software reverse engineering (SRE) framework
中文介绍 美国国家安全局开源的软件逆向工程框架，提供反汇编、反编译和数据分析工具。支持多种架构和文件格式，适合安全研究员、逆向工程师分析二进制程序、漏洞挖掘和恶意代码研究。
TypeScript · ★ 145,499 · 🍴 23,478 · 📈 165 stars today
Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.
中文介绍 Anthropic 推出的终端 AI 编码代理，能理解代码库结构，执行常规任务、解释复杂代码和辅助 Git 工作流。适合开发者集成到日常编码流程，提升开发效率和代码理解能力。
Python · ★ 50,596 · 🍴 4,812 · 📈 260 stars today
We write your reusable computer vision tools. 💜
中文介绍 Roboflow 推出的计算机视觉工具库，提供可复用的目标检测、跟踪、分类等实用工具。以模块化设计简化 CV 项目开发，适合计算机视觉工程师快速搭建应用原型，降低重复造轮子的成本。
Rust · ★ 4,414 · 🍴 273 · 📈 1,017 stars today
Turn your coding agents into research agents
中文介绍 将编码代理转变为研究代理的开源框架，支持自动化文献调研、实验设计和论文生成。适合科研人员和实验室利用 AI 加速研究流程，从数据收集到论文撰写实现端到端辅助。
TypeScript · ★ 109,722 · 🍴 14,041 · 📈 120 stars today
The Postgres development platform. Supabase gives you a dedicated Postgres database to build your web, mobile, and AI applications.
中文介绍 Supabase 是开源 Postgres 开发平台，为 Web、移动和 AI 应用提供专用数据库、认证、存储和实时功能。一站式替代 Firebase，适合全栈开发者快速构建数据驱动的应用，无需管理底层数据库。
Python · ★ 2,549 · 🍴 184 · 📈 80 stars today
All in one plugin for Hermes Agent ⚚ the coding intelligence, a long-term memory system and model optimized workflow packages
中文介绍 Hermes Agent 的一体化插件包，整合编码智能、长期记忆系统和模型优化工作流。提升 AI 代理的持续学习和任务执行能力，适合需要长期记忆和复杂工作流编排的智能体应用开发者。
Go · ★ 25,289 · 🍴 3,476 · 📈 1,197 stars today
Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.
中文介绍 腾讯开源的 LLM 知识平台，可将原始文档转换为可查询的 RAG 系统、自主推理代理和自维护 Wiki。适合企业构建内部知识库和智能问答系统，实现文档知识自动沉淀与检索。
Python · ★ 5,766 · 🍴 741 · 📈 367 stars today
claude-red is a curated library of offensive security skills designed for the Claude skills system. Each skill is a structured SKILL.md file that primes Claude with expert-level methodology for a specific attack surface — from SQLi to shellcode, EDR evasion to exploit development.
中文介绍 针对 Claude 技能系统整理的攻击性安全技能库，每个技能以结构化 SKILL.md 形式定义专家级方法论。适合安全研究人员和渗透测试人员通过 Claude 执行红队演练、漏洞评估和攻防模拟任务。
Python · ★ 9,368 · 🍴 1,008 · 📈 332 stars today
YuE2: frontier music generation with symbolic planning, zero-shot covers, and agentic music editing.
中文介绍 YuE2 是前沿音乐生成模型，支持符号规划、零样本翻唱和智能体音乐编辑。适合音乐创作和 AI 音乐研究场景，通过多模态技术实现风格迁移、旋律生成和歌词同步创作。
JavaScript · ★ 95,452 · 🍴 10,111 · 📈 658 stars today
Production-grade engineering skills for AI coding agents.
中文介绍 面向 AI 编码代理的生产级工程技能集合，涵盖代码审查、重构、测试和部署等实战能力。适合团队将企业工程规范封装为可复用技能，提升 AI 代理在企业级项目中的可靠性和专业性。
TypeScript · ★ 68,373 · 🍴 7,384 · 📈 112 stars today
Autonomous coding agent as an SDK, IDE extension, or CLI assistant.
中文介绍 Cline 是自主编码代理，可作为 SDK、IDE 扩展或 CLI 助手使用。支持多步骤任务规划、代码生成和终端操作，适合希望将 AI 编码能力深度集成到开发工具链的工程师和团队。
👍 424
As AI agents become participants in the development of their successors, they reshape both the production of intelligence and the role of human researchers. We introduce Atria Dawn Preview, a foundation agentic language model designed for scientific research and engineering workflows, with the goal
中文介绍 提出 Atria Dawn Preview，面向科研与工程流程的基座 agentic 语言模型，探索 AI agent 参与后续模型开发时的角色转变，目标是推动科学发现与工程实现中的智能体自主能力。
👍 302
In this work, we present ZGCM-1, a fully open 7B dense foundation model trained from scratch with extreme data, system, and algorithmic efficiency. ZGCM-1 is founded on a core premise: compact models cannot passively memorize the open web, but can overcome parametric capacity limits by coupling deli
中文介绍 ZGCM-1 是全开放 7B 稠密基础模型，从零训练，主打数学与 agentic 检索能力，强调数据、系统与算法极致效率。核心理念是紧凑模型通过精细耦合外部检索弥补参数量限制，而非被动记忆开放网络。
👍 287
Language models may need to internalize information that arrives over time and retain it through many subsequent updates. To study this challenge, we introduce long-horizon memorization, a setting in which a model learns 100 query-answer tasks through continual supervised fine-tuning without retaini
中文介绍 提出 long-horizon memorization 评测设定，模型通过持续 SFT 学习 100 个问答任务且不保留过往数据。研究多种持续学习机制的组合方式，验证它们可叠加以实现跨长期更新的稳定记忆。
👍 107
Foundation models, alongside advances in learned game-world models, are reshaping AI across the game lifecycle. Beyond playing games, recent systems model players and game dynamics, support design and development, adapt player-facing experiences at runtime, and evaluate resulting artifacts. Yet thes
中文介绍 综述基础模型时代的游戏 AI 进展，涵盖玩家建模、游戏动力学模拟、设计与开发辅助、运行时体验适配与产物评估。指出这些方向仍存在统一框架与评测标准的缺口。
👍 91
Realtime spoken interaction demands deep reasoning, prompt responses, and fluid turn-taking. We present StepAudio 3 Realtime, an audio-language foundation model organized around a continuous listen-converse-think-act loop. Deep Perception captures rich acoustic cues to interpret user intent, while S
中文介绍 StepAudio 3 Realtime 是音频-语言基座模型，围绕持续 listen-converse-think-act 循环设计。Deep Perception 捕捉声学线索以理解用户意图，支持实时语音交互中的深度推理、快速响应与流畅轮次切换。
👍 86
Recursive self-improvement (RSI) enables AI systems to turn experience and feedback into persistent changes that improve both their capabilities and the process of future improvement. We first use the Headroom-Closed Index (HCI) to reveal the problems of existing LLMs, then introduce the RSI concept
中文介绍 提出真正的递归自我改进（RSI），用 Headroom-Closed Index 揭示现有 LLM 不足，再引入 RSI 概念与框架，使系统能把经验与反馈转化为持久改进，同时提升自身能力和后续改进流程。
👍 70
We introduce StepAudio 3 Music, a large-scale, long-form music generation model that supports explicit musical planning and open-domain text-controlled generation. The StepAudio Music Tokenizer represents audio as a 50-Hz stream from a 65536-entry single codebook, using semantically informed self-su
中文介绍 StepAudio 3 Music 是大规模长音频音乐生成模型，支持显式音乐规划与开放域文本控制。其 Music Tokenizer 将音频表示为 50Hz 单码本 65536 条目流，采用语义引导的自监督训练。
👍 70
Digital agents must often adapt to new environments whose interfaces, tools, and failure modes are not fully captured by pretrained models. We introduce RSIAgent, a training-free multi-agent framework for recursive self-improvement through autonomous memory construction. RSIAgent coordinates curricu
中文介绍 RSIAgent 是无训练多智能体框架，通过自主记忆构建在新环境中实现递归自我改进。协调课程生成、工具探索与记忆管理等智能体，使数字 agent 能适配接口、工具和失败模式未知的场景。
👍 64
The KV cache is a primary bottleneck for Transformer decoding: its memory footprint and cache-read traffic grow with sequence length. Grouped-query attention (GQA) reduces this cost by sharing key-value heads, but still stores both a key and a value at every step. We introduce Grouped Value Attentio
中文介绍 提出 Grouped Value Attention，通过按需重建 key 来进一步压缩 Transformer 解码的 KV cache。相较 GQA 只存 value 头，在保持精度的同时降低显存占用与 cache 读取开销。
👍 46
Video diffusion models are stochastic and hard to control: precise content often requires repeated sampling without guaranteed success, and long-horizon scenes drift in appearance, interactions, and temporal coherence. Agentic visual creation provides explicit references, editable 3D scenes, or exec
中文介绍 LynnReal-Omni 是原生多模态视频生成模型，面向 agentic 视觉工作流。相比随机采样的扩散模型，它接收显式参考、可编辑 3D 场景或执行指令，提升长程内容一致性与可控性。
👍 29
Orthrus is a hybrid autoregressive-diffusion architecture that accelerates autoregressive language-model inference by generating multiple tokens in parallel while using a frozen autoregressive backbone. Its central claim is that an intra-model consensus mechanism enables lossless speculative decodin
中文介绍 研究 Orthrus 中无损 speculative decoding 的真实性，聚焦数值精度对同模型共识机制的影响。分析发现浮点误差可能破坏其声称的无损加速，揭示混合自回归-扩散架构的精度瓶颈。
👍 22
Interactive video world models must maintain broad scene context under camera motion while producing high-fidelity observations with low latency. Existing approaches face a representation trade-off: perspective models operate on local views and must preserve off-screen content over long rollouts, wh
中文介绍 AlayaVista 提出从全景状态到视角视频的流式世界建模，解决交互式视频世界模型在相机运动下维持广域场景上下文与低延迟高保真观察的矛盾。用全景表示兼顾视野与局部精度。
👍 17
We introduce and release ScienceBuddy, an interactive scientific research workspace that brings continually improving scientific agents into researchers' everyday workflows. ScienceBuddy supports researchers in carrying out scientific tasks while transforming their requests, feedback, and execution
中文介绍 发布 ScienceBuddy 交互式科研工作空间，实现递归嵌套的自我改进。科学 agent 在协助研究者完成任务的同时，将请求、反馈和执行经验转化为持续改进，融入日常科研工作流。
👍 17
Native 3D generators turn one image into a single mesh. TRELLIS.2 and its peers deliver high-fidelity non-watertight geometry with materials, but the output is one fused object, while downstream work such as editing, rigging and simulation operates on part-level assets. A naive idea is to run a 3D s
中文介绍 Kaininja 将原生 3D 生成器扩展到部件级。原生生成器输出单一融合网格，下游编辑、绑定与仿真需部件资产；该方法改进分解流程，生成带材质的高保真、可编辑的多部件网格。
👍 15
Embodied navigation requires agents to interpret visual observations, accumulate spatial knowledge, and execute actions to follow instructions or locate objects. Training-based methods face generalization challenges, while training-free methods exploit multimodal large language models (MLLMs) but of
中文介绍 HarnessVLN 通过 agent harness 统一无训练的具身导航。利用多模态大语言模型处理视觉观察、积累空间知识并执行动作，在指令跟随与寻物任务上实现免训练泛化。
👍 14
Computer-use agents increasingly interact with browsers, terminals, file systems, and external services, introducing safety risks that emerge through runtime behavior rather than generated content alone. Existing guard models target static prompts and responses and are poorly suited to agent executi
中文介绍 HazardAuditor 面向 computer-use agent 的安全审计，从可执行威胁出发识别运行时行为风险。现有 guard 模型仅针对静态提示与响应，该方法适配 agent 与浏览器、终端、文件系统等交互。
👍 12
We demonstrate that a general-purpose agent can directly drive a physical robot throughout task execution without any task-specific or environment-specific training. We introduce Agent as Policy (AGP), which places task planning and execution under the agent's control. Given a task and a robot inter
中文介绍 提出 Agent as Policy（AGP），通用 agent 无需任务或环境特定训练即可直接驱动物理机器人完成任务。任务规划与执行均由 agent 控制，根据指令和机器人接口实时决策动作。
👍 12
Large language model (LLM) agents allocate test-time compute adaptively as they revise solutions, use tools, explore alternatives, and decide when to stop. This test-time strategy makes it difficult to measure how agent performance scales. We study open-ended tasks that provide continuous scores for
中文介绍 通过 Elo-per-token 分析研究 LLM agent 的 test-time 策略，理解 agent 如何分配算力于修订解、使用工具、探索替代方案及决定停止。在连续评分的开放任务上量化性能与算力的扩展关系。
👍 9
We demonstrate that training LLMs with RL does not improve performance equally across a dataset. RL shows large improvements on easy problems that an LLM is already good at solving, but small improvements on hard problems. We call this the Matthew Effect in RL for LLMs, after the phenomenon of cumul
中文介绍 揭示 RL 训练 LLM 中的「马太效应」：易题提升大、难题提升小。提出永不放弃的训练策略，针对难题设计机制以改善 RL 在困难样本上的收益不均问题。
👍 7
Backdoor poisoning attacks add poisoned examples to otherwise-clean finetuning data, pairing a trigger with a target behavior that the model learns to produce when the trigger appears. Existing evaluations typically fix the number of poisoned examples and sample them at random from a candidate pool.
中文介绍 提出学习选择毒集以增强 LLM 后门攻击的方法。现有评估固定毒例数量并随机采样，该方法学习挑选更具影响力的毒例，使模型在触发词出现时更稳定地产出目标行为。
@mustafasuleyman · 1.1M 粉丝 · 204.3K 阅 · 502 赞 · 86 转
AIs are not conscious. They do not feel, experience, or suffer. However, there’s a growing movement in support of model welfare; the idea that we might soon owe them a duty of care. I think this
中文介绍 Mustafa Suleyman 讨论「模型福利」（model welfare）概念：AI 无意识，但业界正兴起一种观点，认为人类可能对前沿模型负有关怀义务。他提出需要认真审视这一伦理新议题。
@LingoAITech · 411.4K 粉丝 · 156.5K 阅 · 524 赞 · 227 转
From the Semantic Web to the Human-Centric Agentic Economy Data3.0 + Finance3.0: Sovereign Data, Personal AI, Open Agents and Programmable Value Abstract Artificial intelligence is changing the role
中文介绍 梳理 AI-native Web3.0 架构愿景：从语义网走向以人为核心的 Agentic Economy，涵盖 Data3.0 主权数据、个人 AI、开放 Agent 与可编程价值，探讨 AI 如何重塑数据与金融协作模式。
@jack · 12.1M 粉丝 · 153.5K 阅 · 2.2K 赞 · 304 转
the frontier is the edge of what we know. no company owns what comes next. i want more people to be able to advance it. i favor open releases that people can examine, use, and improve together without
中文介绍 Sam Altman 表态支持开放前沿模型：前沿是知识边界，不应由单一公司垄断。他主张发布可被他人检查、使用并共同改进的开放模型，以让更多人参与推进 AI 进步。
@zachlloydtweets · 20.4K 粉丝 · 123.1K 阅 · 500 赞 · 47 转
The software factory approach (a closed agentic loop that runs in the cloud) is growing in popularity, but it can be daunting to adopt. In this post I’ll go through “crawl, walk, run” steps for making
中文介绍 介绍「软件工厂」模式落地路径：云端运行的闭环 agentic loop 正流行但上手门槛高。作者用 crawl、walk、run 三步拆解企业采用软件工厂的渐进式实践方法。
@mustafasuleyman · 1.1M 粉丝 · 117.2K 阅 · 526 赞 · 123 转
AI must be subordinate and always in service of people. Today we're publishing a Code of Conduct for governing MAI Models as they approach the frontier. https://microsoft.ai/news/mai-code-of-conduct/ This is a first draft
中文介绍 Microsoft 发布《人文主义 AI 行为准则》初稿，治理其 MAI 前沿模型：AI 必须从属并服务于人类。这是 Microsoft 在模型接近前沿能力时推出的具体治理框架。
@EricSimons · 28.7K 粉丝 · 84.0K 阅 · 500 赞 · 94 转
This weekend the leaders of the largest closed AI labs agreed the industry should slow down. I want to talk about the other direction: acceleration to the frontier. The gap is closing For two years,
中文介绍 Eric Simons 反向发声：闭源大实验室周末刚同意「减速」，他主张加速开放权重模型走向前沿。过去两年差距在缩小，开放阵营需要更快迭代，而非放慢脚步。
@JacquelineSYC19 · 1.0K 粉丝 · 72.4K 阅 · 501 赞 · 47 转
At the time of writing, Artie is a team of 17. Every one of us works alongside Hermes, an open-source AI agent harness from @NousResearch. It runs on a single physical machine in Germany, has a
中文介绍 Artie 团队 17 人全员使用 Nous Research 的开源 AI agent harness Hermes 协作，部署在德国的单台物理机上。分享了如何用一套 harness 支撑整个团队的日常工作流。
@GoogleCloudTech · 1.3M 粉丝 · 61.0K 阅 · 515 赞 · 34 转
A friend asked me an inconsequential question at work and I had an agent research my recent chats and docs and it found the right answer but it cost me $38!!!. That was a bad ROI. Sure my time is
中文介绍 对比 Claude Fable 5.1 与 Gemini 3.8 Flash 团队协作效果：一次让 agent 检索近期聊天记录和文档回答无关紧要的问题，却花了 38 美元，ROI 极差，揭示调用成本隐患。
@donnfelker · 29.9K 粉丝 · 45.9K 阅 · 513 赞 · 53 转
If you’re like me, you likely thought software engineering was a safe and secure job … In fact, I never thought I’d have to worry about a job again as long as I stayed on top of my craft… I was dead
中文介绍 Donny Felker 感叹软件工程正在「蒸发」：曾经以为技术扎实就永远安全，如今 Agentic Builder 崛起，传统工程角色面临根本性重塑，从业者需要重新思考职业路径。
@GoogleCloudTech · 1.3M 粉丝 · 43.8K 阅 · 505 赞 · 60 转
When developers first work on harness engineering for agentic coding systems, they often fall into the same trap: they run common end-to-end benchmarks like Terminal-Bench and DeepSWE, watch a
中文介绍 解析 Agentic 编码系统的 harness engineering：开发者常陷入只跑 Terminal-Bench、DeepSWE 等端到端基准的陷阱。文章讲如何评估、迭代并为 AI 编码 Agent 加护栏。
@ShaneLegg · 82.3K 粉丝 · 40.6K 阅 · 585 赞 · 86 转
We are on the cusp of a profound transformation. Today’s AI systems have impressive capabilities and the rapid pace of innovation suggests we’re now approaching artificial general intelligence (AGI),
中文介绍 Google DeepMind 成立 DeepMind Institute，宣称当前 AI 创新节奏已逼近 AGI，需要设立专门研究机构应对这一深度转型，强调前沿能力的系统性治理与长期研究。
@GoogleAIStudio · 203.4K 粉丝 · 35.5K 阅 · 506 赞 · 51 转
New Gemini Audio models are available for developers to build more intelligent conversational experiences via the Gemini API and Google AI Studio. Today, we released new Gemini Live models in the
中文介绍 Google AI Studio 推出新的 Gemini Audio 模型，支持开发者基于 Gemini API 构建实时语音应用，包括 Gemini 3.8 Live 与 3.5 Transcribe，用于打造更智能的对话体验。
@GoogleAIStudio · 203.4K 粉丝 · 35.1K 阅 · 585 赞 · 47 转
Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking are our most advanced live dialogue models yet. Major upgrades in intelligence and parallel reasoning make it more intuitive to collaborate and
中文介绍 Google 发布 Gemini 3.8 Live 与 3.8 Live Extended Thinking：最新实时对话模型，智能与并行推理能力显著提升，让多轮协作更自然，是其在 live 对话方向的又一次重大升级。
@Hrushikeshhhh · 420 粉丝 · 30.2K 阅 · 563 赞 · 37 转
TLDR; understanding the importance of evals, and how they can make yours and your agent life easier without spending exponential tokens and time to complete a task. pre - Agent = Model + Harness.
中文介绍 作者提倡「就算一开始很糙，也要先写 evals」：Agent = Model + Harness，理解 evals 的重要性能让任务和 agent 工作更顺畅，避免指数级 token 与时间浪费，早写早受益。
We sit down with AIUC’s CEO on their Series A!
中文介绍 AIUC首席执行官接受访谈，介绍该公司为可追责的智能体提供承保服务。该公司已完成A轮融资，旨在为高风险AI系统构建保险机制，降低智能体部署的财务与法律责任风险。
OpenAI shares a framework for tracking, investigating, and disclosing model misalignment, alongside six reports of unexpected or concerning model behavior.
中文介绍 OpenAI发布模型错配报告框架，用于追踪、调查和披露模型行为偏差，并公开了六份关于模型出现异常或令人担忧行为的报告，以提升AI安全透明度。
OpenAI and AARP are bringing free, hands-on ChatGPT workshops to 1,000 older adults across 10 U.S. cities to build practical AI skills safely.
中文介绍 OpenAI与AARP合作，在美国10个城市为1000名老年人提供免费的ChatGPT实操工作坊，帮助其安全掌握实用AI技能，拓展日常生活中的应用。
Explore new AI-powered advertising experiences from OpenAI, including Sponsored Agents, tools for marketers, and integrations with HubSpot and Shopify.
中文介绍 OpenAI推出AI驱动的广告新方案，包括赞助智能体、营销工具，以及与HubSpot和Shopify的集成，助力广告主重塑数字营销体验。
The AI boom is becoming a materials challenge. As AI pushes computing into new territory, the materials behind that infrastructure are becoming just as crucial as the algorithms running on it. Semiconductors and data centers are approaching physical limits around performance, thermal management, ele
中文介绍 MIT科技评论指出，AI发展正面临材料层面的挑战。半导体和数据中心在性能与功耗方面逼近物理极限，基础设施材料的重要性已与算法本身同等关键。
Learn how ChatGPT Work and Codex analytics help teams understand AI usage and spend, identify training needs, and connect adoption to business outcomes.
中文介绍 OpenAI介绍ChatGPT Work与Codex分析工具，帮助团队追踪AI使用情况和支出，识别培训需求，并将AI采用率与业务成果关联，提升投资回报评估能力。
congrats to TypeSafe!
中文介绍 TypeSafe公司发布Jev模型，定位为「系统一模型」，仅负责决策、分类、路由和评分任务。相比小型前沿LLM，Jev速度超100倍、成本降低超200倍。
New OpenAI Economic Research shows how workers use AI beyond traditional roles and which new activities become recurring parts of their work.
中文介绍 OpenAI经济研究报告显示，员工使用AI不仅限于传统岗位，还涌现出新型常态化工作活动，AI正在重塑知识工作者的日常流程与角色定位。
中文介绍 本期AI日报涵盖三项更新：TypeSafe的Jev模型、Google推出Periodic Neon工具，以及Gemini 3.8 Live的发布与实时功能。
Good Start Labs trained an AI on a railroad game — and one version improved at financial research. The difference was the training design.
中文介绍 Good Start Labs通过铁路游戏训练AI，其中采用特定训练设计的版本展现出金融研究能力提升，说明游戏环境习得的技能可迁移至现实任务，关键在于训练设计方式。
Listen to the session or watch below Employees at the world’s leading AI labs are saying there’s a real possibility that advanced AI could destroy humanity. Are they right? Or is this more scaremongering and hype? Watch a conversation unpacking AI extinction fears: where they come from, whether they
中文介绍 MIT科技评论圆桌讨论AI灭绝风险：多家领先AI实验室员工认为先进AI可能带来毁灭性威胁。讨论辨析这些担忧源于真实风险还是恐慌营销，梳理风险来源与应对思路。
中文介绍 Google DeepMind发布Gemini 3.8 Live与Gemini 3.8 Live Extended Thinking两款模型，增强实时交互与扩展思维推理能力。
中文介绍 IBM研究探讨AI智能体任务完成后的一致性问题：即使某次任务成功，智能体在重复执行时表现可能不稳定，研究聚焦于提升智能体行为的可靠性与一致性。
When Jessica Wachter, a finance professor at the University of Pennsylvania’s Wharton School, wanted to assess AI’s impact on the economy over the next few years, she faced a long list of business and technical uncertainties. So she started with what she calls a “remarkable fact” that is not in ques
中文介绍 宾大沃顿商学院教授Jessica Wachter分析AI万亿美元基建投资的经济风险，指出在技术与商业高度不确定性下，评估AI经济影响需从关键事实出发，警惕投资泡沫。
Pacing gathers pace.
中文介绍 第三方AI评估标准AEF-1正式形成，xAI、OpenAI和Anthropic均签署认可。该标准旨在统一第三方模型安全评估流程，推动AI治理框架标准化。
整体市场呈现分化格局。美股方面，SPY(-0.90%)与QQQ(-0.63%)短期承压、RSI 均在44-45中性偏弱区，但AAPL、MSFT、NVDA、META仍维持SMA20>SMA50>SMA200多头排列，其中META RSI 71.6已超买、GOOGL刚出现MACD金叉。加密市场恐慌贪婪指数=50中性，总市值2.62T(24h -1.49%)、BTC主导率58.4%、ETH 11.3%，BTC-USD、ETH-USD、SOL-USD均线多头排列但均位于SMA20下方、动量边际走弱。中概股集体走弱，BABA、PDD、0700.HK呈空头排列，RSI 33-44偏弱，JD RSI 33.6接近超卖。商品外汇端，黄金GC=F(-2.97%)回调至SMA20下方，WTI原油CL=F单日-4.23%但仍多头排列RSI 66.4；USDCNY=X RSI 21.1超卖+接近52周低+空头排列。宏观方面^TNX RSI 74.1超买逼近52周高、DXY金叉接近52周高、VIX升至17.71(+7.59%)，风险偏好边际收敛，整体技术面呈高位波动、风险资产内部分化、避险与美元走强的组合。
META当前价673.31，5日+9.75%领涨成分股，MACD 19.794 远高于信号线9.355，柱状为强正红柱，SMA20 595.32、SMA50 603.96、SMA200 623.66被价格显著站上形成多头排列，但RSI 71.6已进入超买区间，短线动量极强但存在回调风险。
AAPL现价332.41，5日+5.12%，价格站上SMA20 319.27与SMA50 318.84并显著高于SMA200 285.51，触发多头排列信号；MACD 4.3044>信号2.7078为红柱扩张，RSI 14 61.3处于多头但未超买的健康区间，距52周高-3.53%接近突破位。
CL=F现价101.35，5日+5.52%显著反弹，价格大幅高于SMA20 91.36、SMA50 85.25、SMA200 80.25维持多头排列，MACD 5.2129>信号3.8396红柱扩张，RSI 66.4强势但未超买；单日-4.23%属于高位回调，52周低+84.34%涨幅巨大，趋势动能仍偏上行。
BABA现价107.27，5日-4.78%、单日-1.79%持续走弱，价格显著低于SMA20 116.10、SMA50 117.77、SMA200 134.04构成空头排列；MACD -3.1421低于信号-2.2852绿柱扩张，RSI 36.8偏弱尚未到超卖，距52周高-44.32%下行幅度大。
USDCNY=X现价6.70，价格已击穿SMA20 6.72、SMA50 6.74、SMA200 6.84呈空头排列，RSI 21.1深度超卖、接近52周低，同时出现MACD金叉(-0.0106 > -0.0107)信号，多空信号相互冲突，趋势下行但短线反弹概率上升，定义为中性。
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
本报告仅基于公开行情数据的技术指标读数，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，技术指标存在滞后性与失效风险，请读者独立判断。
Rates were hiked in a unanimous decision despite fierce opposition from President Donald Trump, who had called for a cut.
中文摘要 美国联邦储备系统一致投票决定加息，为三年来首次加息。尽管总统特朗普强烈反对并曾要求降息，美联储仍作出加息决定。
Follow the day’s news live Get our breaking news email, free app or daily news podcast Barnaby Joyce says Labor is copying and pasting One Nation’s homework because the measures we know so far around blocking the families of international students and preventing visa hopping have some similarity to
中文摘要 澳大利亚政治快讯：议员Lambie在议会发表情绪化演讲后，欢迎工党在退伍军人辅助医疗上限问题上让步。Barnaby Joyce则指责工党照搬一方党的移民限制政策。
Trump, who has frequently clashed with Canada's PM, threatens to cut trade with the European Union in response to the proposal.
中文摘要 欧盟领导人支持加拿大成为「准成员国」的计划。美国总统特朗普与该提案多次与加拿大总理交恶，威胁将对欧盟削减贸易以回应此提案。
Warning comes as forecast for nation’s economic growth downgraded amid concerns over rising debt and inflation Follow our Australia news live blog for latest updates Get our breaking news email, free app or daily news podcast The International Monetary Fund has called on Australia’s federal and stat
中文摘要 国际货币基金组织敦促澳大利亚政府削减开支，并下调该国经济增长预期。报告警示债务上升和通胀压力，另一次加息可能逼近。
The bloc’s offer of “associate member” status to Ottawa shows how U.S. allies are pulling closer together. But their plans face significant obstacles.
中文摘要 欧盟向渥太华提出「准成员国」地位，显示美国盟友正加速抱团。但相关计划面临重大障碍，且特朗普对此表示不满。
Indigenous authorities issue condolences after two adults and a child swept away by floodwaters in southwestern state.
中文摘要 美国新墨西哥州纳瓦霍族领地发生洪水，两名成年人和一名儿童被冲走溺亡。原住民当局对此表示哀悼，事件发生在该州西南部地区。
President Trump says the US is nearing the end of the conflict with Iran and claims direct talks with Tehran.
中文摘要 伊朗战争直播：总统特朗普称美国正接近结束与伊朗的冲突，并声称已与德黑兰展开直接会谈。美伊关系动态持续受到关注。
Evan Speigel says it would be "an important step forward for the industry" in an interview with the BBC.
中文摘要 Snapchat负责人Evan Spiegel在接受BBC采访时表示，公司愿意为青少年实施每日使用时间限制，并称这将是「该行业向前的重要一步」。
The Trump administration plans to deliver 60,000 of the powerful munitions at a time when opposition is growing in Washington to arms sales to Israel.
中文摘要 特朗普政府计划向以色列交付6万枚重磅弹药（一吨级炸弹）。在华盛顿对对以军售的反对声日益增强的背景下，以色列寻求更多此类武器。
Both sides agreed to expand air travel and resume direct commercial flights.
中文摘要 以色列与摩洛哥同意互设大使馆，升级双边外交关系。双方还同意扩大航空旅行并恢复直航商业航班，进一步加强经贸联系。
The comments from a US CENTCOM spokesperson come amid continued tensions over shipping through the vital waterway.
中文摘要 美国中央司令部发言人表示，在持续封锁背景下，霍尔木兹海峡仍然保持开放。该关键水道的航运通行问题仍面临紧张局势。
Internal reports, including one from the Pentagon’s Defense Intelligence Agency, say China could acquire U.S. jet technology through spying or cooperation with Saudi Arabia.
中文摘要 美国情报机构警告称，中国可能通过间谍活动或与沙特合作获取F-35战机技术。五角大楼国防情报局等内部报告提出了这一风险警示。
Higher diesel and petrol prices are pushing up transport and food costs across an already impoverished Syria.
中文摘要 叙利亚柴油和汽油价格上涨推高了交通和食品成本，加剧了该国本就贫困的经济困境。民众表示缺乏支付能力，经济痛苦加深。
The unanimous decision comes as prices have remained stubbornly high and despite Trump's demands for lower rates.
中文摘要 美联储一致投票决定加息，为三年来首次。决定在物价持续高位不下背景下作出，尽管特朗普总统多次要求降低利率，美联储仍坚持加息立场。
Rising temperatures contributed to the disaster by destabilizing the permafrost and ice around the collapsed rock, scientists concluded.
中文摘要 一份研究报告发现，人为导致的气候变暖对喜马拉雅地区洪水起到了「非常强烈」的作用。气温上升导致冻土和岩体周围的冰层不稳定，从而引发灾害。
Rates were hiked in a unanimous decision despite fierce opposition from President Donald Trump, who had called for a cut.
中文摘要 美联储以一致投票通过三年来首次加息，此举违背了特朗普总统要求降息的强烈呼吁。
Berkshire Hathaway Inc. is considering increasing its holdings in Japanese trading houses, according to the head of the lobby group for the firms.
中文摘要 三井物产表示，伯克希尔·哈撒韦公司正在考虑增持其持有的日本综合商社股份。
President Donald Trump directed agencies to remove Canadian goods from government contracts, executing an earlier threat as tensions between the two countries flare.
中文摘要 特朗普总统指示政府机构将加拿大商品从政府合同采购中剔除，两国关系随之趋于紧张。
Malaysia’s struggling bond market may see an outflow of Japanese capital as its yield premium shrinks.
中文摘要 随着马来西亚债券收益率溢价收窄，该国动荡的债券市场面临日本资金大幅撤离的风险。
Gold held a decline after the Federal Reserve raised interest rates for the first time in three years and signaled another hike was likely before year-end.
中文摘要 美联储三年来首次加息并暗示年底前可能再次加息，受鹰派立场影响，金价维持跌势。
Evan Speigel says it would be "an important step forward for the industry" in an interview with the BBC.
中文摘要 Snapchat首席执行官伊万·斯皮格尔表示，公司愿意为青少年实施每日使用时间限制，并视其为行业前进的重要一步。
Kris Singh says market sentiment against data centres had dented flotation
中文摘要 Holtec创始人克里丝·辛格称，因AI数据中心领域市场情绪低迷形成“完美风暴”，公司已取消IPO计划。
Nvidia, OpenAI and Anthropic are all now selling their own limited edition fashion lines
中文摘要 英伟达、OpenAI和Anthropic等科技巨头近期纷纷推出各自的限量版时装系列，引发了科技品牌跨界服饰的讨论。
From coffee chats to LinkedIn messages, here's how to network your way to your next job.
中文摘要 BBC报道了一位求职者通过向200家公司发送私信和领英信息等方式成功找到工作，分享了职场社交获聘的经验。
Forecasts of further rises in the inflation rate mean some analysts expect the Bank to act by the end of the year.
中文摘要 英国央行按预期维持利率不变，但因通胀率可能进一步上升，部分分析师预计央行将在年底前采取行动。
Cross-party group says term has become ‘lightning rod for opposition to climate action’
中文摘要 英国跨党派议员呼吁将“净零排放”从相关部门名称中移除，称该术语已成为反对气候行动的对立焦点。
Severe weather disruptions will amplify food security risks in Asia, potentially triggering civil unrest and political instability in vulnerable countries such as India, Indonesia and the Philippines, according to consultancy Verisk Maplecroft.
中文摘要 咨询公司Verisk Maplecroft警告，厄尔尼诺现象与战争正威胁亚洲粮食供应，印度、印尼和菲律宾等国面临动荡风险。
4 回复 · Apple 节点
4 回复 · Apple 节点
5 回复 · 程序员 节点
4 回复 · Apple 节点
70 回复 · 程序员 节点
7 回复 · 程序员 节点
4 回复 · 程序员 节点
8 回复 · 程序员 节点
10 回复 · 程序员 节点
28 回复 · Apple 节点
该源今日无内容。