C · ★ 30,923 · 🍴 3,318 · 📈 868 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 Colibri 是一个纯 C、零依赖的极简推理引擎，能在普通消费级硬件上运行前沿 MoE 大模型，通过将专家模块按需从磁盘流式加载，显著降低显存占用。适合算力有限的开发者和研究者本地跑 MoE 大模型，实现低成本推理。
Go · ★ 24,234 · 🍴 1,783 · 📈 443 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 Alibaba Open Code Review 是在阿里生产环境验证过的混合架构代码审查工具，结合确定性流水线与 LLM Agent 生成行级精准评论，内置 NPE、线程安全等多语言规则。面向企业开发团队，用于提升代码质量与审查效率。
Python · ★ 8,036 · 🍴 888 · 📈 487 stars today
YuE2: frontier music generation with symbolic planning, zero-shot covers, and agentic music editing.
中文介绍 YuE2 是前沿音乐生成框架，支持符号化规划、零样本翻唱和代理式音乐编辑。适合音乐创作、AI 配乐、二次创作等场景，可让开发者或音乐人快速完成编曲、改词与风格迁移。
Python · ★ 28,026 · 🍴 3,445 · 📈 2,632 stars today
VoiceStudio is the open-source, fully-local ElevenLabs alternative — voice cloning, voice design, video dubbing, dictation, transcription & audiobook creation in 646 languages.
中文介绍 VoiceStudio 是开源、完全本地运行的 ElevenLabs 替代品，支持声音克隆、声音设计、视频配音、口述转文字、转录与有声书生成，覆盖 646 种语言。适合对隐私敏感的播客、内容创作者和企业本地化语音需求。
Python · ★ 72,789 · 🍴 11,266 · 📈 524 stars today
A Simple and Universal Swarm Intelligence Engine, Predicting Anything. 简洁通用的群体智能引擎，预测万物
中文介绍 MiroFish 是一个简洁通用的群体智能引擎，用模拟多智能体协作的方式对各类问题做预测。适合数据分析、社会预测、决策支持等场景，为研究者和企业用户提供可解释的预测框架。
Python · ★ 80,705 · 🍴 7,031 · 📈 640 stars today
Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
中文介绍 Agent-Reach 让 AI Agent 具备浏览全网的能力，通过一个 CLI 工具读取和搜索 Twitter、Reddit、YouTube、GitHub、B 站、小红书等平台，无需 API 调用。适合自主 AI 助手开发者构建实时信息获取管道。
JavaScript · ★ 66,420 · 🍴 10,840 · 📈 706 stars today
Extracted system prompts from Anthropic - Claude Fable 5.1, Opus 5, Claude Design, Claude Code. OpenAI - ChatGPT GPT-6-Astra, Codex. Google - Gemini 3.8 Flash, 3.1 Pro, Antigravity. xAI - Grok, Grok Bot, Cursor, Kimi and more! Updated regularly.
中文介绍 该仓库收集了 Anthropic Claude、OpenAI GPT、Google Gemini、xAI Grok 等主流大模型的泄露系统提示词，覆盖 Claude Opus、Codex、Gemini 3.8 Flash 等版本。适合研究提示工程、安全分析和模型行为逆向工程。
Python · ★ 1,797 · 🍴 151 · 📈 52 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 oh-my-hermes 是 Hermes Agent 的一体化插件，集成编码智能、长期记忆系统和模型优化工作流包。适合 AI Agent 开发者打造具备持续学习和专业编码能力的智能体。
Dart · ★ 90,968 · 🍴 5,078 · 📈 213 stars today
An open-source cross-platform alternative to AirDrop
中文介绍 LocalSend 是开源跨平台 AirDrop 替代品，通过局域网在 macOS、Windows、Linux、Android、iOS 之间无线传输文件。适合用户跨设备快速分享文件、照片与文档，无需云端账号。
Rust · ★ 67,407 · 🍴 3,204 · 📈 110 stars today
Unofficial Bitwarden compatible server written in Rust, formerly known as bitwarden_rs
中文介绍 Vaultwarden 是用 Rust 编写的非官方 Bitwarden 兼容服务器，允许自托管密码管理。适合注重隐私的个人和团队自建密码保险库，与 Bitwarden 客户端无缝对接。
Python · ★ 105,627 · 🍴 20,244 · 📈 756 stars today
TradingAgents: Multi-Agents LLM Financial Trading Framework
中文介绍 TradingAgents 是面向金融交易的多智能体 LLM 框架，让多个 AI Agent 协同进行市场分析、决策与交易模拟。适合量化研究者与金融科技团队实验 LLM 驱动的自动化交易策略。
Rust · ★ 93,616 · 🍴 12,415 · 📈 370 stars today
π RuView turns commodity WiFi signals into real-time spatial intelligence, vital sign monitoring, and presence detection — all without a single pixel of video.
中文介绍 RuView 利用普通 WiFi 信号实现实时空间感知、生命体征监测和存在检测，无需摄像头。适合智能家居、养老监护和安防场景，在隐私友好的前提下提供非视觉感知能力。
TypeScript · ★ 5,880 · 🍴 508 · 📈 265 stars today
The secure, validated skill registry for professional AI coding agents. Extend Antigravity, Claude Code, Cursor, Copilot and more with absolute confidence.
中文介绍 agent-skills 是面向专业 AI 编码 Agent 的安全技能注册表，为 Antigravity、Claude Code、Cursor、Copilot 等工具提供经验证的技能扩展。适合企业开发团队统一管理和分发 Agent 能力。
Python · ★ 37,194 · 🍴 4,228 · 📈 204 stars today
VoxCPM2: Tokenizer-Free TTS for Multilingual Speech Generation, Creative Voice Design, and True-to-Life Cloning
中文介绍 VoxCPM2 是无分词器（Tokenizer-Free）的多语言 TTS 模型，支持创意声音设计与高保真语音克隆。适合本地化内容制作、有声书、配音等需要自然语音合成的场景。
Python · ★ 165,782 · 🍴 34,568 · 📈 152 stars today
🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
中文介绍 Hugging Face Transformers 是覆盖文本、视觉、音频和多模态任务的模型定义框架，支持主流 SOTA 模型的推理与训练。是机器学习研究者和开发者的核心工具库，广泛应用于 NLP 与多模态 AI 项目。
TypeScript · ★ 5,639 · 🍴 970 · 📈 191 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，集成 ERP、CRM、HRM、ATS、项目管理等模块。适合中小企业一站式管理客户、员工、招聘、项目和财务，降低多工具切换成本。
TypeScript · ★ 36,719 · 🍴 3,672 · 📈 26 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 是离线优先的知识与教育服务器，内置维基百科、数千本书籍、课程、地图，并可运行本地 AI，完全无需联网。适合学校、偏远地区、军用或灾难场景下的本地化知识访问。
TypeScript · ★ 8,123 · 🍴 1,015 · 📈 279 stars today
A modern platform for visual, flexible, and extensible graph-based investigations. For cybersecurity analysts and investigators.
中文介绍 Flowsint 是现代可视化的图调查平台，灵活可扩展，专为网络安全分析师和调查人员设计。适合追踪威胁情报、关联攻击链、构建实体关系图谱等安全调查任务。
Swift · ★ 3,317 · 🍴 239 · 📈 314 stars today
Free, open-source Sidecar/Duet alternative — use your iPhone or iPad as a true second monitor for your Mac over USB or WiFi. Low latency H.264, Retina HiDPI, touch input.
中文介绍 OpenDisplay 是免费开源的 Sidecar/Duet 替代品，通过 USB 或 WiFi 把 iPhone/iPad 变成 Mac 的真第二屏，采用 H.264 低延迟传输，支持 Retina HiDPI 和触控输入。适合需要扩展屏幕的 Mac 用户。
Python · ★ 4,402 · 🍴 608 · 📈 506 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 Claude-Red 是面向 Claude 技能系统的攻击性安全技能库，每个技能以结构化 SKILL.md 形式封装专家级方法论。适合红队、渗透测试人员借助 Claude 执行专业的攻防任务。
👍 451
Automating empirical research is a long-standing direction of AI. Recent automatic research (AutoResearch) agents bring this goal within reach, as modern LLMs show the capability to independently implement solutions and learn from the execution outcomes. Behind these gains, post-training (especially
中文介绍 针对 AutoResearch 智能体扩展难的问题，提出基于 World Models 提升自动科研智能体的能力，使 LLM 能独立实现方案并从执行结果中学习，推动自动化实证研究的可行性。
👍 303
We introduce NCP-ArchPreview, a latent-space language model that pushes autoregressive pretraining beyond standard next-token prediction (NTP). Alongside NTP, the model learns through Next Concept Prediction (NCP) to predict discrete concepts that span multiple tokens, introducing an explicit and mo
中文介绍 提出 NCP-ArchPreview，一个潜空间语言模型，在传统 next-token prediction 基础上引入 Next Concept Prediction（NCP），预测跨多个 token 的离散概念，为预训练提供显式结构化信号，超越标准 NTP 范式。
👍 134
Large Vision-Language Models (LVLMs) have achieved strong performance on diverse visual tasks, yet their ability to reconstruct and reason about the 3D structure of the scene depicted in 2D images -- referred to as spatial intelligence -- remains limited. Existing approaches attempt to address this
中文介绍 提出 SpatialBlock，通过合成积木堆叠问题提升 LVLMs 的空间智能。针对视觉语言模型难以重建二维图像中三维场景结构的问题，利用合成的空间推理任务进行训练，增强场景空间理解与推理能力。
👍 93
Large language model (LLM)-based multi-agent systems (MAS) achieve strong performance by employing specialized multiple agents, yet their performance depends on the prompt design of each agent. For MAS prompt optimization, textual gradient methods that guide prompt updates using natural-language fee
中文介绍 提出 AgentGrad，一种干预引导的多智能体系统提示优化方法。利用自然语言反馈构建文本梯度指导每个智能体的提示更新，提升 LLM 多智能体协作系统的整体性能，减少对人工提示调优的依赖。
👍 72
Training capable cyber agents is often treated primarily as a problem of model scale, yet open-weight post-training is constrained more directly by the cost of executable environments, reliable multi-turn supervision, and access to strong teachers. We present a data-centric framework that addresses
中文介绍 提出 Feyospace-v1，一个数据为中心的安全模型训练框架，解决开源安全模型后训练受限于可执行环境成本、多轮监督质量和强教师模型获取的问题，训练出能力接近前沿的安全 AI 模型。
👍 59
Large Language Model (LLM) agents are turning language into real-world effects, making safety necessary against both indirect prompt injections and direct harmful requests. System-level safety harnesses add an enforcement layer beyond model-level defenses, but existing harnesses are usually designed
中文介绍 提出 EvoSafeHarness，通过进化算法生成模型特定和领域特定的安全约束 Harness，在 LLM 智能体系统层面部署安全防护层，防御间接提示注入和直接有害请求，突破通用型 harness 的局限。
👍 56
Agent usage is shifting toward long-horizon tasks such as coding and scientific discovery, among which terminal tasks are especially important. We introduce T1, a Mixture-of-Experts model of 122B total trained with reinforcement learning, operating a real shell in a cloud sandbox for up to 300+ tool
中文介绍 提出 T1，一个 122B MoE 模型，通过强化学习训练，可在云端沙箱操作真实 shell 完成超过 300 步工具调用的长程终端任务，面向代码和科学发现等长周期智能体场景。
👍 49
Robot foundation models achieve strong in-distribution performance but often degrade under visual distribution shifts. When learning to generate actions from pretrained visual representations, models may exploit task-irrelevant visual cues that correlate with demonstrated actions within the training
中文介绍 提出 Latent Interface Training，解决机器人基础模型因视觉分布偏移导致的性能下降问题。通过潜在接口训练避免模型利用与任务无关的视觉捷径，提升动作生成的泛化能力。
👍 46
Benchmark researchers and developers of large language models (LLMs) and other AI systems need to find relevant evaluations, locate their benchmark datasets and code, and understand the settings behind reported scores. We present Benchmark Radar, a living database and search engine for retrieval and
中文介绍 提出 Benchmark Radar，一个 AI 评测基准的活数据库和搜索引擎，帮助研究人员检索相关评测、定位数据集和代码，理解分数背后的实验设置，提升基准评测的可发现性和可追溯性。
👍 45
Post-training attention sparsification reduces the quadratic cumulative attention cost of pretrained Transformers by selecting a small set of context units (tokens or blocks) for each query. Existing trainable methods usually use a lightweight selector to score context units, followed by hard Top-K
中文介绍 提出 SAS（Simple Attention Sparsification），通过端到端优化上下文排序实现注意力稀疏化，降低 Transformer 二次注意力成本，避免传统 Top-K 硬选择方法带来的信息损失问题。
👍 43
Reducing audio-encoder depth lowers the inference cost of speech large language models, but removing complete blocks perturbs the embeddings consumed by the decoder and can cause deletion and premature end-of-sequence errors. We introduce X-AuT, a progressive framework that selects layer combination
中文介绍 提出 X-AuT，一种渐进式音频编码器压缩框架，通过跨尺度蒸馏选择层组合来缩减语音 LLM 的推理成本，避免直接移除完整层导致的删除和过早序列终止错误。
👍 37
Mainstream robotic policies often adopt a Markovian formulation, but many complex real-world manipulation tasks are inherently non-Markovian, requiring long-horizon memory beyond the current observation. Existing memory mechanisms often rely on language summaries, growing visual windows, or their co
中文介绍 提出 Memory as Plans，一种基于记忆锚定的规划方法用于世界-动作建模，解决机器人策略中标记马尔可夫假设不适用于复杂长程操作任务的问题，通过显式记忆机制超越当前观测约束。
👍 32
We introduce MetroLLM-Bench, a 955-case benchmark for testing language models as the policy layer of a transit kiosk. It covers six real metro systems, ranging from 37 to 414 stations, and eleven categories that include routing, fare calculation, disruptions, accessibility, and adversarial input. In
中文介绍 提出 MetroLLM-Bench，一个 955 用例的基准，测试 LLM 作为地铁自动售票机策略层的能力，覆盖六个真实地铁系统（37 至 414 站），包含路线规划、票价计算、中断处理和对抗输入等十一类任务。
👍 31
On-Policy Self-Distillation (OPSD) has emerged as a popular paradigm for large language model (LLM) self-improvement, allowing models to act as their own teachers by leveraging privileged information such as ground-truth solutions. However, recent findings indicate that OPSD can severely degrade the
中文介绍 提出 Negative Self-Distillation，通过在推理时规避缺陷进行自蒸馏学习，解决 On-Policy Self-Distillation 严重退化模型能力的已知问题，用反例引导 LLM 自我改进推理质量。
👍 31
Quantization has been widely adopted in LLM training and inference to reduce cost and improve efficiency. However, low-bit quantization of the attention module often introduces large errors at very low bit-widths, causing performance degradation. Existing methods mainly rely on smoothing techniques
中文介绍 提出 HyQuant，一种混合精度量化方法针对 LLM 注意力模块，解决极低比特量化引入大误差导致性能下降的问题，相比基于平滑技术的现有方法，在低比特下保持注意力计算精度。
👍 29
Temporal reasoning benchmarks for Video-LLMs are often mediated by language, leaving room for linguistic shortcuts from option wording, answer correlations, or language priors. To reduce such shortcuts, we introduce TempCloze, a video cloze benchmark for evaluating visual temporal reasoning in Video
中文介绍 提出 TempCloze，一个视频完形填空基准，用于评估 Video-LLMs 的视觉时序推理能力。通过填空形式减少语言捷径（如选项措辞、答案相关性），迫使模型依赖视觉时序理解而非语言先验。
👍 28
Neurosymbolic systems rely on mathematical solvers to guarantee reasoning correctness, yet solvers are fundamentally blind to whether a formal translation maintains strict reference-equivalence to a designated formalization. We formalize this vulnerability as Verdict-Preserving-Unfaithfulness (VPU):
中文介绍 提出 Generative Reward Models for Autoformalization，解决神经符号系统中求解器无法验证形式化翻译是否保持严格引用等价性的问题，形式化该漏洞为 VPU，并提出生成式奖励模型来弥补验证盲区。
👍 28
Code world models represent worlds as executable programs, but this representation alone does not determine how to construct a complex world. We introduce Recursive Code World Models (RCWM), a framework for reconstructing complex 3D worlds in code from a single reference image. RCWM couples a Recurs
中文介绍 提出 Recursive Code World Models（RCWM），一个通过递归场景程序从单张参考图像重建复杂三维世界的框架，将世界表示为可执行程序，通过递归结构构建复杂场景的代码表示。
👍 27
Reasoning language models have made substantial advances on a variety of complex tasks, yet their capabilities remain overwhelmingly English-centric: models primarily reason in English regardless of the language they are prompted in. This is inaccessible for non-English-speaking users, risks losing
中文介绍 提出数据混合策略以构建多语言推理桥梁，解决推理语言模型能力严重偏向英文、非英文用户无法获得同等推理支持的问题，通过合理的多语言数据配比提升各语言内部的推理泛化能力。
👍 26
Large language model (LLM) agents can benefit from reusable skills distilled from prior task experience, yet existing skill optimization methods often rely on costly execution-based evaluation and substantial task data. We introduce COBRA-Skills, an efficient framework that formulates skill optimiza
中文介绍 提出 COBRA-Skills，一种基于上下文赌博机引导的进化框架，用于 LLM 智能体技能优化。将技能优化形式化为赌博机问题，避免昂贵的执行评估和大量任务数据需求，实现高效可复用技能蒸馏。
@IAmPascio · 124.8K 粉丝 · 812.5K 阅 · 881 赞 · 140 转
The superintelligence, the singularity, AGI, whatever you call it, the one we're so scared of is, ironically, also the one we most likely won't see coming. This essay is going to probably be a pretty
中文介绍 思辨长文，提出一个反常识假设：AGI可能已出现，但其载体不是超级模型，而是儿童群体协作。作者分析「奇点」为何会被集体忽视，属于概念推演与未来预测，适合喜欢AI哲学讨论的人。
@builtbysketch · 1.4K 粉丝 · 717.2K 阅 · 518 赞 · 79 转
I built a full game with GPT-6 Astra, and it kinda blew up on X. Initially it was just the demo, but 4 days later we have the full playable Okay, not just a crappy one-run demo. Check PaperRoute
中文介绍 博主用 Blender 配合 GPT-6 Astra 从零做出完整可玩游戏 PaperRoute，不是只跑一局的 demo，而是 4 天内把原型扩成完整版。属于 AI 游戏开发实战案例，展示生成式工具链已能支撑端到端游戏生产。
@mayukh_panja · 9.8K 粉丝 · 311.6K 阅 · 501 赞 · 75 转
There seems to be a widespread impression that OpenAI has somehow “solved the Navier–Stokes equations,” and that this represents some enormous breakthrough for fluid dynamics and engineering. It
中文介绍 澄清 OpenAI 并未「解决 Navier-Stokes 方程」，社区对近期流体动力学相关工作的解读存在严重夸大。作者梳理原始论文内容与工程应用的差距，属于辟谣型科普，帮助辨别 AI 进展的宣传话术。
@docmilanfar · 116.7K 粉丝 · 126.0K 阅 · 512 赞 · 63 转
In AI circles, Recursive Self-Improvement (RSI) is talked about in almost magical terms: a model tweaks itself, becomes smarter, uses that intellect to rewrite itself again, and this leads to
中文介绍 讨论递归自我改进（RSI）是否真会指数爆发，作者认为智能提升存在物理和信息层面的速度上限，模型重写自身的增益会递减。属于 AI 对齐领域的理论反思，挑战「智能爆炸」叙事。
@mbeisen · 73.0K 粉丝 · 113.7K 阅 · 528 赞 · 68 转
Over the last few months, the cancer-curing capabilities of LLMs have improved dramatically, to the point that they have cured previously untreatable types of cancer and are on the cusp of developing
中文介绍 反讽性质的推文，声称 LLM 在癌症治愈能力上近期突飞猛进、已攻克此前不可治的癌症类型。实为夸张叙事或测试性内容，提醒读者对「AI 治愈癌症」类标题保持警惕。
@charliejhills · 14.4K 粉丝 · 97.6K 阅 · 510 赞 · 70 转
Harness, loop, context, graph engineering. It's tiring, right? Every month there's a new AI system you MUST implement. I do this full-time and I still struggle to keep up. But all of them matter (to
中文介绍 用面向普通人的语言串讲当前热门 AI 系统概念：Harness、loop、context、graph engineering 等，解释它们为何每月都在冒出、以及如何互相替代。适合作为概念速查与术语入门。
@EXM7777 · 142.6K 粉丝 · 94.6K 阅 · 501 赞 · 56 转
I'm basically giving you the full system i'd use to automate content creation for ANY business... every tool in the chain, every step in order, and where it breaks because a content agent is just a
中文介绍 完整课程式分享，覆盖为任意业务自动化内容生产的整条工具链，包括每一步的具体工具、执行顺序，以及 content agent 容易失效的环节。实操导向，适合想搭建内容自动化流水线的创作者。
@OpenAIDevs · 421.3K 粉丝 · 51.6K 阅 · 677 赞 · 54 转
We shipped GPT-6 Astra. Builders took it from there: Explore 2,234 modeled anatomical parts in 3D: A Manhattan recreation built in Unreal Engine: Six Van Gogh paintings become a walkable
中文介绍 OpenAI 官方展示 GPT-6 Astra 发布后社区创作成果：2234 个解剖部件的 3D 模型、Unreal Engine 中的曼哈顿复刻、梵高画作变可走场景等。属于新模型首发后的生态案例集。
@AndrewYNg · 1.9M 粉丝 · 43.5K 阅 · 722 赞 · 105 转
When you’re skilled at AI Engineering, your best work won’t be merely implementing a product that someone else spec’ed out. Instead, you will actively shape the build. Before modern AI tools
中文介绍 吴恩达发布 AI 工程技能地图，强调成熟工程师的价值不在于实现他人 spec，而在于主动塑造产品构建方向。结合现代 AI 工具前后期能力差距，梳理能力分层，适合职业规划参考。
@anshuc · 15.6K 粉丝 · 41.4K 阅 · 544 赞 · 36 转
Most people say to use the big, expensive model as an orchestrator, delegating grunt work to small, cheap workers. That's the opposite of what you should do. Here's why. Comparing different approaches
中文介绍 提出「傻老板式」策略获取更多 Astra/Fable 额度：与主流做法相反，不用大模型做 orchestrator 把活分给小 worker，而是反向调度。属于 prompt 编排反直觉技巧，附实测对比。
@alexwg · 68.7K 粉丝 · 23.4K 阅 · 501 赞 · 47 转
The Singularity just got its first formal request to slow down from the people building it, and the strangest part is the guest list. Dario Amodei published "We Must Pace the Frontier," arguing that
中文介绍 2026 年 9 月 13 日 AI 圈日报，核心事件是 Dario Amodei 发表「We Must Pace the Frontier」呼吁放缓前沿研究，且签名者名单出人意料。属于行业热点综述，关注 AI 治理与减速论。
@Didotxyz_ · 63.6K 粉丝 · 17.9K 阅 · 516 赞 · 388 转
Model access is becoming easier to buy, compare, route, and replace. Public API menus now include premium reasoning models, lower-cost variants, cached inputs, batch processing, and models optimized
中文介绍 观点文：模型调用正越来越便宜、易替换，但上下文成本并未同步下降。分析公共 API 菜单、缓存输入、批处理等趋势，指出推理基础设施的新瓶颈在 context 而非模型本身。
Perplexity uses Astra to write communications, change software, and monitor production systems, and checks in much less frequently than with earlier models.
中文介绍 Perplexity采用GPT-6 Astra用于撰写通信、修改软件及监控生产系统，人工介入频率较此前模型显著降低。
Before co-founding Kepler, Vinoo Ganesh led Spark at Palantir and built Project Frontline — a pioneering program for Forward Deployed Engineers. He takes us through the best practices of FDEs.
中文介绍 Latent Space访谈Kepler联合创始人、曾任Palantir Spark负责人及Project Frontline架构者的Vinoo Ganesh，探讨前线部署工程师（FDE）的最佳实践。
We agree with Sebastian: this should have been DeepSeek v5
中文介绍 DeepSeek v4.1-Flash发布，采用763B参数、8B隐层、16B解码的因果式Encoder–Decoder架构并支持视觉，被评价为「鲸鱼的回归」。
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? Join MIT Technology Review executive editor Niall Firth for a conversation with senior AI editor Will Douglas Heaven and AI
中文介绍 MIT Tech Review圆桌对话，讨论顶级AI实验室员工认为先进AI可能毁灭人类的观点是否成立，由主编Niall Firth主持。
GPT‑6 Astra improves Devin’s ability to test software and show that it works, with the goal of helping engineers review less code and ship more.
中文介绍 Cognition使用GPT-6 Astra增强Devin的软件测试能力，验证产出可用性，以减少工程师代码审查、加快交付。
Learn how OpenAI evolved Habitat from a Python library into a globally distributed storage platform serving 1 billion ChatGPT users and 22M requests per second.
中文介绍 OpenAI将Habitat从Python库演进为全球分布式存储平台，支撑超过10亿ChatGPT用户及每秒2200万次请求。
中文介绍 TLDR AI日报汇总Agents API、Cognition SWE-2与Muse Shared Agents等AI领域动态。
César de la Fuente’s lab uses Codex and ChatGPT to search living and extinct genomes for antimicrobial candidates to fight drug-resistant infections.
中文介绍 César de la Fuente实验室借助Codex与ChatGPT在现存及已灭绝生物基因组中检索抗菌分子候选，以应对耐药感染。
Meet the Data agent in ChatGPT Work. Connect company data, uncover insights, and build interactive dashboards with AI using natural language.
中文介绍 OpenAI推出ChatGPT Work中的数据智能体，支持连接企业数据、挖掘洞察，并通过自然语言构建交互式仪表盘。
On July 22, 2026, a transmission line fault in Ashburn, Virginia—the heart of the world’s largest data center cluster—knocked more than 3 gigawatts of load off the grid in seconds. And it wasn’t the first time. Two years earlier, a single failed surge arrester dropped roughly 60 Virginia facilities
中文介绍 MIT Tech Review指出AI供电是架构问题：2026年7月22日弗吉尼亚Ashburn电网故障致超3吉瓦负荷秒级脱网，两年前单个失效避雷器曾致约60%掉电。
Introducing ChatGPT for Financial Services, combining built-in financial data and GPT-6 Astra for research, modeling, and client-ready materials.
中文介绍 OpenAI发布面向金融服务的ChatGPT，整合内置金融数据与GPT-6 Astra，用于研究、建模及生成客户交付材料。
OpenAI and GSA will offer eligible federal, state, local, and tribal governments $0 license fees, 50% off usage, and expanded cyber defense support.
中文介绍 OpenAI与美国GSA合作，为符合条件的联邦、州、地方及原住民政府提供0元授权费、50%用量折扣及强化网络防御支持。
a quiet day
中文介绍 Latent Space AINews刊文称当日AI领域无重大动态。
Build and launch cloud agents with the Agents API, a managed service powered by the Codex harness for orchestration, long-running sessions, and tool use.
中文介绍 OpenAI推出Agents API，基于Codex运行时的托管服务，支持编排、长时会话与工具调用，用于构建与发布云端智能体。
GPT‑Live‑1 brings natural, full-duplex voice conversations to the API, with stronger instruction following, custom voices, and telephony support.
中文介绍 OpenAI在API中上线GPT-Live-1，提供全双工自然语音对话，强化指令遵循，支持自定义音色与电话集成。
美股三大指数技术面偏强但短期回调压力显现：SPY 764.29贴近52周高(-1.93%)且多头排列，QQQ 714.88距52周高-4.51%，两者近5日均小幅回落，MACD 相对信号走弱。科技权重股分化，AAPL、META 动量占优，NVDA 触发 MACD 死叉，TSLA、GOOGL 趋势中性。加密方面，恐慌贪婪指数=57(贪婪)、总市值 2.65T USD(24h -1.3%)、BTC 主导率 58.9%、ETH 11.6%，BTC-USD 77860 与 ETH-USD 2512 维持多头排列但距52周高分别-38.3%/-47.17%，回调空间较大。中概股整体空头排列偏弱势，BABA、PDD、0700.HK 均呈下行结构，JD 中性。商品外汇端 CL=F 强势超买、GC=F 中性、USDCNY=X 深度超卖走弱；宏观面 ^VIX 单日+11.74%、^TNX 4.97 触发超买接近52周高，显示风险偏好边际收敛、期限溢价抬升。
近5日涨幅11.96%、近1日+2.37%，价格102.42明显高于SMA20=89.52、SMA50=83.9、SMA200=79.8，多头排列成立；MACD 4.6492 高于信号3.0962，红柱扩张；但RSI14=73.8已入超买区，短线上行动能强但追高存在回调风险。
价格77.81低于SMA20=84.32、SMA50=85.65、SMA200=97.3，呈空头排列；近5日-4.68%，MACD=-2.2718低于信号-1.3506，绿柱继续扩大；RSI14=30.3已接近超卖边界，趋势与动量均偏空。
6.7低于SMA20=6.72、SMA50=6.75、SMA200=6.85，空头排列；距52周低仅0.06%，MACD=-0.0119与信号-0.0112接近且死叉信号在列；RSI14=14.3深度超卖，下行趋势持续但已临近超卖反弹阈值。
价格218.29略低于SMA20=220.16但高于SMA50=212.58与SMA200=197.47，中期多头排列尚存；近5日-4.45%，MACD 2.5719低于信号3.1035已出现死叉，RSI14=49.8贴近50中性轴，短中期信号矛盾。
价格332.27高于SMA20=316.63、SMA50=317.97、SMA200=284.95，多头排列；近1日+1.75%、近5日+1.24%，MACD 3.227高于信号1.9023红柱扩张，RSI14=62.8尚在中性偏强区间，距52周高仅-3.57%。
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与形态复述，过去走势不代表未来表现，不构成任何投资建议；所有描述仅供技术指标解读参考，实际决策需结合基本面与个人风险承受能力。
Talks between the United States and Iran hit a dead end weeks ago, leaving Arab countries struggling to manage their ties to Tehran amid ongoing attacks.
中文摘要 伊朗与阿拉伯国家原定会议延期，中东外交陷入僵局。数周前美国与伊朗谈判已破裂，海湾阿拉伯国家在持续袭击背景下难以处理与德黑兰的关系，地区外交努力受阻。
U.K. and Irish politicians on the left and the right are in rare agreement over President Trump's suggestion that Ireland and Northern Ireland reunite — he doesn't know what he's talking about.
中文摘要 特朗普暗示爱尔兰与北爱尔兰应重新统一后，英国及爱尔兰左右两翼政界罕见达成共识，纷纷质疑其表态，认为他对此议题并不了解，相关外交表态引发争议。
朝鲜官方周一宣布成功举行涵盖导弹、火炮与无人机的多兵种联合实弹演习，并宣称展示出“巨大破坏力”。即便美方近期主动缩减美韩联合军演规模以寻求外交突破，平壤依然密集测试整合武器控制系统。
中文摘要 朝鲜官方周一宣布成功举行多兵种联合实弹演习，试射导弹、火炮与无人机，并宣称展示「巨大破坏力」。尽管美方近期缩减美韩联合军演规模以求外交突破，平壤仍密集测试整合武器控制系统。
The parliament isn’t set to meet until late September to elect a new speaker, and then prime minister But leading Social Democrats are not exactly impressed by Kristersson’s chill approach. The party’s foreign policy spokesperson, Morgan Johansson, accused him to trying to “cling on” to power, sayin
中文摘要 瑞典选举在已计95%选票的情况下，中间偏左反对党将领先优势扩大至三个席位。议会将于九月底召开选举新议长及首相。社会民主党外交发言人Morgan Johansson批评Kristersson试图「抓住」权力。
China's Ministry of Foreign Affairs on Monday hit back at a call from the head of Anthropic for the U.S. to curb China's artificial intelligence capabilities.
中文摘要 中国外交部周一回应Anthropic首席执行官要求美国限制中国人工智能能力发展的言论，中方予以反驳。北京强调AI发展应遵循公平竞争，反对单方面遏制措施。
The election, in an area long known for violent separatists, was marred by a shootout that killed three, officials said.
中文摘要 菲律宾棉兰老岛穆斯林自治区举行具有里程碑意义的大选，该地区长期因暴力分裂活动动荡不安。官员称，选举期间发生枪击事件，造成三人死亡，选举过程受到干扰。
At least 51 people rescued as search operation continues off the island of Gavnos near Crete.
中文摘要 希腊克里特岛附近加夫诺斯岛海域一艘载有移民的船只沉没，至少1人死亡、25人失踪，已救出51人。搜救行动仍在持续进行中，救援机构继续搜寻失踪者。
Cutting-edge technologies have reshaped the Russian capital, making it safer, more convenient — and more closely watched.
中文摘要 新莫斯科借助前沿科技重塑首都，使其更安全、便利，同时也成为监控重地。数字技术大幅加强了对市民的日常监视，构成一个安全与监控并存的双重城市。
Yemeni government forces have advanced in western Taiz after successfully repelling a Houthi offensive.
中文摘要 也门政府军在成功击退胡塞武装进攻后，向塔兹西部推进。同时，胡塞武装宣称对沙特阿拉伯发动了袭击，也门战事持续升级，沙特面临外部安全威胁。
STC’s 2027 program also features Celeste Barber in Amy Sedaris and David Sedaris play, Tim Minchin as Scrooge and the return of The Picture of Dorian Gray Six years after the Australian historian Jenny Hocking won her legal battle to release secret correspondence between Queen Elizabeth II, Malcolm
中文摘要 澳大利亚历史学家Jenny Hocking胜诉后，伊丽莎白二世与惠特拉姆的秘密通信被公开，激发悉尼剧院公司（STC）2027年剧目创作。STC 2027节目还包括Celeste Barber、Tim Minchin及《道林格雷画像》回归。
At least six people were killed and more than 100 rescued after the vessel sank in the Java Sea on Sunday, the authorities said.
中文摘要 一艘渡轮周日在印度尼西亚爪哇海沉没，造成至少6人死亡、超过120人下落不明，救援当局已救出100余人。该事件发生在婆罗洲附近海域，救援工作仍在进行。
It’s about more than just shamelessness.
中文摘要 文章探讨为何丑闻已不再能推翻政客，认为原因远超单纯「不知羞耻」。政治极化、媒体碎片化及选民群体分化等因素共同削弱了丑闻对政治生涯的冲击。
Current chatbots aren’t likely to help lone actors invent a deadly pathogen, experts say. But newer models, trained in biology, may need stronger safeguards.
中文摘要 专家指出当前聊天机器人不太可能帮助个人制造致命病原体，但经过生物学训练的新模型可能带来更高风险，需要更强有力的安全措施，以防人工智能助推生物战威胁。
As global temperature increase shoots past 1.5 degrees Celsius, the vulnerable countries that fought for that Paris Agreement target face a brutal new reality.
中文摘要 全球气温上升已突破《巴黎协定》设定的1.5摄氏度上限，为达成此目标而奋斗的脆弱国家面临严峻新现实，气候危机后果进一步显现，引发新一轮应对挑战。
Racial and political tensions have flared in Trinidad and Tobago after members of an ethnic minority were accused of an assassination plot based on a private phone call.
中文摘要 特立尼达和多巴哥爆发种族与政治紧张局势，一名女性在私人通话中批评政府后遭拘押数月，其后被控参与针对政府的暗杀阴谋，事件引发社会强烈关注。
Just three years ago, global bankers would fly into the Gulf to corner a slice of the world’s hottest market for new share sales. Now, even local advisers are being forced to hunt for business outside the region as deals dwindle.
中文摘要 三年前全球银行家争相赴海湾抢滩IPO业务，如今IPO交易萎缩，连本地顾问也不得不走出该地区寻找新业务机会，海湾市场热度明显降温。
India’s inflation accelerated in August, pushing closer to the top of the central bank’s 2%-6% target range and potentially narrowing the scope for policymakers to remain on hold for long.
中文摘要 印度8月通胀加速上行，逼近央行2%-6%目标区间上限，可能压缩印度储备银行（RBI）维持利率不变的空间，政策暂停余地收窄。
Africa's richest man, Aliko Dangote, is set for a $23 billion leap in his net worth with the IPO of his refinery business. Jennifer Zabasajja explains why the continent's biggest listing promises to supercharge the Nigerian's wealth. (Source: Bloomberg)
中文摘要 非洲首富Aliko Dangote旗下炼油企业计划IPO，其净资产有望增加230亿美元，被视为非洲最大规模上市之一。
Investors were already fretting that AI hype had inflated a bubble in stocks.
中文摘要 科技投资者因AI发展放缓风险感到焦虑，市场担忧此前围绕AI热潮已被推高的股票泡沫可能受到冲击。
South Korea’s tech-heavy Kospi slides as Wall Street is set to open lower
中文摘要 在大型AI企业呼吁放缓技术发展的消息刺激下，韩国Kospi指数等科技股下挫，美股预计低开后市。
The commercial mortgage backed securities market, long a mainstay of financing for America’s offices, apartments and malls, is being reshaped by a surge in data-center deals, forcing investors to grapple with an entirely new set of risks.
中文摘要 数据中心交易激增正在重塑美国商业抵押贷款支持证券（CMBS）市场，令投资者面临一组全新的风险因素。
Equities fell around the world after leaders of the biggest artificial-intelligence firms proposed slowing the technology’s development, while a jump in oil prices sapped risk sentiment further. Tech stocks led declines as traders fretted that efforts to rein in the advance of cutting-edge AI could
中文摘要 最大AI企业高管提议放缓技术发展，叠加油价上涨，全球股市下跌，科技股领跌，欧洲股市承压。
Prince Michael of Liechtenstein accuses EU of ‘bureaucratic zeal’ over anti-money laundering rules
中文摘要 列支敦士登米夏埃尔亲王批评欧盟反洗钱披露规则「过度推进」，指责其体现「官僚狂热」，认为监管措施已走得太远。
An American miner is partnering with Rwanda’s government to source and process tungsten from the African nation, extending a push to build a Western-aligned supply chain for the strategic metal that reduces reliance on China.
中文摘要 一家美国矿企与卢旺达政府合作开发钨资源，旨在拓展西方供应链、降低对中国依赖，强化关键金属供应安全。
Anthropic and OpenAI have called for slowing AI development in the name of safety. But they may face pushback from the tech industry, financial markets and the Trump administration. Newley Purnell explains. (Source: Bloomberg)
中文摘要 Anthropic与OpenAI出于安全考虑呼吁放缓AI发展，但可能遭到科技行业、金融市场及特朗普政府的反对。
Danish pharmaceutical group is under pressure from investors to regain ground in obesity market
中文摘要 丹麦制药集团Novo Nordisk（诺和诺德）拟从名称中去除Nordisk，试图推进复兴，同时面临投资者要求其在减重市场夺回份额的压力。
Take part in a live Ask an Expert Q&A with John Gapper, the FT’s chief UK business columnist, on Thursday September 17 at 1pm BST
中文摘要 《金融时报》首席英国商业专栏作家John Gapper将于9月17日13时（BST）举办在线问答，讨论英国企业对预算案的需求。
23 回复 · 程序员 节点
14 回复 · 程序员 节点
20 回复 · 程序员 节点
34 回复 · 程序员 节点
10 回复 · 程序员 节点
52 回复 · 程序员 节点
17 回复 · Apple 节点
20 回复 · Apple 节点
39 回复 · 程序员 节点
62 回复 · Apple 节点
该源今日无内容。