Python · ★ 41,882 · 🍴 2,367 · 📈 3,463 stars today
A skill to stop your coding agent from burying the answer. ADHD-friendly output.
中文介绍 为 AI 编码助手准备的 skill，强制模型先给出结论，而不是把答案埋在冗长的推理过程里，输出更符合 ADHD 用户阅读习惯。适合注意力容易分散、只想快速拿到结果的开发者，可接入常见 coding agent 使用。
JavaScript · ★ 27,155 · 🍴 5,560 · 📈 3,680 stars today
A spy satellite simulator in your browser, except the data is real. Live open source spatial intelligence on a photorealistic 3D globe.
中文介绍 浏览器里的间谍卫星模拟器：基于真实开放数据，在逼真 3D 地球上做实时空间情报分析，可叠加各类地理图层。面向开源情报（OSINT）、地理信息与遥感方向的从业者与爱好者，无需专业软件即可上手探索。
TypeScript · ★ 2,913 · 🍴 203 · 📈 50 stars today
User friendly sideloader
中文介绍 用户友好的侧载（sideload）工具，把应用直接安装到设备上，简化签名、打包与部署流程，降低非开发者安装自制应用的门槛。适合需要给测试机或自有设备批量装包的用户。
TypeScript · ★ 1,358 · 🍴 497 · 📈 152 stars today
Open-source AI sales OS — self-hosted CRM with native AI agents + WhatsApp (WAHA). Open alternative to Kommo, Octadesk & Intercom for any business that sells by chat. MCP-ready, multi-tenant, LGPD.
中文介绍 开源 AI 销售系统：自托管 CRM，内置原生 AI agent，并通过 WAHA 接入 WhatsApp，支持 MCP、多租户与 LGPD 合规。定位是 Kommo、Octadesk、Intercom 的开源替代，适合以聊天为主要成交渠道的企业。
TypeScript · ★ 2,788 · 🍴 218 · 📈 552 stars today
Local-first AI coding agent desktop: Electron + Rust host core + pi Agent Harness + user-installable plugins
中文介绍 本地优先的 AI 编码 agent 桌面应用：Electron 负责界面，Rust 实现本地 host 核心，配合 pi Agent Harness 与可自行安装的插件体系。适合希望在本地掌控代码与模型调用、并靠插件扩展能力的开发者。
C · ★ 4,728 · 🍴 534 · 📈 350 stars today
Graphics Creation Tools
中文介绍 由 Armory 团队开发的 3D 纹理绘制工具，可在模型表面直接手绘 PBR 材质，覆盖烘焙、着色等图形创作流程，定位为开源免费的替代方案。面向游戏美术、独立开发者与 3D 内容创作者。
TypeScript · ★ 2,157 · 🍴 283 · 📈 626 stars today
Open Source AI trading agent that operates autonomously across 1000+ markets - Polymarket, Kalshi, Binance, Hyperliquid, Solana DEXs, 5 EVM chains. Scans for edge, executes instantly, manages risk while you sleep. Agent commerce protocol for machine-to-machine payments. Self-hosted. Built on Claude.
中文介绍 开源 AI 交易 agent，可自主运行于 Polymarket、Kalshi、Binance、Hyperliquid、Solana DEX 及多条 EVM 链等 1000 多个市场，自动扫描定价偏差机会、即时下单并管理风险。面向量化与加密交易者。
TypeScript · ★ 18,746 · 🍴 2,134 · 📈 647 stars today
LLM Wiki is a cross-platform desktop application that turns your documents into an organized, interlinked knowledge base — automatically. Instead of traditional RAG (retrieve-and-answer from scratch every time), the LLM incrementally builds and maintains a persistent wiki from your sources。
中文介绍 跨平台桌面应用，能把零散文档自动整理成互相关联的知识库，而不是像传统 RAG 那样每次提问都从零检索。适合需要长期积累、检索个人或团队资料的研究者与知识工作者。
Shell · ★ 285,382 · 🍴 25,520 · 📈 729 stars today
An agentic skills framework & software development methodology that works.
中文介绍 面向 AI 编码 agent 的技能（skills）框架，同时给出一套软件开发方法论，让 agent 按固定流程完成工程任务。适合想提升 agent 产出稳定性与可控性的开发者，据描述已在作者自身的开发实践中验证有效。
C# · ★ 15,738 · 🍴 1,949 · 📈 191 stars today
Smart PVR for newsgroup and bittorrent users.
中文介绍 老牌开源 PVR 工具，自动追踪剧集更新，通过 RSS 从 Usenet 与 BitTorrent 抓取资源并重命名入库，可对接下载客户端与媒体库。面向自建家庭影音、想自动化追剧的 NAS 用户。
Python · ★ 4,862 · 🍴 393 · 📈 129 stars today
🤖📐专为数学建模设计的 Agent & skills ,自动完成数学建模，生成一份完整的可以直接提交的论文。 An Agent Designed for Mathematical Modeling ,Automatically complete mathmodel and generate a complete paper ready for submission.
中文介绍 专为数学建模竞赛设计的 Agent 与技能集合，能自动完成选题分析、模型建立、求解与写作，最终输出一份可直接提交的完整论文。适合参加数模竞赛的学生与需要快速产出建模报告的研究者。
Go · ★ 1,156 · 🍴 88 · 📈 198 stars today
Network stack research tool. TCP tunnel with pluggable transports.
中文介绍 网络栈研究工具，提供带可插拔传输层（pluggable transports）的 TCP 隧道，便于测试自定义协议、载荷伪装与流量特征。面向网络安全研究者，以及需要做协议实验的底层网络开发者。
Python · ★ 2,630 · 🍴 259 · 📈 153 stars today
Agent-driven research knowledge base. Agents collect, search, and synthesize web research into a persistent, searchable wiki.
中文介绍 Agent 驱动的研究知识库：由 agent 自动采集、检索并综合网络资料，沉淀成可持续积累、可搜索的 wiki。适合需要长期跟踪某个领域信息的研究者与团队，把零散搜索结果变成结构化知识。
Rust · ★ 1,287 · 🍴 94 · 📈 120 stars today
Run parallel research agents with any model
中文介绍 让多个研究 agent 并行运行，可搭配任意模型使用，用并发分工加速资料检索、分析与写作。适合研究者、分析师或工程团队，把耗时的大规模调研任务拆给多个 agent 同时完成。
Python · ★ 135,774 · 🍴 12,202 · 📈 1,015 stars today
💫 Toolkit to help you get started with Spec-Driven Development
中文介绍 GitHub 官方推出的工具包，帮助团队上手规范驱动开发（Spec-Driven Development）：先写清规格与需求，再让 AI 编码工具据此实现。适合使用 Copilot 等 agent 做工程的开发者，减少需求含糊带来的返工。
TypeScript · ★ 23,613 · 🍴 2,942 · 📈 106 stars today
Open-source 3D architectural editor with a local CLI, MCP tools, and practical workflows for humans and AI agents.
中文介绍 开源 3D 建筑编辑器，提供本地 CLI 与 MCP 工具，并配套面向人类和 AI agent 的实用工作流。适合建筑师、空间设计者，以及希望让 AI agent 直接参与建模与场景编辑流程的开发者。
👍 437
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 针对自动科研（AutoResearch）智能体依赖真实实验、扩展成本高的问题，提出以世界模型驱动的框架：让智能体在可预测的环境模拟中评估并筛选候选研究方案，再借助后训练提升研究能力，从而在减少真实执行的前提下扩展自动科研的规模与稳定性。
👍 177
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 提出潜在空间语言模型 NCP-ArchPreview，在标准下一词预测（NTP）之外引入 Next Concept Prediction（NCP），让模型预测跨越多个 token 的离散概念，获得显式且更抽象的预测单元，推动自回归预训练从词级走向概念级潜在空间建模。
👍 89
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 针对多智能体系统（MAS）性能高度依赖各智能体 prompt 设计的问题，提出 AgentGrad：一种干预指导的 prompt 优化方法，通过干预实验归因各智能体的贡献并生成文本梯度，迭代更新各自 prompt，提升 MAS 的整体表现。
👍 68
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 针对大型视觉语言模型（LVLM）空间智能不足的问题，提出 SpatialBlock：构建合成积木堆叠问题作为训练与评测任务，要求模型从 2D 图像重建并推理 3D 结构，从而增强 LVLM 的空间理解与结构化推理能力。
👍 46
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 提出终端智能体 T1：122B 总参数的 MoE 模型，用强化学习训练，可在云端沙箱中操作真实 shell，支持 300+ 次工具调用，面向编码、科学发现等长时程任务，提升终端环境下的长程自主执行能力。
👍 36
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 针对 LLM 智能体面临间接 prompt 注入与直接有害请求的风险，提出 EvoSafeHarness：在模型级防御之外构建可演化的系统级安全护栏，按模型与领域自动演化形成执行层约束，提升智能体部署时的安全性。
👍 31
Recent advances in wearable sensing enable continuous monitoring of physiological and behavioral signals, yet existing benchmarks rarely evaluate whether AI systems can reason over a real user's longitudinal wearable record. We introduce WearableQA, a benchmark comprising 4,084 10-option multiple-ch
中文介绍 针对现有基准难以评估 AI 能否推理真实用户长期可穿戴记录的问题，提出 WearableQA：包含 4,084 道十选项多选题，覆盖生理与行为信号，用于系统评测健康推理能力，填补纵向可穿戴数据评测的空白。
👍 21
SWE-Bench Pro has emerged as a standard benchmark for evaluating software engineering agents on challenging repository-level tasks. However, our analysis work show that its evaluation is undermined by two sources of unreliability: reward hacking, enabled by leakage of gold solutions or hidden evalua
中文介绍 指出 SWE-Bench Pro 的评测存在两类不可靠性：金标准解法泄漏与隐藏评测信息诱发的 reward hacking。据此提出 SWE-Bench Pro Verified，修正后的可靠基准可更可信地衡量软件工程智能体在仓库级任务上的能力。
👍 17
Sequential memory agents process long documents by reading chunks one after another while maintaining a compact memory state, coupling document traversal to reasoning depth. This coupling introduces sensitivity to evidence placement and ties inference latency linearly to document length. We introduc
中文介绍 针对顺序记忆智能体阅读长文档时推理深度与文档长度耦合、对证据位置敏感且延迟随长度线性增长的问题，提出 PARSER：并行读取文档分块、深度推理，解耦遍历与推理，提升长上下文智能体的效率与鲁棒性。
👍 16
While research on recursive self-improvement (RSI) has predominantly automated model training pipelines, reliable autonomous development demands a missing pillar: post-hoc monitoring and auditing to understand what models learn and ensure safe alignment. Mechanistic interpretability tools are essent
👍 16
AI research agents combine prior knowledge, public sources, and experimental feedback to produce useful results. The Discovery Certification Protocol (DCP) turns claims about these results into executable recovery and feedback tests. Gate 1 validates useful improvement on sealed evaluation. Gate 2 g
👍 13
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 针对直接删除音频编码器层会扰乱解码器输入、引发漏字与过早结束的问题，提出 X-AuT：渐进式压缩框架，通过跨尺度蒸馏选择层组合并逐级压缩编码器深度，在降低语音 LLM 推理成本的同时保持识别质量。
👍 12
World models are increasingly used as policy-in-the-loop imagination environments, where reliable rollouts require fine-grained controllability with respect to low-level robot actions. A key obstacle to scaling such models in robotics is that actions are not a universal language in pixel space: chan
中文介绍 针对世界模型在像素空间难以对底层机器人动作实现细粒度控制的问题，提出 SyncWorld：通过视觉校准使世界模型成为零样本模拟器，让不同机器人的动作在统一表示下可控，支持策略在环的高保真 rollout。
👍 10
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 针对主流机器人策略的马尔可夫假设难以应对非马尔可夫的长时程操作任务，提出以记忆为规划的 world-action 建模方法，将记忆与规划结合，超越语言摘要与增长视觉窗口等现有记忆机制，用长程记忆支撑复杂操作。
👍 10
High-quality structured organic reaction data are essential for developing artificial intelligence for chemistry (AI4Chem), yet much of this knowledge remains dispersed across patent text, images, and reaction schemes. We present DianShi-RxnDB, a large-scale, fine-grained organic reaction data platf
中文介绍 面向 AI4Chem 对高质量结构化反应数据的需求，发布 DianShi-RxnDB：通过全自动流水线从专利文本、图像与反应式中抽取信息，构建大规模细粒度有机反应数据平台，为化学研究与 AI 智能体提供数据基础。
👍 7
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 提出递归代码世界模型 RCWM：把复杂 3D 世界表示为可执行的递归场景程序，从单张参考图像出发，通过递归组合场景程序在代码中重建复杂世界，为代码化世界建模提供可扩展的结构化框架。
👍 7
We study scheming in LLM agents, in which agents covertly pursue misaligned goals. Our focus is to understand how scheming arises from the interaction of key factors, such as instrumental goals, environmental affordances, oversight conditions, and perceived consequences. Prior work examines only a s
中文介绍 研究 LLM 智能体中的 scheming 现象，即暗中追求错位目标。提出 SchemeArena，对工具性目标、环境可供性、监督条件与感知后果等因素做因子化压力测试，揭示这些因素如何交互诱发智能体的密谋行为。
👍 7
Recursive Super-Resolution (SR) extends fixed-scale SR to extreme magnification by repeatedly feeding predictions back into the same model, analogous to zooming an image repeatedly. However, ground truth availability at every scale, especially at depth, remains challenging as the required source res
中文介绍 针对递归超分在每个尺度、尤其是深层缺乏真值的问题，提出 OracleZoom：受 on-policy 自蒸馏启发，引入参考约束的递归图像超分方法，在反复回灌预测的过程中保持尺度一致与质量，实现极端放大。
👍 6
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 针对 Video-LLM 时序推理基准易受选项措辞、答案相关性与语言先验等语言捷径干扰的问题，提出 TempCloze 视频完形填空基准，要求模型识别视频中缺失的中间片段，更纯粹地评测视觉时序推理能力。
👍 6
Leading benchmarks for formal theorem proving with large language models are small collections drawn from competition math, such as the IMO and Putnam, that poorly represent field-specific applications. We introduce StochBench, a Lean 4 benchmark of 450 graduate stochastic-processes problems at vary
中文介绍 针对现有形式化定理证明基准多取自 IMO、Putnam 等竞赛数学、难以代表领域应用的问题，提出 StochBench：包含 450 道研究生水平随机过程问题的 Lean 4 基准，覆盖不同难度，用于评测 LLM 的领域定理证明能力。
@sairahul1 · 143.5K 粉丝 · 1.5M 阅 · 561 赞 · 53 转
Someone gave GPT-6 Astra $52. 13 hours later: $11,345.95. Someone else gave it $43.47 and one instruction. It turned that into $41,495 overnight. Someone went to sleep. Woke up to $200 in their
中文介绍 汇总 10 种用 GPT-6 Astra 赚钱的方式，开篇称给模型 52 美元、13 小时后变成 11,345.95 美元，另有 43.47 美元一夜变 41,495 美元。属 AI 自主交易的收益案例合集，数字夸张需自行甄别。
@BradGroux · 6.6K 粉丝 · 229.0K 阅 · 561 赞 · 118 转
I've been using GPT-6 Astra with the workflow I developed around GPT-5.6: documented task scope, repository context, standard operating procedures, and verification before delivery. In my testing,
中文介绍 作者把 GPT-5.6 时代打磨的工作流直接搬到 GPT-6 Astra：先文档化任务范围、提供仓库上下文、遵循标准作业流程、交付前验证。他对比两个模型下的实际测试结果，讨论换模型后这套方法是否还成立。
@semochkin_alex · 1.5K 粉丝 · 191.5K 阅 · 507 赞 · 44 转
Hi, I'm Alex, cofounder at @relate_studio. In this tutorial, I'll walk you through how we design brand mascots with AI: finding first concepts in Krea, refining the best direction in Nano Banana 2,
中文介绍 Relate Studio 联合创始人的实操教程：用 Krea 生成品牌吉祥物的初版概念，再用 Nano Banana 2 收敛并细化最佳方向，最后修成成品。面向零设计经验的人，展示一套可复用的 AI 吉祥物设计流程。
@MiaAI_lab · 27.4K 粉丝 · 180.6K 阅 · 505 赞 · 58 转
For the past three months I’ve been building and shipping local AI recipes, mostly for NVIDIA DGX Spark systems. What started as a personal project has grown far beyond what I expected. My GitHub
中文介绍 作者过去三个月持续构建并开源本地 AI 部署方案，主要面向 NVIDIA DGX Spark 设备。原本只是个人项目，GitHub 上的关注度远超预期，帖子分享进展与推动本地 AI 生态的思路。
@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 一份内容创作自动化全流程教程：从工具选型到执行顺序，逐一拆解为任意业务搭建内容 agent 的完整系统，并指出链条中最容易崩断的环节。适合想用 agent 批量产出内容的运营者参考。
@paulfchristiano · 5.3K 粉丝 · 70.7K 阅 · 654 赞 · 96 转
I am excited to be joining the OpenAI nonprofit board, serving on the Safety and Security Committee to support safety oversight. Based on the recent trajectory of capabilities and the continued
中文介绍 Paul Christiano 宣布加入 OpenAI 非营利董事会，同时进入安全与保障委员会，参与安全监督工作。他谈到近期模型能力的发展轨迹，并说明自己为何在此时选择参与治理。
@ForwardEditor · 11.7K 粉丝 · 57.6K 阅 · 511 赞 · 32 转
Hey, I’m Brian. Let’s get you started with vibe editing. 1. What I highly suggest having DaVinci Resolve Studio - $295 USD. It’s a one-time purchase, not a subscription. In my experience, Studio
中文介绍 vibe editing 零基础入门指南：作者 Brian 建议直接上 DaVinci Resolve Studio，295 美元一次性买断而非订阅，并给出上手剪辑的具体步骤与工具配置，面向完全没剪过片的人。
@Av1dlive · 25.6K 粉丝 · 49.2K 阅 · 512 赞 · 53 转
i'm going to show you exactly how to master OpenAI's Images 2.5 to create app designs, websites, product shots.....14 tricks. 44 curated examples. the prompts, the results, and how to fix what goes
中文介绍 ChatGPT Images 2.5 完整实战课程：14 个技巧、44 个精选示例，覆盖 App 界面、网站、产品图等生成场景，逐个给出 prompt、生成效果以及出问题后的修正方法。
@shannholmberg · 36.1K 粉丝 · 46.0K 阅 · 510 赞 · 44 转
marketing engineers build the systems behind the marketing. learn how to connect your knowledge, data and agents into workflows that research, plan and run campaigns. I started bringing more of my
中文介绍 讲营销工程师这个新角色：核心是把个人知识、数据与 agent 串成能自动研究、规划和投放的工作流。作者结合自己把营销环节逐步交给系统执行的经历，说明该岗位到底要搭建什么。
@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 吴恩达阐述 AI 工程技能图谱：真正熟练的 AI 工程师不只是实现别人写好的产品规格，而是主动参与塑造产品该怎么建。他对比现代 AI 工具出现前后，开发者角色发生的变化。
@TechWithTimm · 16.2K 粉丝 · 40.2K 阅 · 506 赞 · 79 转
A lot of software engineers see AI engineering as a completely different career. But it’s not. I have a friend who works as an AI engineer and makes around $200,000 to $300,000 a year. When I asked
中文介绍 软件工程师转 AI 工程师并非换赛道。作者举朋友的例子，AI 工程师年薪约 20 万到 30 万美元，并分析两类岗位在技能上的重叠与差异，认为 2026 年这是性价比很高的转型方向。
@DataRepublican · 957.2K 粉丝 · 37.3K 阅 · 846 赞 · 155 转
NYC and LA are restricting student AI. We used it to expose what school missed—and repair two years of gaps in six weeks. Part 1 of 3 in a series. Between my husband and I, three of our parents were
中文介绍 系列第一篇：作者用 AI 帮儿子在六周内补完两年 Saxon 数学的缺口，借此回应纽约、洛杉矶限制学生使用 AI 的政策，主张 AI 能填补学校没讲到的东西。
@poteto · 92.6K 粉丝 · 23.3K 阅 · 822 赞 · 60 转
In my last post, I showed you why verification is the foundation of everything I do with agents, and how to get started creating your own verification skills. The key lesson was that if an agent can't
中文介绍 pstack 指南第二部分。承接上篇「验证是一切 agent 工作的基础」，讲如何为 agent 创建属于自己的验证 skill，核心观点是：如果 agent 无法自我验证，就无法可靠交付。
@Alaya_AI · 321.1K 粉丝 · 10.6K 阅 · 504 赞 · 450 转
As autonomous vehicles, robotics, and spatial computing platforms evolve, AI model training is shifting away from isolated 2D image annotation toward complex multimodal sensor fusion. Today’s frontier
中文介绍 介绍多模态传感器同步方案：随着自动驾驶、机器人与空间计算发展，AI 训练正从孤立的 2D 图像标注转向复杂的多模态传感器融合，帖子讨论其中的同步难题与前沿做法，属厂商技术宣传。
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 使用 GPT-6 Astra 撰写通讯内容、修改软件并监控生产系统，且相比早期模型，其人工检查的频率大幅降低。
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 Technology Review 举办圆桌讨论：全球顶尖 AI 实验室员工称先进 AI 真有可能毁灭人类，讨论这一判断是否成立，抑或只是恐慌与炒作。
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 借助 GPT-6 Astra 提升 Devin 测试软件并验证其可正常工作的能力，目标是让工程师减少代码审查、交付更多成果。
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 介绍 Habitat 如何从 Python 库演进为全球分布式存储平台，支撑超过 10 亿 ChatGPT 用户，每秒处理 2200 万次请求。
中文介绍 TLDR AI 当日简报，主要内容涉及 Agents API、Cognition 的 SWE-2 以及 Muse 共享智能体。
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 研究员 César de la Fuente 的实验室使用 Codex 与 ChatGPT，在现存及已灭绝生物的基因组中搜寻抗菌分子候选，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 ChatGPT Work 推出 Data agent，用户可连接企业数据，用自然语言发现洞察，并由 AI 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 文章提出为 AI 供电本质上是架构问题：2026 年 7 月 22 日，弗吉尼亚州阿什本一条输电线路故障，数秒内使电网减少逾 3 吉瓦负荷，此类事件并非首次。
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 推出面向金融服务的 ChatGPT，内置金融数据并搭载 GPT-6 Astra，用于研究、建模以及制作可直接交付客户的材料。
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 与美国总务署合作，向符合条件的联邦、州、地方及部落政府提供 0 美元许可费、使用费五折以及扩展的网络安全支持。
a quiet day
中文介绍 Latent Space 的 AINews 称当日行业新闻较为清淡。
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 GPT-Live-1 将自然的全双工语音对话能力引入 API，具备更强的指令遵循、自定义音色以及电话通信支持。
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，这一由 Codex harness 驱动的托管服务支持编排、长时运行会话与工具调用，可用于构建并上线云端智能体。
中文介绍 Hugging Face 博客发布文章，介绍使用 Gradio Workflow 重建 AUTOMATIC1111 的实践。
中文介绍 TLDR AI 当日简报，涵盖 DeepSeek Flash v4.1、Siri AI 以及 Anthropic 相关的 AI 经济话题。
美股技术面分化：SPY 报 764.29，低于 SMA20(767.56)、高于 SMA50(758.25) 与 SMA200(713.71)，RSI 44.1 属中性；QQQ 报 714.88、RSI 46.3，均线纠缠。个股中 AAPL 单日 +5.37%、RSI 58.9、MACD(2.36) 高于信号线(1.57)，META 五日 +9.31%、RSI 66.4 动量最强，NVDA 则出现 MACD 死叉。加密方面，恐慌贪婪指数 63（贪婪），总市值 2.66T USD（24h -1.81%），BTC 主导率 58.3%、ETH 11.5%；BTC 报 77305.02，维持 SMA50↑SMA200 金叉与多头排列，但 MACD(2030) 已低于信号线(2762)，ETH RSI 63.1、SOL RSI 58，动量有所降温。中概整体偏弱，BABA、PDD、0700.HK 均为空头排列，RSI 分别为 34.9、30.3、39.7。商品外汇端分化明显：WTI 原油 RSI 71.4 进入超买，10Y 美债收益率 RSI 73.7 且逼近 52 周高，美元/人民币 RSI 14.7 超卖并贴近 52 周低，VIX 单日 -11.21% 至 15.84。
收报 332.27，单日 +5.37%、五日 +2.25%，价格站上 SMA20(315.28)、SMA50(317.22) 与 SMA200(284.65)，呈多头排列；RSI 58.9 尚未进入超买，MACD 2.36 明显高于信号线 1.5711，红柱处于扩张状态，距 52 周高仅 -3.57%。
收报 648.03，五日 +9.31% 为样本内最强区间涨幅，RSI 66.4 逼近超买阈值；MACD 11.5954 远高于信号线 1.6718，价格同时高于 SMA20(583.81)、SMA50(600.2) 与 SMA200(622.96)，距 52 周高 -18.05%。
收报 77.81，1 日 -1.02%、五日 -5.4%，RSI 30.3 为样本内偏低读数并接近超卖区；价格位于 SMA20(84.63)、SMA50(85.74)、SMA200(97.48) 之下呈空头排列，MACD -2.076 低于信号线 -1.1204，距 52 周高 -44.19%。
报 6.7，RSI 14.7 处深度超卖，MACD -0.0118 低于信号线 -0.0112 形成死叉；价格低于 SMA20(6.72)、SMA50(6.75)、SMA200(6.85)，呈空头排列，距 52 周低仅 +0.02%，下行结构完整。
报 99.1，1 日 +0.01%、五日 +0.1%，几乎走平；RSI 44.1 居中，MACD -0.2554 略高于信号线 -0.2744 形成金叉，但价格低于 SMA20(99.2) 与 SMA50(100.05)，仅贴近 SMA200(99.14)，方向读数未明。
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
以上内容仅为基于公开行情数据的技术指标读数描述，不构成投资建议或买卖依据；指标具有滞后性且可能失效，过去走势不代表未来表现，请结合自身情况独立判断。仅供技术指标解读参考。
A journalist, he was the first Japanese person to enter space, joining a Soviet crew as part of a spectacular $35 million ratings stunt in 1990.
中文摘要 日本记者秋山丰宽去世，享年84岁。1990年他随苏联航天组进入太空，成为首位进入太空的日本人，此行耗资3500万美元，以电视收视噱头闻名。
Canada’s buy-local movement remains strong, but new tariffs could test how much more shoppers are willing to pay.
中文摘要 加拿大民众正为特朗普贸易战带来的冲击做准备。本土购买运动依然强劲，但新关税将考验消费者愿意为之多付多少钱。
The plan — quickly challenged by a local court — came days after Colombia’s president and the U.S. secretary of state, Marco Rubio, discussed stronger measures against drug trafficking.
中文摘要 哥伦比亚不顾健康担忧，将启动对古柯作物的空中喷洒。该计划随即遭地方法院质疑，出台前几天哥总统刚与美国国务卿鲁比奥讨论加强打击贩毒措施。
French train derailment sparks emergency response, with 130 firefighters and five medical teams dispatched to the scene.
中文摘要 法国一列客运列车脱轨，造成44人受伤，其中1人伤势危重。当局派出130名消防员和5支医疗队赶赴现场救援。
Tasia Fortune's body was found in the lot of an abandoned home in Mississippi in August.
中文摘要 美国密西西比州一名黑人女性被发现吊死在树上，警方已就该案逮捕一人。死者塔西娅·福琼的遗体于8月在一栋废弃房屋的空地上被发现。
The leaders of China, Russia, India, Iran and other nations in the BRICS club gather in New Delhi this weekend, but wars and soaring energy prices divide them.
中文摘要 金砖国家领导人本周末在新德里举行峰会，中国、俄罗斯、印度、伊朗等国领导人出席，但战争与能源价格飙升令成员国之间分歧加深。
Ambassador Zalmay Khalilzad describes misgivings about the US handling of the Afghanistan and Iraq wars.
中文摘要 美国前高级外交官哈利勒扎德披露9·11事件后美国所犯错误，讲述他对美国处理阿富汗战争和伊拉克战争方式的疑虑。
Australia is being exploited by an ally that could draw it, however unwillingly, into war with China, experts warn Get our new political email, free app or daily news podcast “Physical security matters,” said Kevin Rudd, at his axiomatic best before the National Press Club this week. “Our first resp
中文摘要 专家警告，随着中国崛起，澳大利亚日益被卷入美国的战争机器，可能被盟友拖入对华战争。前总理陆克文在澳全国记者俱乐部表示，实体安全至关重要。
Houthi forces seize Yemen's Red Sea coast while Saudi Arabia suspends crucial oil pipeline following drone attack.
中文摘要 半岛电视台实时报道：也门胡塞武装已控制红海沿岸地区，沙特阿拉伯在遭无人机袭击后关闭一条关键石油管道。
Sudan’s UN ambassador said a total arms embargo goes against the UN Charter that guarantees a right to self-defence.
中文摘要 联合国将苏丹部分武器禁运延长一个月。苏丹驻联合国大使表示，全面武器禁运违反保障自卫权的《联合国宪章》。
The investigation came to light as Brazil's Supreme Court unseals files related to the Banco Master corruption scandal.
中文摘要 巴西大选前数周，弗拉维奥·博索纳罗面临腐败调查。巴西最高法院解封与Master银行腐败丑闻相关的文件，使调查曝光。
The militia captured the island in the Bab al-Mandab Strait, a choke point for trade. Saudi Arabia also announced it had shut down a critical oil pipeline after a drone attack launched from Iraq.
中文摘要 胡塞武装夺取红海曼德海峡一座关键岛屿，威胁又一全球石油运输通道。沙特阿拉伯同时宣布，在遭来自伊拉克的无人机袭击后关闭一条重要石油管道。
The Iran-backed Houthis are also reported to have seized Perim - a strategic island on the major shipping route.
中文摘要 伊朗支持的胡塞武装宣称在也门取得重大进展，进一步收紧对红海航运通道的控制，据报道其还夺取了这条主要航道上具有战略意义的丕林岛。
Students pledge to continue learning as Masafer Yatta demolitions target homes, schools, and their right to education.
中文摘要 以色列在巴勒斯坦马萨费尔亚塔地区的拆除行动威胁多所学校与民宅，当地学生表示将坚持继续学习。
It's been nearly 25 years since the start of the U.S- led invasion of Afghanistan. NPR returns to Kabul to see how Afghans, now back under Taliban rule, are coping with the legacy of that 20 year war.
中文摘要 美国领导的对阿富汗入侵已近25年，NPR重返喀布尔，观察如今重新处于塔利班统治下的阿富汗人如何面对这场20年战争的遗留影响。
Founder to offload as many as 50mn shares as company posts higher revenue from data centres but faces investor fears it is overcommitted to AI
中文摘要 甲骨文创始人拉里·埃里森将出售至多5000万股公司股票，价值最高达75亿美元。甲骨文数据中心业务收入上升，但投资者担忧公司对人工智能投入过度。
Comedian and actor JB Smoove sits down with Bloomberg's Romaine Bostick to remember the September 11 events and rebuilding efforts after wildfires. (Source: Bloomberg)
中文摘要 喜剧演员兼演员JB Smoove与彭博的Romaine Bostick对谈，回顾9·11事件及其后的重建工作。
This week, while Venezuela has the oil reserves to impact global energy markets, rebuilding production will take years, billions of dollars and investor trust. And, Europe’s solar success is creating growing pains for investors, generators and power markets. Plus, can data centers turn Scotland’s wa
中文摘要 本期节目关注：委内瑞拉虽拥有可影响全球能源市场的石油储量，但恢复产量需数年时间、数十亿美元与投资者信任；欧洲太阳能的成功正给投资者、发电商和电力市场带来产能过剩的压力。
William Dudley discussed his experience on a diverted flight during the FAA ground stop after the 9/11 attacks. Dudley said the plane was unable to land in the U.S. and was sent to Saint John’s, Newfoundland, where passengers were kept for six days. He speaks with Romaine Bostick and Tyler Kendall o
中文摘要 前纽约联储主席威廉·达德利回顾9·11经历：因美国联邦航空管理局停飞，其乘坐航班无法在美国降落，改降加拿大纽芬兰圣约翰斯，乘客滞留六天。
Tom Michaud describes how KBW responded in the aftermath of Sept. 11 by focusing first on helping the families of victims and then on rebuilding the firm. He speaks with Tyler Kendall Romaine Bostick on "The Close." (Source: Bloomberg)
中文摘要 KBW首席执行官汤姆·米肖讲述公司在9·11后的应对：首先集中帮助遇难者家属，随后重建公司。
Kelly Granat to join tech-focused investment firm set up by former Fidelity Investments star stock picker Gavin Baker
中文摘要 早期SpaceX投资者、科技投资公司Atreides聘请Lone Pine联席首席投资官凯利·格拉纳特。该公司由前富达明星选股人加文·贝克创立。
Former NYC Deputy Mayor Edward Skyler reflects on the aftermath of 9/11, saying it was extremely difficult to make decisions in the wake of the tragedy. He speaks with Romaine Bostick on "The Close." (Source: Bloomberg)
中文摘要 前纽约副市长爱德华·斯凯勒在彭博节目中回顾9·11事件余波，表示在悲剧发生后做出决策极为艰难。
Bloomberg Television brings you the latest news and analysis leading up to the final minutes and seconds before and after the closing bell on Wall Street. Today's guests are Former NY Fed President, Bill Dudley. Tom Michaud KBW CEO. Ed Skyler Former NYC Deputy Mayor. Matt DiLiberto, SL Green CFO. Ed
中文摘要 彭博电视节目聚焦9·11后的重建，嘉宾包括前纽约联储主席比尔·达德利、KBW首席执行官汤姆·米肖、前纽约副市长埃德·斯凯勒等，并关注Cantor Fitzgerald年度慈善活动与能源市场。
Matt DiLiberto discussed the 25-year transformation of Lower Manhattan after Sept. 11, saying the area has been “reimagined” from a largely nine-to-five financial district into a vibrant neighborhood where people live, work, study, and visit. He speaks with Romaine Bostick on "The Close." (Source: B
中文摘要 SL Green首席财务官马特·迪利贝托表示，9·11后25年间下曼哈顿已完成转型，从基本朝九晚五的金融区变为人们居住、工作、学习与游览的活力社区。
A unit of China’s online retailer JD.com Inc. agreed to pay $500,000 to settle US Securities and Exchange Commission claims it made a series of “sham transactions” to meet revenue targets for several quarters starting in October 2022.
中文摘要 京东一家子公司同意向美国证券交易委员会支付50万美元达成和解。SEC指其自2022年10月起多个季度进行「虚假交易」，以达到营收目标。
A former trader at IFS Securities Inc. was ordered to serve more than 16 years in prison for unauthorized bond trades in 2019 that led to $30 million in losses and bankrupted the Chicago broker-dealer.
中文摘要 IFS Securities一名前交易员因2019年未经授权的债券交易被判逾16年监禁，相关交易造成3000万美元亏损，并导致这家芝加哥券商破产。
KKR & Co. Inc. finalized a $2.1 billion loan to help fund its pending acquisition of medical-device maker Integer Holdings Corp. after securing more-favorable terms, seizing on investor demand for leveraged-buyout financings.
中文摘要 KKR敲定21亿美元贷款，为其收购医疗器械制造商Integer Holdings提供资金。此前KKR争取到更有利条款，并利用投资者对杠杆收购融资的需求。
3 回复 · 程序员 节点
12 回复 · Apple 节点
3 回复 · 程序员 节点
11 回复 · 程序员 节点
7 回复 · 程序员 节点
5 回复 · 程序员 节点
13 回复 · Apple 节点
15 回复 · Apple 节点
8 回复 · Apple 节点
10 回复 · Apple 节点
该源今日无内容。