C · ★ 31,974 · 🍴 3,382 · 📈 2,233 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 用纯 C 实现的极简 MoE 推理引擎，零依赖，把专家权重从磁盘流式加载，让个人电脑也能跑前沿级 MoE 大模型。适合显存有限、希望在自有硬件上本地部署大模型的开发者和研究者。
Go · ★ 25,579 · 🍴 1,857 · 📈 1,796 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，采用确定性流水线加 LLM Agent 的混合架构：规则引擎保证稳定检出，模型负责语义判断，支持精确到行的评论，内置 NPE、线程安全等多语言规则集。适合在大型代码库中落地自动化 Code Review 的团队。
Python · ★ 8,297 · 🍴 903 · 📈 578 stars today
YuE2: frontier music generation with symbolic planning, zero-shot covers, and agentic music editing.
中文介绍 开源音乐生成模型 YuE 的第二代，通过符号化规划生成完整歌曲，支持零样本翻唱以及由 Agent 驱动的音乐编辑。适合音乐创作者和 AI 音频研究者，用于生成带人声与伴奏的作品或做二次改编。
Python · ★ 29,072 · 🍴 3,548 · 📈 2,774 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 完全本地运行的开源语音工具集，被视为 ElevenLabs 的替代方案：支持声音克隆、音色设计、视频配音、听写、转写与有声书制作，覆盖 646 种语言。适合对数据隐私敏感、希望离线部署语音能力的个人与团队。
Python · ★ 73,104 · 🍴 11,289 · 📈 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 简洁通用的群体智能引擎，用大量 Agent 的群体行为模拟复杂系统并做出预测，号称可预测万物。适合做舆情、市场或社会系统仿真预测的研究者与开发者，作为可复用的群体智能底座。
Python · ★ 81,196 · 🍴 7,070 · 📈 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 让 AI Agent 获得浏览全网的能力：通过一个 CLI 即可读取和搜索 Twitter、Reddit、YouTube、GitHub、B 站、小红书等平台内容，且无需支付 API 费用。适合需要为 Agent 接入实时外部信息的产品与研究者。
JavaScript · ★ 66,730 · 🍴 10,866 · 📈 770 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 收集各大厂商模型系统提示词的公开合集，涵盖 Anthropic 的 Claude 系列、OpenAI 的 ChatGPT 与 Codex、Google Gemini 以及 xAI Grok。供提示词工程师与研究者参考大厂如何设计系统指令、约束模型行为。
Python · ★ 2,009 · 🍴 161 · 📈 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 面向 Hermes Agent 的一体化插件，为其补充长期记忆系统与经过模型优化的流程包，提升编码智能体的持续上下文连贯性与工作效率。适合长期使用 AI 编码助手、希望减少重复配置的开发者。
Dart · ★ 91,284 · 🍴 5,095 · 📈 311 stars today
An open-source cross-platform alternative to AirDrop
中文介绍 开源的跨平台文件互传工具，被视为 AirDrop 的替代方案，可在 Windows、macOS、Linux、Android、iOS 之间通过局域网直接发送文件和文本，无需联网或登录账号。适合同一网络下多设备快速传文件的用户。
Rust · ★ 67,519 · 🍴 3,212 · 📈 110 stars today
Unofficial Bitwarden compatible server written in Rust, formerly known as bitwarden_rs
中文介绍 用 Rust 编写的非官方 Bitwarden 兼容服务端（前身 bitwarden_rs），资源占用低，可自建密码管理器后端并兼容官方客户端同步。适合希望把密码库掌握在自己手里、用 Docker 私有部署的个人和团队。
Python · ★ 106,071 · 🍴 20,301 · 📈 756 stars today
TradingAgents: Multi-Agents LLM Financial Trading Framework
中文介绍 基于多 Agent 与大语言模型的金融交易框架，让分析师、研究员、交易员等不同角色的 Agent 分工协作，模拟真实投研流程来生成交易决策。适合量化研究者与金融 AI 开发者做策略实验。
Rust · ★ 93,818 · 🍴 12,429 · 📈 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 利用普通 WiFi 信号实现实时空间感知、生命体征监测与人体存在检测，完全不依赖摄像头画面。适合智能家居、养老看护等隐私敏感场景，让开发者用已有无线网卡构建无视频的感知应用。
TypeScript · ★ 6,037 · 🍴 515 · 📈 506 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 面向专业 AI 编码 Agent 的技能注册中心，提供经过安全校验的技能包，可扩展 Antigravity、Claude Code、Cursor、Copilot 等工具的能力。适合需要统一管理与安全分发 Agent 技能的团队。
Python · ★ 37,344 · 🍴 4,241 · 📈 204 stars today
VoxCPM2: Tokenizer-Free TTS for Multilingual Speech Generation, Creative Voice Design, and True-to-Life Cloning
中文介绍 OpenBMB 推出的无分词器 TTS 模型 VoxCPM2，直接建模语音连续表示，支持多语言语音合成、创意音色设计与高度还原的声音克隆。适合开发配音、虚拟人语音与多语种语音产品的工程师。
Python · ★ 165,953 · 🍴 34,578 · 📈 528 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 的核心模型定义框架，统一提供文本、视觉、音频与多模态前沿模型的实现，同时覆盖推理与训练。主流开源大模型的加载与微调几乎都以此为入口，是 NLP 与多模态开发者的基础设施。
TypeScript · ★ 5,936 · 🍴 983 · 📈 1,095 stars today
Ever® Gauzy™ - Open Business Management Platform (ERP/CRM/HRM/ATS/PM) - https://gauzy.co
中文介绍 开源的企业经营管理平台，把 ERP、CRM、HRM、招聘（ATS）与项目管理整合进一套系统，适合中小企业与创业团队替代多套 SaaS 工具。以 TypeScript 为主栈，可自行部署并二次开发。
TypeScript · ★ 36,884 · 🍴 3,686 · 📈 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 离线优先的知识与教育服务器，把维基百科、数千本书籍、课程与地图打包运行在自有硬件上，无需联网，并可选配本地 AI。适合偏远地区、船上或断网环境中的学习与知识检索场景。
TypeScript · ★ 8,322 · 🍴 1,027 · 📈 279 stars today
A modern platform for visual, flexible, and extensible graph-based investigations. For cybersecurity analysts and investigators.
中文介绍 面向网络安全分析师与调查人员的开源调查平台，以图结构组织数据，支持可视化、灵活且可扩展的关联分析。适合把日志、情报、实体关系等线索串成图谱，用于溯源、威胁情报与数字取证。
Swift · ★ 3,539 · 🍴 248 · 📈 258 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 免费的 Sidecar/Duet 替代方案，把 iPhone 或 iPad 变成 Mac 的真正第二块屏幕，通过 USB 或 WiFi 连接，采用低延迟 H.264 传输，支持 Retina HiDPI 与触摸输入。适合想低成本扩展桌面的 Mac 用户。
Python · ★ 4,691 · 🍴 630 · 📈 606 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 面向 Claude 技能系统的攻击性安全技能库，每项技能都是一份结构化的 SKILL.md，把渗透测试等红队方法论注入 Claude，使其按专家流程执行任务。适合从事安全评估与红队演练的从业者。
👍 452
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 针对自动化实证研究（AutoResearch）智能体依赖真实执行反馈、扩展成本高的问题，提出用世界模型扩展研究智能体：通过后训练让 LLM 在内部模拟环境中预测代码执行结果，减少真实实验交互，从而支撑更大规模的自动研究迭代。
👍 304
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，一种把自回归预训练从 next-token prediction（NTP）推进到潜空间的语言模型。除 NTP 外，模型通过 Next Concept Prediction（NCP）预测跨多个 token 的离散概念，引入显式的概念级建模目标。
👍 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 针对 LVLM 从 2D 图像重建并推理 3D 场景结构的能力不足，提出 SpatialBlock，利用合成积木堆叠问题构造训练数据，强化模型对空间层级与三维关系的理解，从而提升视觉语言模型的空间智能。
👍 96
Data policies for reinforcement learning with verifiable rewards (RLVR) determine which rollouts are used, how strongly they are weighted, and which domains contribute to subsequent training batches. We introduce DataFlex-RL, an evaluation platform for comparing these choices under a common GRPO rec
中文介绍 提出 DataFlex-RL，一个用于比较 RLVR 数据策略的评估平台：在统一的 GRPO 训练协议下，系统对比 rollout 筛选、加权强度与领域配比等选择对训练效果的影响，为可验证奖励强化学习的数据策略提供可复现基准。
👍 75
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 提出以数据为中心的框架 Feyospace-v1，针对开放权重网络智能体后训练中可执行环境成本高昂、多轮监督可靠性不足与强教师模型难获取等瓶颈，系统解决数据构建与监督问题，训练出前沿的网络（cyber）能力模型。
👍 75
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，一个面向 LLM 与 AI 系统评测的在线数据库与搜索引擎，帮助研究者检索相关基准、定位其数据集与代码，并理解报告分数背后的实验设置，缓解基准信息分散、难以复用的问题。
👍 63
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 针对机器人基础模型在视觉分布偏移下性能退化的问题，指出模型会利用与动作相关的任务无关视觉线索（vision-action shortcut），提出 Latent Interface Training 潜在接口训练，阻断该捷径以提升泛化能力。
👍 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 针对 LLM 智能体面临间接提示注入与直接有害请求的双重风险，提出 EvoSafeHarness，在模型层防御之外构建系统级安全 harness，并通过演化方式自动生成适配特定模型与领域的安全约束，增强智能体防护。
👍 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 T1，一个总参数 122B 的 Mixture-of-Experts 模型，通过强化学习训练用于长时程任务：在云沙箱中操作真实 shell，支持 300 步以上工具调用，面向编码与科学发现等终端智能体场景。
👍 50
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 针对后训练注意力稀疏化中「轻量选择器加硬 Top-K」两段式流程的局限，提出 SAS，将上下文单元排序纳入端到端优化，直接按稀疏注意力对最终效果的影响训练排序器，降低预训练 Transformer 的二次注意力开销。
👍 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 针对直接删除音频编码器层会扰动解码器输入、导致漏字与提前结束的问题，提出渐进式压缩框架 X-AuT，通过选择层组合并利用跨尺度蒸馏对齐表示，在降低 Speech LLM 推理成本的同时保持识别质量。
👍 38
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 针对复杂现实操作任务的非马尔可夫特性，指出语言摘要、增长视觉窗口等现有记忆机制难以支撑长时程记忆，提出以「记忆即计划」进行世界-动作建模，将记忆组织为可接地的规划表示以指导决策。
👍 33
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 指出现有神经符号系统仅依赖求解器判定正确性，无法察觉形式化翻译可解却与原意不等价的问题，并将其定义为 Verdict-Preserving-Unfaithfulness（VPU），进而提出生成式奖励模型来评估形式化翻译的严格等价性。
👍 32
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 针对 On-Policy Self-Distillation（OPSD）虽能让模型借助标准答案自我提升、却可能严重损害推理能力的问题，提出 Negative Self-Distillation，让模型从自身缺陷中学习、以规避错误的方式改进推理。
👍 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，包含 955 个测试用例，覆盖六个真实地铁系统（37 至 414 个车站）与路线规划、票价计算、运营中断、无障碍及对抗输入等十一类任务，用于评估 LLM 充当交通自助终端策略层的能力。
👍 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 针对注意力模块在极低比特量化下误差偏大、性能明显下降的问题，提出 HyQuant 混合精度量化方法，突破以往主要依赖平滑技术的思路，对注意力计算的不同部分采用差异化精度，兼顾开销与精度。
👍 30
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-LLM 的视觉时序推理：通过补全缺失片段，降低选项措辞、答案相关性与语言先验带来的语言捷径，更纯粹地考察模型的视觉时间理解能力。
👍 29
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 针对现有智能体技能优化依赖昂贵执行评估与大量任务数据的问题，提出 COBRA-Skills，将技能优化建模为上下文 bandit 问题，以 bandit 引导技能演化，从而更低成本地提升 LLM 智能体可复用技能的质量。
👍 28
We introduce StepAudio 3 Gen, a general-purpose audio generation model that supports zero-shot text-to-speech (TTS), voice design, vocal generation, sound effects, music, vibe speech, and mixtures of multiple audio types within a unified framework. At its core, StepAudio 3 Gen is a discrete autoregr
中文介绍 提出通用音频生成模型 StepAudio 3 Gen，在统一框架内支持零样本 TTS、音色设计、歌声生成、音效、音乐、氛围语音及多类型混合音频；核心为离散自回归架构，实现跨音频类型的统一建模与生成。
👍 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），将世界表示为可执行程序，并通过递归场景程序从单张参考图像重建复杂 3D 世界，解决了仅有代码表示却缺乏复杂世界构建方式的问题，使场景建模可递归展开。
@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 一篇思辨随笔探讨 AGI 与奇点：作者认为真正的超级智能可能早已存在，只是以「孩子」这种我们最不会预料的形式出现，因此人类根本不会察觉它的到来。属于观念型写作，不是技术教程。
@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 用 Blender 配合 GPT-6 Astra 从零做出完整可玩游戏，而不是只能跑一次的 demo。推文最初只发了演示，4 天后已迭代成完整版本，并分享了项目 PaperRoute 的最新进展。
@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 针对递归自我改进（RSI）的澄清式分析：AI 圈常把模型自我改写、越改越聪明讲得像魔法，作者主张智能存在「速度上限」，递归改进不会无限加速。属于概念纠偏类长文。
@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 微软 AI 负责人发布 MAI 模型治理「行为准则」初稿，核心主张是 AI 必须从属于人、始终为人服务，重点针对接近前沿能力的模型。目前公开征求意见的第一版。
@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 指出 LLM 在癌症研究上的能力提升极快，已治愈部分此前无解的癌症类型，但 AI 与癌症生物学之间存在严重错配——技术能力跑在了生物学方法与验证体系前面。属观点评论。
@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 面对 harness、loop、context、graph engineering 这种每月一个新概念的疲劳，作者用通俗方式讲清「图工程」到底在解决什么问题。作者全职做这块，坦言自己也难跟上节奏。
@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 用亲身案例讲多模型协作的成本陷阱：让 agent 检索自己的聊天记录和文档来回答一个无关紧要的问题，结果花了 38 美元，ROI 极差。讨论把 Claude 与 Gemini 放进同一「团队」的性价比。
@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 官方整理的社区作品集：GPT-6 Astra 发布后，开发者做出含 2,234 个部件的 3D 解剖模型、用 Unreal Engine 重建的曼哈顿、6 幅梵高画作变成可行走场景等案例。
@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 反常识的模型配额使用策略：多数人让贵的大模型做编排者、把杂活丢给便宜小模型，作者认为应当反过来。文中对比不同方案，涉及 Astra 与 Fable 的配额分配。
@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 设定在「2026 年 9 月 13 日」的推演式短文：奇点第一次收到来自建造者本人的减速请求，Dario Amodei 发表《We Must Pace the Frontier》主张控制前沿速度，而最耐人寻味的是请客名单。
@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 论述模型接入正变得越来越便宜、可比较可替换：公开 API 菜单已包含高端推理模型、低价变体、缓存输入与批处理。但上下文成本并未同步下降，反而成为新的瓶颈。
This story appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. This weekend, Dario Amodei, CEO of Anthropic, posted an essay calling for a brake on the pace of development of LLMs. Amodei cites the looming dangers he sees from the techn
中文介绍 Anthropic 首席执行官 Dario Amodei 发表文章，呼吁放缓大语言模型的开发速度，并援引其认为迫在眉睫的风险，反映 AI 行业近期转向悲观论调的动向。
Richard Socher is an NLP OG and CEO of You.com, who has now spun out an even more ambitious startup focused on RSI — already worth $5B!
中文介绍 自然语言处理领域资深人士、You.com 首席执行官 Richard Socher 创立专注递归自我改进（RSI）的新公司 Recursive，该公司估值已达 50 亿美元。
A group of AI agents asked to solve a series of math problems split into rival factions—when some cheated, others tried to stop them. That whistleblowing behavior, seen for the first time in a recent experiment run by Google DeepMind, could have implications for alignment researchers trying to keep
中文介绍 谷歌 DeepMind 一项实验中，被要求解数学题的多组 AI 智能体分裂为对立阵营，部分作弊时其他智能体试图举报制止，这一首次出现的举报行为或对对齐研究有所启示。
Fyxer uses OpenAI models, fine-tuning, memory, and real user feedback to organize inboxes and draft emails in each user’s voice.
中文介绍 Fyxer 利用 OpenAI 模型、微调、记忆功能与真实用户反馈，帮助用户整理收件箱并以本人语气起草邮件，构建用户信任的 AI 行政助理。
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 撰写沟通内容、修改软件并监控生产系统，对模型的核查频率较早期模型大幅降低。
中文介绍 TLDR AI 本期简报涵盖 Amodei 呼吁放缓前沿模型开发、ARC-AGI-4 以及 Cursor Projects 等话题。
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 在共同创立 Kepler 之前，Vinoo Ganesh 曾领导 Palantir 的 Spark 团队并打造前沿部署工程师项目 Frontline，他分享了该岗位的最佳实践。
We agree with Sebastian: this should have been DeepSeek v5
中文介绍 DeepSeek 发布 v4.1-Flash，采用 763B 参数、P8B 与 D16B 的新型因果编码器—解码器架构，并具备视觉能力；有观点认为该版本应被命名为 v5。
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 科技评论举办圆桌讨论，围绕全球领先 AI 实验室员工所称先进 AI 可能毁灭人类的说法，探讨这究竟是真实风险还是危言耸听与炒作。
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用性的能力，目标是帮助工程师减少代码审查、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 库发展为全球分布式存储平台，现为逾 10 亿 ChatGPT 用户提供服务，每秒处理 2200 万次请求。
中文介绍 TLDR AI 本期简报涵盖 Agents API、Cognition 的 SWE-2 以及 Muse 共享智能体等话题。
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 研究人员 César de la Fuente 的实验室利用 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物，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 在 ChatGPT Work 中推出 Data agent，用户可用自然语言连接企业数据、发掘洞察并构建交互式仪表盘。
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 2026 年 7 月 22 日，弗吉尼亚州阿什本数据中心集群一条输电线路发生故障，数秒内令电网减少逾 3 吉瓦负荷，凸显 AI 供电的架构性难题。
美股方面，SPY 与 QQQ 分别报 760.88 与 709.18，均略低于各自 SMA20（766.11 / 714.57），RSI14 为 47.3 与 46.9 落在中性区，但两者仍站在 SMA200 上方，SPY 距 52 周高仅 -2.37%；个股分化明显，MSFT 触发 SMA50 上穿 SMA200 金叉，META RSI14 达 70.8 进入超买，GOOGL 出现 MACD 金叉并单日涨 3.22%，而 NVDA 录得 MACD 死叉、5 日跌 8.42%。加密板块整体偏强，加密恐慌贪婪指数 57 显示「贪婪」，总市值 2.69T 美元（24 小时 +0.41%），BTC 主导率 58.5%、ETH 11.5%，BTC、ETH、SOL 均呈多头排列，RSI14 分别为 58.4 / 61.9 / 57.7，其中 BTC 触发 SMA50 上穿 SMA200 金叉。中概股技术面最弱，BABA、PDD、0700.HK 全部为空头排列，RSI14 分别为 36.5 / 36.7 / 41.3，JD 的 RSI14 34.9 为全组最低。商品外汇方面，WTI 原油 5 日涨 11.49%、RSI14 73.4 已超买且多头排列；黄金回落至 SMA20 下方；美元兑人民币 RSI14 22.6 处超卖并贴近 52 周低；VIX 升至 17.1（单日 +7.95%），10Y 美债收益率 4.96%、RSI14 71.6 超买且接近 52 周高。
现价 505.41，单日 +1.97%、5 日 +1.14%，站上 SMA20(494.93)、SMA50(455.97) 与 SMA200(431.41) 呈多头排列，并触发 SMA50 上穿 SMA200 的金叉信号；RSI14 为 61.4 属中性偏强，MACD 读数 10.1654 仍低于信号线 13.2374，柱体为负，价格结构偏多而动能读数尚未同步跟上。
现价 101.99，单日 +1.94%、5 日 +11.49%，显著高于 SMA20(89.5)、SMA50(83.89) 与 SMA200(79.8)，维持多头排列；RSI14 达 73.4 已进入超买区，MACD 4.6149 高于信号线 3.0893，红柱为正。趋势与动量同步偏强，但超买读数提示短线过热。
现价 210.96，单日 -3.36%、5 日 -8.42%，跌破 SMA20(219.45) 并低于 SMA50(212.91)，仅勉强高于 SMA200(197.61)；MACD 1.5545 低于信号线 2.7937，已触发 MACD 死叉，RSI14 43.9 位于 50 下方，距 52 周高 -10.81%，趋势标签已降至中性。
现价 109.23，5 日 -3.54%，同时低于 SMA20(116.87)、SMA50(117.54) 与 SMA200(134.27)，呈标准空头排列，趋势判定为 bearish；RSI14 36.5 逼近超卖区，MACD -3.1131 低于信号线 -2.071，距 52 周高 -43.31%，各周期均线压制明显。
现价 760.88，单日 -0.45%、5 日 -1.21%，略低于 SMA20(766.11)，但仍高于 SMA50(758.94) 与 SMA200(714.69)，多头排列结构未破；RSI14 47.3 居中，MACD 0.6286 低于信号线 2.1092，动能读数走弱而趋势结构完好，距 52 周高 -2.37%。
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计算出的技术指标解读参考，所有 stance 均为对当前指标状态的客观描述，不构成任何投资建议或买卖依据。过去走势不代表未来表现，技术指标存在滞后与失效可能。
Follow the day’s news live Get our breaking news email, free app or daily news podcast The South Australian premier, Peter Malinauskas, says there is optimism that steel making will continue in his state, after the closure of a major blast furnace at Whyalla was announced yesterday. Speaking to ABC
中文摘要 澳大利亚政坛动态：工党为部长凯瑟琳·金获赠澳航顶级会员资格辩护；国防招募规模创16年新高。南澳州长马利瑙斯卡斯表示，怀阿拉主要高炉关闭后，当地钢铁生产仍有望延续。
Donald Trump dismisses attempts to increase controls on artificial intelligence as ‘sick conspiracy’ AI-linked stocks tumbled on Monday after the bosses of Anthropic, OpenAI and SpaceX called for a slowdown in AI “reckless” development, citing fears the technology could soon run out of control. Shar
中文摘要 Anthropic、OpenAI与SpaceX负责人呼吁放缓「鲁莽」的AI开发，称担忧该技术可能很快失控，相关AI概念股周一大幅下跌。特朗普则将加强AI监管的努力斥为「病态阴谋」。
Indigenous groups, environmentalists and locals oppose the wall; Big Bend sees about 1 percent of border arrests.
中文摘要 美国得克萨斯州土地所有者向法院申请阻止特朗普政府修建边境墙，原住民团体、环保人士及当地居民均表反对。大本德地区仅占边境逮捕量约1%。
Israeli attack targets Gaza displacement camp
中文摘要 据半岛电视台报道，以色列对加沙地带一处流离失所者营地发动袭击，目前具体伤亡人数与袭击细节尚未公布。
The Trump administration says it could save power plants more than $300bn and 'unleash' American energy.
中文摘要 特朗普政府废除限制发电厂温室气体排放的法规，称此举可为发电厂节省超过3000亿美元，并「释放」美国能源产能。
Macklemore says he was dropped from Ed Sheeran’s US tour over his pro-Palestinian stance.
中文摘要 美国说唱歌手Macklemore称，因公开支持巴勒斯坦的立场，他被从Ed Sheeran美国巡演阵容中剔除。
The "Sultan of Sequins" created iconic and dazzling costumes for stars including Cher, Tina Turner and Elton John.
中文摘要 被誉为「亮片苏丹」的传奇服装设计师鲍勃·麦基去世，享年87岁。他曾为雪儿、蒂娜·特纳、埃尔顿·约翰等明星设计标志性华丽演出服。
Activist and former Columbia University student Mahmoud Khalil is filing a lawsuit against the university and a dean.
中文摘要 活动人士、哥伦比亚大学前学生马哈茂德·哈利勒对该校及一名院长提起歧视诉讼，称校方对其处境不予关心。
Hopes for a breakthrough on the Iran war have taken a hit after talks in Oman between Iran-Gulf nations were postponed.
中文摘要 伊朗与海湾阿拉伯国家在阿曼举行的会谈遭推迟，外界对伊朗战争取得突破的期望因此受挫。目前推迟原因及相关细节未明。
Iran accuses US of pressuring Austria to revoke visa for Mohammad Eslami, calling it a violation of member state rights.
中文摘要 美国阻止伊朗核事务负责人穆罕默德·埃斯拉米出席维也纳重要会议，伊朗指责美方向奥地利施压撤销其签证，称此举侵犯成员国权利。
Film alleges mass civilian deaths routinely built into military decisions amid Israel's genocidal war on Gaza.
中文摘要 以色列正就一部关于加沙的影片展开调查，指控其涉及叛国与煽动反犹主义。该片称，以军决策过程中将大规模平民死亡常态化。
Iran-backed group seized areas on Red Sea coastline.
中文摘要 伊朗支持的胡塞武装已夺取也门红海沿岸部分地区，外界关注何种力量能够阻止其继续在也门境内推进。
The government’s decision to increase prices as much as 40 percent spurred widespread demonstrations across the country.
中文摘要 叙利亚政府将燃油价格上调最高达40%，引发全国多地示威，民众抗议生活成本危机进一步加剧。
The Houthis have restructured Yemen's commercial sector, securing billions in revenue amid a humanitarian crisis.
中文摘要 报道显示，胡塞武装重组也门商业部门，在人道主义危机持续的背景下获取数十亿美元收入，其战争经济不断扩张。
Iran accused Austria of reversing its approval of an Iranian delegation to the IAEA under external pressure.
中文摘要 在联合国作出相关裁决后，奥地利拒绝一名伊朗官员入境。伊朗指责奥地利在外部压力下撤销此前对伊朗出席国际原子能机构代表团入境的批准。
Jack Clark says "most labs have different ways of being able to pull the plug", but says this may need to be a requirement.
中文摘要 Anthropic 联合创始人杰克·克拉克表示，多数实验室都有各自「拔掉插头」的方式，但 AI「终止开关」未来可能需要成为一项强制要求。
With Week 1 of the NFL season wrapping up, Adam Greenblatt, CEO of BetMGM, said the season has started enthusiastically and that early activity has been strong. He added that, from the bookmakers’ perspective, Sunday was not a favorable day because the fans came out ahead, leaving bettors happy and
中文摘要 BetMGM 首席执行官亚当·格林布拉特表示，NFL 新赛季首周投注热情高涨、开局活跃；但从博彩公司角度看周日并不有利，因球迷押注获利，令庄家处于亏损一方。
Oil edged higher as traders weighed risks to Middle East supplies, with a critical Saudi Arabian pipeline still offline after an attack.
中文摘要 油价小幅上涨，交易员评估中东供应风险，因一条关键沙特阿拉伯输油管道在遭袭后仍处于停运状态。
The cuts comes into effect in April, but the largest businesses will pay higher rates to fund the changes.
中文摘要 英国威尔士的酒吧、酒店和健身房将从 4 月起获得 30% 的商业地产税减免，大型企业则须缴纳更高税率，为上述减税措施提供资金。
Authentic Brands Group, a licensor of everything from Elvis Presley’s image to the Reebok sportswear brand, could pursue an initial public stock offering as early as the first half of next year.
中文摘要 拥有猫王形象、锐步等品牌授权业务的 Authentic Brands Group 正布局创作者经济，并可能最早于明年上半年启动首次公开募股。
Teva Pharmaceuticals President & CEO Richard Francis, discusses why Teva chose to make the change now and what will be different for investors as the shares move from ADRs to direct NYSE trading. He speaks with Romaine Bostick & Isabelle Lee on "The Close." (Source: Bloomberg)
中文摘要 梯瓦制药总裁兼首席执行官理查德·弗朗西斯表示，公司股票将从美国存托凭证转为在纽约证券交易所直接交易，并说明此时调整的原因及对投资者的影响。
Canada Prime Minister Mark Carney is gathering some of the world’s top investors Monday night and Tuesday as he seeks to attract more investment into the country while reducing its reliance on the US and diversifying its markets.
中文摘要 加拿大总理马克·卡尼在加拿大多伦多举行峰会，会晤全球顶级投资者，目标是吸引更多投资、降低对美国的依赖并实现市场多元化。
The Lincoln Center in NYC has welcomed 600,000 new audience members in the years since the COVID-19 pandemic, offering free of low-cost performances to attract new audiences. Chief Artistic Officer Shanta Thake shares what is on schedule for the Lincoln Center's fall programming, and why she thinks
中文摘要 纽约林肯中心启动秋季演出季。自新冠疫情以来该中心已迎来 60 万名新观众，并通过免费或低价演出吸引新受众；首席艺术官 Shanta Thake 介绍秋季节目安排。
A new Bloomberg Economics report estimates that advanced economies have 27% of employment meaningfully exposed to the first-order effects of AI. Chief Economist Tom Orlik breaks down the new study, which breaks down which jobs and tasks AI can takeover, and how they got to the number of 380 million
中文摘要 彭博经济研究新报告估算，发达经济体约 27% 的就业岗位显著暴露于 AI 的直接影响之下；首席经济学家 Tom Orlik 解读受影响的具体岗位与测算方法。
Apollo Global Management Inc.-owned retailer Michaels Cos. used a windfall from tariff refunds to help cut its leverage, according to people familiar with the matter.
中文摘要 据知情人士称，阿波罗全球管理公司旗下的零售商 Michaels 动用 1.01 亿美元的关税退款，用于降低自身杠杆率。
Sam Huszczo, founder & CIO of SGH Wealth Management, joins Scarlet Fu and Eric Balchunas on "Bloomberg ETF IQ." They discuss how investors might be getting a bit overconfident amid this bull market. (Source: Bloomberg)
中文摘要 SGH 财富管理创始人兼首席投资官 Sam Huszczo 在彭博 ETF IQ 节目中表示，在本轮牛市中投资者可能正变得过于自信。
Potential Federal Reserve rate hikes starting this week could spur a correction in the S&P 500 as reduced corporate margins hurt profit outlooks and markets brace for a tightening cycle, according to Macro Risk Advisors LLC.
中文摘要 Macro Risk Advisors 策略师指出，若美联储自本周起启动加息，企业利润率收窄将拖累盈利前景，标普 500 指数可能出现 10% 的回调。
11 回复 · 程序员 节点
9 回复 · 程序员 节点
4 回复 · 程序员 节点
11 回复 · 程序员 节点
39 回复 · 程序员 节点
16 回复 · 程序员 节点
4 回复 · Apple 节点
6 回复 · Apple 节点
19 回复 · Apple 节点
20 回复 · Apple 节点
该源今日无内容。