cordiverse/cordis
TypeScript · ★ 4,824 · 🍴 262 · 📈 720 stars today
Meta-Framework of Spatiotemporal Composability
中文介绍 面向时空组合问题的元框架，提供空间与时间维度的统一建模和编排能力。通过组合位置、时序、轨迹等计算单元，构建可复用的协同系统，适用于地理信息、物流调度、实时仿真等场景。核心关注点是如何让时空对象灵活组合与演化。
2026-08-17
← 历史归档basecamp/omarchy
Shell · ★ 25,438 · 🍴 2,596 · 📈 270 stars today
Beautiful, Modern & Opinionated Linux
中文介绍 由Basecamp推出的Linux系统，主打美观、现代与明确的设计主张。在系统外观、交互和默认配置上做了深度定制，开箱即用，适合追求高颜值终端体验的开发者。基于Linux生态，但整体风格与常见发行版差异明显，体现对桌面体验的执着。
unslothai/unsloth
Python · ★ 72,639 · 🍴 6,551 · 📈 572 stars today
Local UI to run and train LLMs and diffusion models, including Qwen3.8, Kimi K3, MiniMax-H3, Gemma 4, DeepSeek-V4, FLUX and more.
中文介绍 提供本地UI来运行和训练大语言模型与扩散模型，覆盖Qwen、Kimi、Gemma、DeepSeek、FLUX等多种架构。无需云端API即可在自有硬件上完成推理和微调，保护数据隐私，降低使用成本。适合研究人员、开发者和个人用户在离线或私有环境中实验AI模型。
OpenCut-app/OpenCut
TypeScript · ★ 83,941 · 🍴 8,290 · 📈 150 stars today
The open-source CapCut alternative
中文介绍 开源替代CapCut的视频编辑工具，目标是为用户提供免费、可控的剪辑方案。支持主流编辑功能和导出，并允许开发者自由修改和扩展。面向不想用商业软件或需要私有化部署的视频创作者，可在本地完成剪辑流程，避免平台限制和付费墙。
public-apis/public-apis
Python · ★ 461,868 · 🍴 51,018 · 📈 1,588 stars today
A collective list of free APIs
中文介绍 收集并分类整理大量免费公开API的清单项目，覆盖天气、金融、音乐、数据等多个领域。开发者可通过该列表快速找到所需接口，省去逐一搜索的麻烦。按类别和认证方式组织，适合做原型验证、学习练习或集成第三方数据时使用。
ToolJet/ToolJet
JavaScript · ★ 40,079 · 🍴 5,331 · 📈 452 stars today
ToolJet is the open-source foundation of ToolJet AI - the enterprise app generation platform for building internal tools, dashboard, business applications, workflows and AI agents 🚀
中文介绍 开源的企业级应用生成平台，用于快速构建内部工具、仪表盘、业务应用、工作流和AI代理。通过可视化拖拽连接数据源，无需从零开发，降低企业应用搭建门槛。它是ToolJet AI的基础，适合开发团队和业务部门在几天内交付管理后台或自动化流程。
cactus-compute/needle
Python · ★ 6,628 · 🍴 435 · 📈 443 stars today
14MB foundation model for tiny devices; phones, wearables, smart home, and robots.
中文介绍 体积仅14MB的基础模型，专为手机、可穿戴设备、智能家居和机器人等微型设备设计。在极小的存储占用下提供核心AI能力，让端侧设备无需联网也能运行智能推理。面向嵌入式开发者，适用于资源受限但需要本地智能的场景，如语音助手、设备状态识别等。
LLMRouter: Unified Infrastructure for Developing, Evaluating, and Deploying LLM Routers
👍 103
No single large language model (LLM) is optimal across all queries and budget constraints, making model routing essential for cost-effective deployment. Existing routers adopt diverse formulations and implementations, making fair comparison and extension difficult. We present a unified formulation o
中文介绍 LLMRouter 提出统一的 LLM 路由基础设施，覆盖路由器开发、评估与部署全流程。针对现有路由方法公式化与实现各异、难以公平比较和扩展的问题，给出统一公式化表示，降低模型路由部署成本，提升性价比。
DreamX-Phi 1.0: Action-Conditioned Video World Model for Robotic Manipulation
👍 92
We present DreamX-Phi 1.0, an action-conditioned video world model for robotic manipulation that, given an observed frame, a language instruction, and a prescribed action sequence comprising end-effector poses and gripper states, predicts the resulting future observations. Yet realism alone does not
中文介绍 DreamX-Phi 1.0 是面向机器人操作的动作条件视频世界模型，输入当前观测帧、语言指令和包含末端执行器位姿与夹爪状态的动作序列，预测未来观测结果。旨在超越单纯真实感，支持操作规划。
DarwinX: Evolving Agent Harnesses Through Natural Selection
👍 84
An LLM agent's capability depends not only on model weights but on its harness: prompts, tools, skills, and control flow. Self-improvement loops already edit harnesses, yet single-lineage search is path-dependent and local wins often regress other tasks. We introduce DarwinX, which treats self-evolu
中文介绍 DarwinX 将 LLM agent 的自演化视为自然选择过程，对包含提示、工具、技能与控制流的 harness 进行群体演化。克服单一线程搜索的路径依赖和局部优化导致其他任务性能回退的问题。
Mechanist: AI as a Scientific Instrument for Discovering the Mechanisms of Intelligence
👍 84
AI models have achieved remarkable success across diverse domains, yet the mechanisms underlying their capabilities and the risks they may pose remain poorly understood. As AI development becomes faster and increasingly automated, mechanistic exploration remains largely manual, widening the gap betw
中文介绍 Mechanist 将 AI 本身作为科学仪器，用于自动发现智能模型能力与风险的底层机制。针对机制探索高度依赖人工、与 AI 自动化发展速度差距拉大的问题，提供自动化研究工具。
SkillZip: Contract-Preserving Graph Compression for Scalable Agent Skill Libraries
👍 74
Large Language Models (LLMs) increasingly act as agents whose procedural knowledge is stored in reusable skill packages and loaded at inference time. As skill libraries grow, a central challenge is to expose the smallest sufficient executable context under a limited context budget. Existing systems
中文介绍 SkillZip 提出保持契约的图压缩方法，用于大语言模型 agent 的技能库扩展。在有限上下文预算下，从不断增长的技能包中压缩出最小充分的可执行上下文，降低推理时加载成本。
Intern-S2-Preview: Scientific Agentic Foundation Model
👍 54
Scientific discovery increasingly requires AI systems that can reason over scientific evidence of heterogeneous modalities, interact with scientific tools and environments, and sustain progress across long task horizons. We present Intern-S2-Preview, a series of scientific agentic foundation models
中文介绍 Intern-S2-Preview 是一系列面向科学发现的智能体基础模型，能够对异构多模态科学证据进行推理，与科学工具和环境交互，并在长任务跨度中保持持续进展，加速科学发现流程。
How Can Rhetoric Reward-Hack AI Reviewers? Dissecting Rhetorical Sensitivity in AI-Based Peer Review
👍 47
As large language models increasingly participate in scientific evaluation, we investigate a potential form of reward hacking: how rhetorical choices shape AI-review judgments when reported scientific content is preserved and how these effects vary across evaluation conditions. We construct a contro
中文介绍 本研究探讨 AI 同行评审中的奖励黑客风险：在保持科学内容不变的情况下，考察修辞表达如何影响大语言模型评审判断，并分析不同评审条件下的差异，揭示 AI 评审对修辞的敏感性。
AutoDesign: Meta-Harness Optimization for Long-Horizon Agentic Design
👍 43
Transforming multimodal sources into condensed and structured media outputs can be fundamentally conceptualized as a long-horizon agentic process centered on a model-harness system. While an ideal harness system should align with human design priors and accumulate reusable experience through empiric
中文介绍 AutoDesign 面向长周期智能体设计提出元 harness 优化方法。将多模态源转化为结构化媒体输出视为以模型-harness 系统为中心的长期过程，自动对齐人类设计先验并通过经验积累实现可复用优化。
PlayWorld: Benchmarking World Models with Agent Players over Long-Horizon Objectives
👍 42
Video world models simulate future states conditioned on current observations and user actions. Recent systems have demonstrated impressive video consistency and action controllability over long sequences. However, fairly comparing these interactive models remains challenging. In practice, a human p
中文介绍 PlayWorld 提出用智能体玩家评测视频世界模型的基准，在长周期目标下考察模型对未来状态的模拟能力。针对现有系统视频一致性与动作可控性难以公平比较的问题，提供标准化评测框架。
Spatial Memory Agent: Experience-Grounded Procedure Memory for Spatial Intelligence
👍 40
Spatial intelligence is becoming a foundation for embodied agents, robotic planning, and multimodal assistants. To improve the spatial reasoning ability of VLM agents, existing work has mainly followed two lines. One line uses post-training methods, such as supervised fine-tuning and reinforcement l
中文介绍 Spatial Memory Agent 提出基于经验的程序记忆机制，增强 VLM 智能体的空间智能。区别于监督微调和强化学习等后训练方法，通过存储和检索空间任务经验，支持具身智能与机器人规划中的空间推理。
Massive Activations in Hybrid Linear Attention Large Language Models: Pre-Attention Spikes and Inter-Spike Plateaus
👍 28
We present the first systematic study of Massive activations (MAs) in layer-interleaved HLA LLMs and uncover two architecture-aligned morphologies: MAs consistently spike immediately before full attention layers, forming pre-attention spikes (PAS), and can persist through intervening linear attentio
中文介绍 首次系统研究混合线性注意力（HLA）大语言模型中的 Massive activations，揭示两种与架构对齐的形态：全注意力层前出现的 pre-attention spikes，以及在线性注意力层间持续的 inter-spike plateaus，为理解激活模式提供新视角。
Full-bandwidth transformer
👍 20
Autoregressive transformers compute along two axes: horizontally across generated tokens, and vertically through model depth. Dense attention gives each token broad horizontal access to the past, but the vertical feedback channel between decoding steps remains narrow: only the sampled token returns
中文介绍 Full-bandwidth Transformer 针对自回归 Transformer 解码步间垂直反馈通道过窄的问题，提出加宽带模型深度方向信息流动的方法。除水平注意力外，让更多信息从采样 token 回传至后续解码过程，提升生成质量。
An AI4AI Framework for Visual Token Pruning
👍 15
Visual-token pruning can substantially reduce the inference cost of multimodal large language models (MLLMs), yet existing methods largely rely on fixed, handcrafted heuristics and costly expert trial and error. As pruning objectives, budgets, and model architectures diversify, manually navigating t
中文介绍 AI4AI 框架利用 AI 自动设计视觉 token 剪枝策略，降低多模态大语言模型推理成本。针对现有剪枝方法依赖固定启发式和专家试错的问题，在多样化剪枝目标、预算和模型架构下自动导航优化。
Thought-Level Beam Search for Reasoning
👍 14
Test-time compute scaling is a primary driver of performance in large reasoning models (LRMs), but extreme inefficiency bounds current approaches, shifting the critical question from how much compute to spend, to where to allocate it. We formalize test-time reasoning as a constrained compute allocat
中文介绍 Thought-Level Beam Search 将测试时推理形式化为受限计算分配问题，在思想级别进行束搜索，以确定计算资源的最佳分配位置。针对大型推理模型计算效率极低的问题，提升测试时计算缩放的有效性。
AVA-Encoder: Towards Agent-Native Video Representation Learning
👍 14
Creative agents still lack an effective way to learn from high-quality human films, limiting their ability to produce cinematic-grade videos. A key challenge is the absence of a structured video representation that is both faithful to film content and directly usable for agentic reasoning and manipu
中文介绍 AVA-Encoder 提出面向智能体的视频表示学习方法，解决创意智能体难以从高质量人类影片中学习的问题。学习到的结构化表示既忠实于影片内容，又能直接用于智能体推理和视频编辑操作，助力生成电影级视频。
OmniScientist: An Omni-Modal Omni-Discipline AI Scientist
👍 14
Recent advances in foundation models have enabled AI scientists to automate increasingly complete research workflows, from hypothesis generation and code execution to manuscript preparation. Yet workflow coverage alone does not provide access to the full evidence on which scientific discovery depend
中文介绍 OmniScientist 提出全模态、全学科的 AI 科学家框架，在自动化研究流程（假设生成、代码执行、稿件撰写）基础上，进一步提供完整科学证据访问，使 AI 科学家能基于全部相关证据进行发现。
Maglev: Sliding Recurrent Memory
👍 13
We introduce , a recurrent Transformer architecture with fixed-size memory that generalizes sliding-window attention while remaining parallelizable during training. consists of two coupled models: a prefiller Q, which leverages full attentionIn practice, we use interleaved full and sliding-window a
中文介绍 Maglev 提出带固定大小记忆的循环 Transformer 架构，通过滑动循环记忆推广滑动窗口注意力，同时保持训练可并行化。它由两个耦合模型构成，实际使用中交替采用全注意力和滑动窗口注意力。
SKILLER: Language-Level Reinforcement Learning for Reusable Skill Extraction in Small Language Models
👍 13
Agent skills represent a standardized format for packaging procedural knowledge and domain expertise, serving within agent harness systems as an essential mechanism to continually constrain a language model's behavior space for repeatable, high-quality task execution. However, because strong closed-
中文介绍 SKILLER 通过语言级强化学习，在小语言模型中提取可复用技能，以标准化格式封装程序性知识与领域专长。针对强闭源模型技能不可得的问题，在 agent harness 系统中持续约束模型行为空间，实现高质量任务执行。
Hybrid-Policy Self-Editing for Composable Unstructured Knowledge Editing
👍 11
Large language models (LLMs) achieve remarkable performance across natural language tasks, yet they are trained on static corpora and their knowledge quickly becomes outdated in a fast-changing world. This motivates knowledge editing (KE), which updates specific knowledge in an LLM without changing
中文介绍 Hybrid-Policy Self-Editing 提出混合策略的自编辑方法，用于大语言模型的可组合非结构化知识编辑。在不改变模型整体行为的前提下更新过时知识，支持多个知识编辑的组合，应对静态语料带来的知识时效性问题。
Knowing When to Quit: Diagnosing and Training LLMs to Abort Futile Reasoning
👍 11
Large language models generate computationally expensive yet semantically void reasoning on beyond-capability tasks, creating risks where plausible-sounding but incorrect derivations mislead users. We characterize this futile reasoning phenomenon through systematic analysis, revealing universal capa
中文介绍 本文系统诊断大语言模型在超出能力任务上产生的“无益推理”现象，即计算开销大但语义空洞的推导。通过分析其普遍特征，训练模型在无益推理时主动停止，避免貌似合理但错误的推导误导用户。
Cursor + SpaceXAI: the fastest iterating team wins
@sarahdingwang · 18.0K 粉丝 · 1.2M 阅 · 514 赞 · 73 转
By @BornsteinMatt, @martin_casado, and myself. AI coding could be the biggest addressable market of our lives. And, unless you think we’ve fully addressed it (which no one thinks; we are more like 1%
中文介绍 联合撰文谈 AI 编程：称这是此生最大的潜在市场，目前开发渗透率仅约 1%，远未饱和。强调「Cursor + SpaceXAI」组合下，最快迭代的团队才能胜出，呼吁抓住这一波范式转移。
Preparing GLM-5.3 for Open Release: A Responsible Path to Cyber Defense
@Zai_org · 136.2K 粉丝 · 42.7K 阅 · 583 赞 · 50 转
When GLM-5.2 helped Hugging Face investigate an incident in which an AI autonomously bypassed its own safeguards, it highlighted a broader shift. AI is becoming part of both cyber offense and cyber
中文介绍 智谱宣布为开源 GLM-5.3 做准备，主打网络防御责任路径。此前 GLM-5.2 曾协助 Hugging Face 调查 AI 自主绕过自身安全机制的事件，凸显 AI 已同时进入网络攻防两端，开源需兼顾安全。
Grok Bot for GTM
@kristaletz · 4.2K 粉丝 · 41.6K 阅 · 513 赞 · 24 转
This is how I use Grok Bot for Enterprise GTM at SpaceXAI. When we were initially dogfooding this product internally, Grok Bot rapidly gained adoption throughout the company, and I created an internal
中文介绍 分享 SpaceXAI 内部使用 Grok Bot 做企业 GTM（市场进入）的实践。产品内测时迅速被全公司采用，作者为此搭建了内部流程，展示 AI 助手如何驱动企业级销售与增长。
The AI Engineering Skills Map
@AndrewYNg · 1.8M 粉丝 · 41.2K 阅 · 716 赞 · 89 转
I am delighted to present The AI Engineering Skills Map. AI allows us to build software very differently today than in 2022, and everyone with the skills to take advantage of this shift has numerous
中文介绍 吴恩达发布《AI 工程技能图谱》。指出相比 2022 年，AI 让软件构建方式彻底改变，掌握这些技能的人将拥有大量机会，为开发者规划学习路径提供系统框架。
What's new in Gemini 3.7 Flash
@GoogleAIStudio · 192.9K 粉丝 · 31.5K 阅 · 535 赞 · 34 转
Gemini 3.7 Flash (gemini-3.7-flash) is generally available (GA) and ready for production use. It is our most intelligent workhorse model yet for software engineering, and agentic workflows. This guide
中文介绍 Gemini 3.7 Flash 正式发布（GA），可投入生产。官方称这是面向软件工程与 Agentic 工作流最智能的「主力模型」，并同步更新使用指南，介绍新特性和应用场景。
The Holy Trinity - pi, herdr and opencode go
@TheGlobalMinima · 4.0K 粉丝 · 31.4K 阅 · 527 赞 · 17 转
tldr; my current AI coding setup is Pi, Opencode Go and Herdr This year, we are seeing a boom in harness engineering. What initially started as simple scaffolding around an LLM is now one of the most
中文介绍 作者分享当前 AI 编程组合：Pi、Opencode Go 与 Herdr 三件套。指出今年迎来「harness 工程」爆发，从最初简单封装 LLM 变成核心环节，强调工具链编排比模型本身更重要。
Qwen3.8-27B Practical Guide: Control Reasoning Depth and Extend Context to 1M Tokens
@QwenDevs · 8.0K 粉丝 · 25.6K 阅 · 559 赞 · 51 转
Qwen3.8-27B is here — a compact, deployment-friendly dense model built on the Qwen3.5 architecture, with native vision-language understanding and flexible thinking control. Two developer-facing
中文介绍 Qwen3.8-27B 发布，基于 Qwen3.5 架构的紧凑稠密模型，支持原生视觉语言理解与灵活思考控制。官方给出实战指南：可调节推理深度，并将上下文扩展至 100 万 token，方便部署。
ORACLE: Confirmed AI Bots Actively Trading on Polymarket — Gains Reaching 200% Within 24 Hours
@OracleFoundtion · 19.4K 粉丝 · 18.8K 阅 · 3.1K 赞 · 615 转
In 2026, autonomous AI agents have become one of the most effective strategies on prediction markets. Over 30% of all activity on Polymarket now comes from algorithmic and AI-powered wallets. We
中文介绍 ORACLE 确认 AI 机器人正活跃交易 Polymarket，24 小时内收益可达 200%。目前平台超 30% 交易量来自算法与 AI 钱包，自主智能体已成为预测市场最有效策略之一。
ORACLE: Confirmed AI Bots Actively Trading on Polymarket — Gains Reaching 200% Within 24 Hours
@OracleFndAI · 23.9K 粉丝 · 2.3K 阅 · 3.1K 赞 · 594 转
In 2026, autonomous AI agents have become one of the most effective strategies on prediction markets. Over 30% of all activity on Polymarket now comes from algorithmic and AI-powered wallets. We
中文介绍 ORACLE 确认 AI 机器人正在 Polymarket 上积极交易，24 小时内收益可达 200%。平台现有超过 30% 的活跃来自算法与 AI 钱包，自主智能体已成为预测市场的高效策略。
React for Agents: Astro Creator Brings Hooks to his Meta-Harness, Flue
Flue 2 takes its inspiration from React. Creator Fred Schott, of Astro fame, tells Latent Space why he added hooks and why agents are defined by their harnesses.
中文介绍 Flue 2 的设计灵感来自 React。Astro 创建者 Fred Schott 在 Latent Space 访谈中解释了为何为 Flue 引入 hooks 机制，并阐述智能体由其所处的 harness（元框架）定义的观点。
[AINews] Gemini 3.7 Flash brings GDM back to the forefront
Down, but not out!
中文介绍 Latent Space 的 AINews 报道称，谷歌发布 Gemini 3.7 Flash，使 GDM 重新成为关注焦点。此前 GDM 一度表现低迷，但并未退出竞争。
State of Open Models: Summer 2026 Observations
中文介绍 Hugging Face 发布 2026 年夏季开放模型现状观察，梳理开源模型领域的最新进展、性能突破与生态变化，为开发者提供参考。
Gemini 3.7 🤖, GPT-5.6 Sol Ultrafast ⚡, Anthropic $2T IPO 💰
中文介绍 TLDR AI 8 月 14 日简报聚焦三大进展：谷歌发布 Gemini 3.7；OpenAI 推出 GPT-5.6 Sol 超快模式；Anthropic 启动估值达 2 万亿美元的 IPO。
Record, train, and deploy from one place with Strands Agents, LeRobot, and Hugging Face Storage Buckets
中文介绍 Hugging Face 博客介绍如何借助 Strands Agents、LeRobot 与 Hugging Face Storage Buckets，在单一平台上完成机器人数据记录、模型训练与部署的全流程。
Introducing Gemini 3.7 Flash
中文介绍 DeepMind 正式发布 Gemini 3.7 Flash 模型。官方博客介绍了该模型的技术特性、性能表现与适用场景，为开发者提供新的模型选择。
Flock is tightening its rules in response to a growing surveillance backlash
The police-tech giant Flock is announcing today that it will change officers’ access to its nationwide network of license plate readers, in an apparent effort to quell a growing backlash and win back contracts lost amid concerns about mass surveillance and police abuse. Several changes aim directly
中文介绍 警用科技巨头 Flock 宣布变更执法人员对其全国车牌读取器网络的访问权限，意在平息日益高涨的监控反对声浪，并挽回因大规模监控与警方滥用担忧而流失的合同。
The builder’s guide to GPT‑5.6
Learn how startups use GPT-5.6 to build faster, more cost-efficient AI agents with smarter model selection and new Responses API capabilities.
中文介绍 OpenAI 发布 GPT-5.6 开发者指南，介绍初创企业如何利用该模型构建更快、更具成本效益的 AI 智能体，涵盖更智能的模型选择与新的 Responses API 功能。
Previewing Ultrafast mode: GPT-5.6 Sol at up to 14X the speed
Preview Ultrafast, a new OpenAI API service tier that runs GPT-5.6 Sol up to 14× faster. Powered by Cerebras, it delivers up to 750 output tokens per second.
中文介绍 OpenAI 预览 Ultrafast 模式，这是由 Cerebras 支持的全新 API 服务层级，可使 GPT-5.6 Sol 运行速度提升最高 14 倍，输出速率达到每秒 750 个 token。
OpenAI appoints Dali Rajic as Chief Revenue Officer
OpenAI appoints Dali Rajic as Chief Revenue Officer to lead its global revenue organization and help businesses realize the full value of AI.
中文介绍 OpenAI 任命 Dali Rajic 为首席营收官（CRO），领导其全球营收组织，并帮助各类企业充分实现人工智能的价值。
How kids feel about AI, in their own words
When we set out to talk to kids about artificial intelligence, we thought we knew what we’d hear. We expected some to tell us they were using it to cheat a little, the way Millennials and Gen Xers opened up CliffsNotes or programmed formulas into their TI-82s, and others to share inspiring ways they
中文介绍 MIT Tech Review 就人工智能话题采访儿童，原以为会有孩子承认用 AI 作弊，如同千禧一代与 X 世代使用 CliffsNotes 或在 TI-82 上编程公式一样。
not much happened today
**Google** rapidly released **Gemini 3.7 Flash** just three weeks after 3.6 Flash, targeting coding, web development, knowledge work, and agentic workflows with a 50% introductory price cut and improved benchmark scores like **DeepSWE 65.3%** and **Code Arena Elo 1588**. The update quickly integrate
中文介绍 Smol AI News 报道，谷歌在 Gemini 3.6 Flash 发布仅三周后推出 Gemini 3.7 Flash，面向编码、网页开发、知识工作及智能体工作流，首发降价 50%，基准测试提升，DeepSWE 得分 65.3%，Code Arena Elo 1588。
[AINews] SpaceXAI Grok 4.6 and Grok @Bot
AI teammate category just had its most significant new entrant yet
中文介绍 Latent Space 的 AINews 报道称，SpaceXAI 发布 Grok 4.6 和 Grok @Bot，成为 AI 队友（AI teammate）类别迄今最重要的新进入者。
What We Learned by Reproducing 2,200 papers from ICML
中文介绍 Hugging Face 博客分享其复现 ICML 上 2200 篇论文的经验与心得，探讨了学术研究可复现性方面的挑战与启示。
Claude Chrome Cowork 🌐, Grok 4.6 🚀, DeepSeek v4-Pro-0813 🧠
中文介绍 TLDR AI 8 月 13 日简报聚焦：Anthropic 的 Claude Chrome Cowork 功能、Grok 4.6 的发布，以及 DeepSeek v4-Pro-0813 模型。
美股大盘技术面最强：SPY 和 QQQ 均高于 20/50/200 日均线，呈多头排列，SPY 距 52 周高点仅 -0.39%，RSI 65.7；但个股分化，MSFT RSI 70.9 超买，AAPL MACD 为负，GOOGL 触发死叉。加密市场情绪偏恐慌，恐慌贪婪指数 31，BTC-USD 空头排列、RSI 43、MACD 绿柱扩大，ETH-USD 围绕 20 日线震荡，SOL 亦空头排列，加密总市值 2.25T，BTC 主导率 56.1%。中概股短线承压：JD 5 日跌 11.86%、RSI 37.8，0700.HK 死叉且空头排列，PDD 跌破 20 日线；BABA 相对抗跌但低于 200 日线。商品与外汇分化：黄金期货 1 日涨 1.64%、RSI 68.5 逼近超买，原油站上所有均线且 MACD 金叉；美元兑人民币 RSI 27.7 超卖，美元指数 99.51 偏弱。整体技术面美股偏强，加密与中概偏弱，商品中性偏强。
今日关注
SPY
标普 500 ETF (SPY)
偏上行
当前价 776.34 高于 SMA20 756.2、SMA50 748.93、SMA200 705.46，均线多头排列；距 52 周高点仅 -0.39%，RSI 65.7 处于正常偏强区，MACD 红柱（8.51>6.26）延续，5 日累计 +0.4%，短中期趋势向上。
NVDA
英伟达 (NVDA)
偏上行
当前价 225.16 站上 SMA20 210.39、SMA50 206.52、SMA200 194.92，多头排列；RSI 63，MACD 值 5.28 高于信号线 3.23，红柱为正，5 日 +0.54%，距 52 周高点 -4.81%，技术动能偏强。
BTC-USD
比特币 (BTC-USD)
偏下行
当前价 63003.44 低于 SMA20 63792.03、SMA50 63572.2、SMA200 69240.28，空头排列；RSI 43 低于中性 50，MACD -234.57 低于信号 -81.25，绿柱扩大，5 日 -0.86%，距 52 周高 -50.08%，叠加恐慌贪婪指数 31，技术面偏弱。
JD
京东 (JD)
偏下行
当前价 29.06 低于 SMA20 31.57、SMA50 29.21、SMA200 29.38，呈空头排列；RSI 37.8 接近超卖，MACD 0.447 低于信号 0.910，死叉后绿柱运行，5 日大跌 11.86%，距 52 周高点 -21.16%，短期下行动能显著。
ETH-USD
以太坊 (ETH-USD)
中性
当前价 1888.38 与 SMA20 1888.48 基本持平，高于 SMA50 1836.51，低于 SMA200 2012.77；RSI 52.5 处于中性区，MACD 8.63 低于信号 13.54，红柱转绿但仍在零轴上方，5 日 +0.37%，方向不明确，技术状态中性。
全部资产
^VIX
VIX 恐慌指数
$14.25
-2.60%
- 5 日
- -4.36%
- 距 52w 高
- -59.6%
- RSI(14)
- 39.7
- 趋势
- 空头
- SMA 20 / 50 / 200
- 16.71 / 17.14 / 18.52
- MACD / 信号
- -0.767 / -0.467
空头排列
^TNX
10Y 美债收益率 (%)
$4.70
+1.19%
- 5 日
- +0.77%
- 距 52w 高
- -1.1%
- RSI(14)
- 49.7
- 趋势
- 中性
- SMA 20 / 50 / 200
- — / — / —
- MACD / 信号
- — / —
接近 52 周高
DX-Y.NYB
美元指数 DXY
$99.51
-0.16%
- 5 日
- -0.30%
- 距 52w 高
- -2.2%
- RSI(14)
- 37.0
- 趋势
- 中性
- SMA 20 / 50 / 200
- 100.33 / 100.56 / 99.19
- MACD / 信号
- -0.281 / -0.209
接近 52 周高
SPY
S&P 500 ETF
$776.34
-0.20%
- 5 日
- +0.40%
- 距 52w 高
- -0.4%
- RSI(14)
- 65.7
- 趋势
- 多头
- SMA 20 / 50 / 200
- 756.20 / 748.93 / 705.46
- MACD / 信号
- 8.515 / 6.257
接近 52 周高多头排列
QQQ
Nasdaq 100 ETF
$731.07
-0.14%
- 5 日
- +1.11%
- 距 52w 高
- -2.3%
- RSI(14)
- 59.6
- 趋势
- 多头
- SMA 20 / 50 / 200
- 704.13 / 712.95 / 650.62
- MACD / 信号
- 5.524 / 1.144
接近 52 周高多头排列
AAPL
Apple
$305.93
+0.22%
- 5 日
- -2.36%
- 距 52w 高
- -11.2%
- RSI(14)
- 43.7
- 趋势
- 中性
- SMA 20 / 50 / 200
- 318.43 / 309.19 / 280.48
- MACD / 信号
- -2.461 / 0.187
MSFT
Microsoft
$495.40
-0.30%
- 5 日
- -0.92%
- 距 52w 高
- -10.5%
- RSI(14)
- 70.9
- 趋势
- 中性
- SMA 20 / 50 / 200
- 450.24 / 412.78 / 432.48
- MACD / 信号
- 28.582 / 25.060
RSI 超买
NVDA
Nvidia
$225.16
-0.06%
- 5 日
- +0.54%
- 距 52w 高
- -4.8%
- RSI(14)
- 63.0
- 趋势
- 多头
- SMA 20 / 50 / 200
- 210.39 / 206.52 / 194.92
- MACD / 信号
- 5.277 / 3.226
多头排列
GOOGL
Alphabet
$345.90
-0.13%
- 5 日
- -2.37%
- 距 52w 高
- -15.3%
- RSI(14)
- 47.0
- 趋势
- 中性
- SMA 20 / 50 / 200
- 346.41 / 353.89 / 331.38
- MACD / 信号
- -1.137 / -0.970
MACD 死叉 (今天)
TSLA
Tesla
$342.27
+0.68%
- 5 日
- +4.17%
- 距 52w 高
- -31.4%
- RSI(14)
- 48.8
- 趋势
- 空头
- SMA 20 / 50 / 200
- 329.14 / 371.09 / 406.08
- MACD / 信号
- -11.370 / -16.326
空头排列
META
Meta
$589.85
-0.86%
- 5 日
- -0.38%
- 距 52w 高
- -25.4%
- RSI(14)
- 47.6
- 趋势
- 空头
- SMA 20 / 50 / 200
- 594.67 / 597.13 / 627.63
- MACD / 信号
- -5.392 / -5.604
MACD 金叉 (今天)空头排列
BTC-USD
Bitcoin
$63,003.44
-0.03%
- 5 日
- -0.86%
- 距 52w 高
- -50.1%
- RSI(14)
- 43.0
- 趋势
- 空头
- SMA 20 / 50 / 200
- 63,792.03 / 63,572.20 / 69,240.28
- MACD / 信号
- -234.573 / -81.250
空头排列
ETH-USD
Ethereum
$1,888.38
+0.40%
- 5 日
- +0.37%
- 距 52w 高
- -61.9%
- RSI(14)
- 52.5
- 趋势
- 中性
- SMA 20 / 50 / 200
- 1,888.48 / 1,836.51 / 2,012.77
- MACD / 信号
- 8.629 / 13.541
SOL-USD
Solana
$75.04
-0.31%
- 5 日
- -1.52%
- 距 52w 高
- -70.4%
- RSI(14)
- 49.7
- 趋势
- 空头
- SMA 20 / 50 / 200
- 74.45 / 76.06 / 81.78
- MACD / 信号
- 0.089 / -0.069
空头排列
BABA
阿里巴巴 (BABA)
$123.81
+1.35%
- 5 日
- -3.58%
- 距 52w 高
- -35.7%
- RSI(14)
- 55.0
- 趋势
- 中性
- SMA 20 / 50 / 200
- 121.81 / 113.76 / 138.95
- MACD / 信号
- 3.520 / 3.681
MACD 死叉 (今天)
PDD
拼多多 (PDD)
$84.79
+0.74%
- 5 日
- -7.60%
- 距 52w 高
- -39.2%
- RSI(14)
- 43.9
- 趋势
- 中性
- SMA 20 / 50 / 200
- 87.58 / 84.03 / 101.76
- MACD / 信号
- 0.904 / 1.434
MACD 死叉 (1 天前)
JD
京东 (JD)
$29.06
-0.82%
- 5 日
- -11.86%
- 距 52w 高
- -21.2%
- RSI(14)
- 37.8
- 趋势
- 空头
- SMA 20 / 50 / 200
- 31.57 / 29.21 / 29.38
- MACD / 信号
- 0.447 / 0.910
MACD 死叉 (2 天前)空头排列
0700.HK
腾讯控股 (0700.HK)
HK$444.00
+0.91%
- 5 日
- -7.77%
- 距 52w 高
- -35.0%
- RSI(14)
- 42.3
- 趋势
- 空头
- SMA 20 / 50 / 200
- 463.25 / 455.02 / 530.39
- MACD / 信号
- -1.189 / 3.450
MACD 死叉 (3 天前)空头排列
GC=F
黄金期货
$4,452.30
+1.64%
- 5 日
- +2.07%
- 距 52w 高
- -20.3%
- RSI(14)
- 68.5
- 趋势
- 中性
- SMA 20 / 50 / 200
- 4,196.72 / 4,157.54 / 4,488.70
- MACD / 信号
- 77.269 / 41.104
CL=F
WTI 原油期货
$82.64
+0.29%
- 5 日
- +0.62%
- 距 52w 高
- -30.8%
- RSI(14)
- 52.9
- 趋势
- 多头
- SMA 20 / 50 / 200
- 82.48 / 79.34 / 76.95
- MACD / 信号
- 0.402 / 0.240
MACD 金叉 (3 天前)多头排列
USDCNY=X
美元 / 人民币
¥6.74
-0.04%
- 5 日
- -0.12%
- 距 52w 高
- -6.3%
- RSI(14)
- 27.7
- 趋势
- 空头
- SMA 20 / 50 / 200
- 6.76 / 6.77 / 6.89
- MACD / 信号
- -0.009 / -0.008
RSI 超卖接近 52 周低空头排列
过去走势不代表未来表现，任何技术指标信号均可能随市场变化而失效。本报告基于公开行情数据，仅供技术指标解读参考，不构成投资建议或价格预测，读者需独立判断并自担风险。
Trump orders cuts to South Korea joint drills despite warnings of North’s growing battlefield experience
President says move is based on ‘very good relationship with Kim Jong-un’ but US military officials have said Ukraine war is boosting regime’s combat capabilities Donald Trump has said he is scaling back joint US military drills with South Korea aimed at deterring North Korea, despite military offic
中文摘要 特朗普下令削减与韩国的联合军事演习，称此举基于他与金正恩的“非常良好关系”。但美国军方官员警告，乌克兰战争正在提升朝鲜的作战能力。
Ferrari's first ever electric car sold for record $40m at auction
The Luce's launch drew backlash from some commentators who said it broke from the brand's signature dramatic styling.
中文摘要 法拉利首款电动汽车“Luce”在拍卖会上以创纪录的4000万美元成交。该车发布后遭到一些评论人士批评，认为其偏离了品牌标志性的戏剧化设计风格。
Australia politics live: Hume denies Chalmers’ claim Coalition would ‘end super as we know it’; Victorian teachers accept pay offer
Vote to accept state government’s latest pay offer ends year-long negotiations. Follow today’s news live Get our breaking news email, free app or daily news podcast Hume says housing the ‘best indicator of economic security’ in retirement over super The deputy Liberal leader, Jane Hume says the Coal
中文摘要 澳大利亚自由党副领袖休姆否认国库部长查默斯关于联盟党将“终结现行退休金制度”的说法，称住房是退休经济保障的最佳指标。维州教师投票接受州政府最新薪资提议，结束长达一年的谈判。
Israeli settlers pitch tents for new illegal outpost in occupied West Bank
UN Special Rapporteur Albanese warns that Israel is 'using settlers to advance ethnic cleansing' in the West Bank.
中文摘要 以色列定居者在被占领的约旦河西岸搭建帐篷，建立新的非法前哨。联合国特别报告员阿尔巴尼斯警告，以色列正“利用定居者推进种族清洗”。
Can the AU make its current Sudan mission count?
The African Union is seeking progress in Sudan, but must overcome deep battlefield and diplomatic divisions.
中文摘要 非洲联盟正寻求在苏丹问题上取得进展，但必须克服战场和外交层面的深刻分歧。非盟当前苏丹任务能否奏效，引发外界关注。
Canadian Petition to Expel U.S. Ambassador Pete Hoekstra Takes Off
Mr. Hoekstra has been President Trump’s faithful messenger in Canada amid a historic low in relations. More than 200,000 Canadians want him gone.
中文摘要 超过20万加拿大人签署请愿书，要求驱逐美国驻加拿大大使皮特·霍克斯特拉。在美加关系处于历史低点之际，霍克斯特拉一直是特朗普在加拿大的忠实传声筒。
About 100 homes swept away after Hurricane Lala lashes Hawaii
The state's governor says there "amazingly" does not seem to have been any loss of life besides a car crash.
中文摘要 飓风“拉拉”袭击夏威夷，约100所房屋被冲毁。该州州长表示，除一起车祸外，“惊人地”似乎没有造成任何生命损失。
Trump says US to ‘substantially reduce’ military drills with South Korea
US president says he's 'not happy' about the drills, citing his 'very good relationship' with North Korea's Kim Jong Un.
中文摘要 美国总统特朗普表示，美国将“大幅减少”与韩国的军事演习。他说自己对军演“不满意”，并提到他与朝鲜领导人金正恩的“非常良好关系”。
Ebola outbreak in Democratic Republic of Congo now deadliest in country’s history
At least 2,325 people have been killed, according to government data, surpassing the death toll from 2018-20 outbreak The Ebola outbreak in the Democratic Republic of Congo is now the deadliest in the country’s history with at least 2,325 people killed, according to government data, surpassing the t
中文摘要 刚果（金）埃博拉疫情成为该国历史上最严重的一次。政府数据显示，至少2325人死亡，超过2018-2020年疫情的死亡人数。
Trump says US to scale back military drills with South Korea after it stayed out of Iran war
Trump cites his "very good relationship" with North Korea's leader in deciding to downsize the exercises.
中文摘要 特朗普表示，由于韩国未参与伊朗战争，美国将缩减与韩国的军事演习。他还援引与朝鲜领导人金正恩的“非常良好关系”作为决定缩小演习规模的原因。
Israel releases 35 Palestinian prisoners back into Gaza
Emotional reunions greeted 35 Palestinians prisoners released by Israel into Gaza.
中文摘要 以色列释放35名巴勒斯坦囚犯返回加沙地带，囚犯们与等候的家人激动团聚，现场欢呼与泪水交织。此次释放受到外界关注。
Iran war live: Tehran-US MoU set to expire; Hamas’s al-Khayya meets Kushner
Hamas calls on Board of Peace to 'compel' Israel to accept the Gaza peace plan after meeting Jared Kushner in Cairo.
中文摘要 伊朗战争动态：德黑兰与美国的谅解备忘录即将到期；哈马斯高级成员哈亚在开罗会见库什纳，呼吁和平委员会“强制”以色列接受加沙和平计划。
UN’s Albanese: Israel outsourcing coercion to settlers in West Bank
UN Special Rapporteur Francesca Albanese says Israel bears responsibility and is outsourcing coercion to settlers.
中文摘要 联合国特别报告员弗兰切斯卡·阿尔巴尼斯表示，以色列在约旦河西岸将胁迫行为外包给定居者，但以方仍应承担责任。
Israeli minister calls for killing ’30 to 40′ Palestinians in Gaza nightly
Far-right National Security Minister Ben-Gvir sits on Israel's security cabinet, which helps shape war and policy.
中文摘要 以色列极右翼国家安全部长本-格维尔呼吁每晚在加沙杀死“30至40名”巴勒斯坦人。作为安全内阁成员，他参与制定战争与政策。
'Ghana was the crime scene' - the country leading the demand for slavery reparations
Ghana has become Africa's leading voice in telling the world that Western nations owe reparations
中文摘要 加纳已成为非洲要求奴隶制赔偿的主要声音，该国表示西方国家对历史上的奴隶贸易负有责任，并称“加纳曾是犯罪现场”。
Silver Traders Brave Red Tape to Capture Indian Premiums
Shipments of silver into India are picking up despite administrative hurdles, as traders seek to capitalize on attractive premiums after a new licensing regime disrupted imports.
中文摘要 尽管面临行政障碍，白银对印度的发货量正在回升，交易商寻求利用新许可制度扰乱进口后出现的诱人溢价获利，以弥补此前贸易中断的影响。
Ferrari's first ever electric car sold for record $40m at auction
The Luce's launch drew backlash from some commentators who said it broke from the brand's signature dramatic styling.
中文摘要 法拉利首款电动汽车Luce在拍卖会上以创纪录的4000万美元成交，但其发布也引发一些评论人士批评，认为它背离了品牌标志性的动感造型设计。
Copper Heads Toward Record High on LME as Supply Tensions Build
Copper extended a seven-week rally to head toward a record, with a key price spread on the London Metal Exchange widening to unusual levels in a clear sign of rising competition for short-term supply.
中文摘要 铜价延续七周涨势，正逼近历史高位，伦敦金属交易所的关键价差扩大至非常罕见水平，暗示短期供应争夺日益激烈，供应紧张情绪持续升温。
Sri Lankan Inflation to Slow Toward Goal, Central Bank Head Says
Sri Lanka’s inflation is expected to slow back toward the central bank’s target of 5% over the second half of this year and into next year, according to Governor Nandalal Weerasinghe.
中文摘要 斯里兰卡央行行长Nandalal Weerasinghe近日表示，今年下半年和明年，该国通胀预计将逐步回落至央行5%的目标水平，期间经济增长也将放缓。
Olympus boss tries to beat buyout firms at their own game
Bob White adopts private equity-style tactics to deflect takeover interest and boost share price
中文摘要 据英国《金融时报》报道，奥林巴斯首席执行官Bob White正采用私募股权式策略，以化解收购兴趣并提振股价，尝试在收购博弈中抢占先机。
Trump Says He’s Curtailing Drills With South Korea
President Donald Trump said he told the Pentagon to “substantially reduce” joint military exercises with South Korea, citing the cost of the drills and the signal they send to North Korea. Bloomberg's John Herskovitz breaks down the situation. (Source: Bloomberg)
中文摘要 美国总统特朗普表示，他已指示五角大楼「大幅减少」与韩国的联合军事演习，理由包括军演成本高企，以及军演可能向朝鲜传递的信号。
The next China shock will come from open-source AI
Countries adopting Chinese models will also absorb Chinese standards and governance
中文摘要 英国《金融时报》评论文章称，采用中国开源AI模型的国家也将吸纳中国的标准和治理方式，下一轮「中国冲击」将来自开源人工智能。
Thai Baht Rally May Fade on Dovish Central Bank, Analysts Say
A nascent rally in the Thai baht may be short-lived as the central bank is likely to maintain a dovish policy stance to support an economy hit by elevated oil prices, according to analysts.
中文摘要 分析师称，泰铢近期的上涨行情可能是昙花一现，因该国央行很可能维持鸽派政策立场，以支持受油价高企冲击的经济，限制泰铢升值空间。
India’s sweltering nights expose power crunch
Rise in evening temperatures and air conditioning demand leaves gap in renewable energy supply
中文摘要 印度夜间气温不断上升，空调电力需求随之增加，暴露出可再生能源在晚间供应不足的缺口，导致电力紧张问题日益凸显，凸显该国电力系统的结构性短板。
Gold Steady Near $4,400 as Traders Weigh US Data and Rate Path
Gold was steady as traders weighed the competing impacts of cooling US economic data and energy-led inflation pressure on the Federal Reserve’s interest-rate path.
中文摘要 金价持稳于每盎司4400美元附近，交易员权衡美国经济数据降温与能源引起的通胀压力对美联储利率路径的竞争性影响，市场保持观望。
If Meta loses this trial, Instagram and Facebook could change forever
Thirty US states have sued Meta to force an overhaul of its platforms for young users.
中文摘要 美国30个州起诉Meta，要求强制其全面改革面向年轻用户的平台。若Meta在此案中败诉，Instagram与Facebook可能迎来永久性改变。
China’s energy strategy vindicated by Iran war
Playbook is likely to be adopted by other countries, shaping markets
中文摘要 英国《金融时报》报道称，伊朗战争印证了中国能源战略的正确性，这一做法可能被其他国家采纳，并对市场格局产生影响。
有没有在 Mac 下跑 ComfyUI 成功的大佬
8 回复 · 程序员 节点
大家在什么场景下还打算使用涨价后的 DeepSeek？
10 回复 · 程序员 节点
寄了， Opencode go v4flash 额度缩水 88%，双倍也结束了
12 回复 · 程序员 节点
想问问朋友们现在都是用的什么 Coding Agent？
18 回复 · 程序员 节点
iPhone 18 Pro 想买首发 想买非国行版本 有什么渠道吗？
25 回复 · Apple 节点
后台关闭程序和自动刷新
6 回复 · Apple 节点
WLOC 改定位「好像」在 iOS27 beta5 失效了
9 回复 · Apple 节点
升级 ios26.6 掉电池健康度
13 回复 · Apple 节点
看见有人说 m1pro 的 mbp,我想说,当初多花 3000 买的 32g 还是很香啊
60 回复 · Apple 节点
看到 AI 让 Linux 变得真正可用了帖子，想问下现在最推荐的是 Linux 发行版是哪个呢
106 回复 · Linux 节点
该源今日无内容。