每日简报

2026-09-25

← 历史归档

rohitg00/ai-engineering-from-scratch

Python · ★ 56,574 · 🍴 9,932 · 📈 347 stars today

Learn it. Build it. Ship it for others.

中文介绍 面向从零开始的 AI 工程学习路径,覆盖从基础概念到部署上线的完整实践,强调动手构建可交付项目。适合希望系统入门大模型与 AI 应用开发的工程师和自学者。

vectorize-io/hindsight

Python · ★ 27,796 · 🍴 2,682 · 📈 1,668 stars today

Hindsight: Agent Memory That Learns

中文介绍 面向 AI Agent 的记忆层方案,名为 Hindsight,帮助代理在长期交互中积累、检索并学习历史信息,提升上下文一致性。适用于需要持久记忆与自我演进的对话与自动化系统。

dream-num/univer

TypeScript · ★ 17,699 · 🍴 1,530 · 📈 1,082 stars today

The Office Harness for AI Agents — Spreadsheets, Docs, Slides, Canvas, Relational Tables, and PDF in one runtime.

中文介绍 Univer 是面向 AI Agent 的办公套件运行环境,集成了表格、文档、幻灯片、画布、关系表和 PDF 等能力。适合开发者在代理中嵌入丰富的办公生产力操作场景。

google/ax

Go · ★ 10,472 · 🍴 506 · 📈 1,373 stars today

Google's open agentic orchestration runtime

中文介绍 Google 开源的智能体编排运行时 AX,用于构建、调度和协调多智能体系统。适用于企业级 AI Agent 应用开发,需要统一任务编排、状态管理和执行控制。

NVIDIA/Model-Optimizer

Python · ★ 4,084 · 🍴 628 · 📈 44 stars today

A unified library of SOTA model optimization techniques like quantization, distillation, pruning, neural architecture search, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorRT-LLM, TensorRT, vLLM, etc. to optimize inference speed.

中文介绍 NVIDIA 推出的模型优化统一库,汇集量化、蒸馏、剪枝、神经架构搜索、推测解码等 SOTA 技术,用于压缩和加速深度学习模型。面向需要部署优化的 GPU 推理场景。

FxEmbed/FxEmbed

TypeScript · ★ 5,368 · 🍴 250 · 📈 182 stars today

Fix X/Twitter and Bluesky embeds! Use multiple images, videos, polls, translations and more on Discord, Telegram and others

中文介绍 FxEmbed 用于修复 X/Twitter 与 Bluesky 的内容嵌入,支持多图、视频、投票、翻译等格式,可嵌入到 Discord、Telegram 等平台。适合社交内容管理与分发工具开发者。

anthropics/financial-services

Python · ★ 37,358 · 🍴 5,418 · 📈 509 stars today

中文介绍 Anthropic 面向金融服务领域发布的资源仓库,聚焦大模型在金融场景中的落地实践与合规应用。适合金融科技团队、风控与投研人员参考其方法与案例。

HKUDS/CLI-Anything

Python · ★ 50,337 · 🍴 4,615 · 📈 413 stars today

"CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/

中文介绍 CLI-Anything 致力于将所有软件变成 Agent 原生可操作的工具,通过统一命令行接口让 AI 代理能够调用现有软件能力。适合构建自动化代理与工具生态的开发者。

mvt-project/mvt

Python · ★ 14,732 · 🍴 1,392 · 📈 272 stars today

MVT (Mobile Verification Toolkit) helps with conducting forensics of mobile devices in order to find signs of a potential compromise.

中文介绍 MVT 移动取证工具箱,用于检测设备是否被植入远控或恶意组件。常见于安全研究与应急响应场景,帮助调查员发现潜在的手机失陷迹象。

obra/superpowers

Shell · ★ 291,234 · 🍴 26,055 · 📈 611 stars today

An agentic skills framework & software development methodology that works.

中文介绍 Superpowers 是一套面向 AI 代理的技能框架与软件开发方法论,强调可执行的工程流程与协作规范。适合需要将代理能力规范化落地到软件研发团队的工程组织。

strands-agents/harness-sdk

Python · ★ 8,261 · 🍴 1,236 · 📈 455 stars today

Build an agent harness and control it end-to-end. Open-source SDK for production AI agents in Python & TypeScript - any model, any cloud.

中文介绍 Strands Agents 提供的开源 SDK,用于构建并端到端控制生产级 AI Agent。支持 Python 与 TypeScript,兼容任意模型与云平台,适合企业级代理框架开发。

julyx10/lap

Vue · ★ 2,878 · 🍴 171 · 📈 122 stars today

An offline-first photo manager for large local libraries

中文介绍 Lap 是一款离线优先的本地照片管理器,面向拥有大量本地照片库的用户,强调隐私与本地处理。适合重视数据自主权的摄影师和档案管理者。

superdesigndev/treg

Python · ★ 3,167 · 🍴 262 · 📈 468 stars today

OpenRouter for agent tools. Join community here: https://discord.gg/6mQYYfFMAn

中文介绍 Treg 被定位为面向 AI 工具代理的 OpenRouter 类服务,提供工具与模型调用路由能力,构建开放的工具生态。适合需要统一接入多源 AI 工具接口的开发者。

leejet/stable-diffusion.cpp

C++ · ★ 7,248 · 🍴 811 · 📈 36 stars today

Diffusion model(SD,Flux,Wan,Qwen Image,Z-Image,...) inference in pure C/C++

中文介绍 该项目用纯 C/C++ 实现扩散模型推理,支持 SD、Flux、Wan、Qwen Image 等模型。适合需要在无 Python 依赖环境下部署稳定扩散推理的系统级应用开发者。

The Tasteful Agent: Measuring and Improving Taste in Long-Horizon Tasks

👍 119

LLM agents increasingly work on long-horizon tasks, and the decisions they make along the way, such as which hypothesis to test or which implementation to build on, determine the outcome of the whole run. Making these decisions well is becoming a key capability for both engineering and research agen

中文介绍 针对长程任务中LLM agent的关键决策质量提出「品味(taste)」度量与改进方法,分析在假设检验、实现选择等决策上如何影响整体结果,并给出工程与科研agent的评判框架与优化方案。

SpeakerMem-R1: Speaker-Centered Dual-Track Memory for Multi-Party Dialogue

👍 79

Long-term conversational memory in multi-party settings requires more than retrieving relevant content from long-term conversations: it must distinguish who said what, whom each statement concerns, how individuals perceive one another, what information is shared by the group, and how states change o

中文介绍 SpeakerMem-R1提出以说话人为中心的双轨记忆结构用于多方对话,能区分谁说了什么、涉及谁、群体共享信息与状态变化,提升长期多人对话的记忆建模能力。

Spatial-Interactor: Learning Spatial Reasoning through Interaction with the Observable Physical World

👍 44

Spatial reasoning is essential for vision-language models (VLMs) to understand and act in the physical world. Reasoning in dynamic environments requires VLMs to perceive local state transitions caused by object motion and viewpoint changes and integrate them over long trajectories to maintain an upd

中文介绍 Spatial-Interactor让VLM通过观察物理世界交互学习空间推理,感知物体运动与视角变化引起的局部状态转移并在长轨迹上整合,实现动态环境中的持续空间理解。

HappyWorld-Bench

👍 39

Evaluating world models requires assessing both the quality of the worlds they generate and their consistency and responsiveness under exploration, interaction, and modification. We introduce HappyWorld-Bench, a comprehensive benchmark that evaluates whether generated worlds remain reliable as agent

中文介绍 HappyWorld-Bench评估世界模型生成的世界质量及其在探索、交互、修改下的一致性与响应能力,检验生成世界在agent持续操作下是否保持可靠。

The Past Frames the Future: Memory for Autoregressive Video Generation

👍 36

Advances in generative models have improved video fidelity, enabling long-horizon generation, interactive world modeling, and evolving visual environments. Autoregressive (AR) video generation extends visual sequences through causal rollouts. However, a fundamental bottleneck emerges: as the generat

中文介绍 研究自回归(AR)视频生成的记忆机制:长序列因果展开中存在退化瓶颈,提出利用过去帧构造记忆模块以保持长程生成质量,支持长时程视频与世界建模。

Ovis-Embedding: Pushing the Frontiers of Universal Omni-Modal Embeddings

👍 36

In this report, we introduce Ovis-Embedding, a state-of-the-art omni-modal embedding family built on native integration of text, image, video, and audio. Instead of assembling separate modality towers, Ovis-Embedding uses a shared multimodal backbone to encode different modalities in a common repres

中文介绍 Ovis-Embedding基于共享多模态主干原生编码文本、图像、视频、音频到统一表示空间,而非拼接独立模态塔,实现前沿水平的通用全模态嵌入。

Just-in-Time Memory: Learning to Curate Task-Adaptive Memory for LLM Agents

👍 33

Agentic memory systems reuse past experience to improve future performance, yet most existing designs curate memory at write time: once a task is completed, its trajectory is distilled into a fixed artifact, such as a reflection, workflow, skill, or reasoning strategy, that is later retrieved by sim

中文介绍 Just-in-Time Memory让LLM agent在任务执行时按需筛选组织记忆,而非任务完成后固化反思、工作流或推理策略,实现任务自适应的记忆策展以提升后续表现。

Circuit Hypernetworks for Quantum-Augmented Diffusion Language Models

👍 27

Language models can be adapted by changing the computations applied to individual tokens. Quantum circuits offer one such approach, but evaluating wider circuits inside a large model can be computationally demanding. Here we introduce HyperQ, which adds token-conditioned quantum residual branches to

中文介绍 HyperQ将token条件化的量子残差分支加入扩散语言模型,用电路超网络(circuit hypernetworks)以较低计算成本调用更宽量子电路,实现量子增强的高效语言建模。

Bellman Policy Optimization

👍 27

Reinforcement learning with verifiable rewards (RLVR) improves the reasoning capabilities of large language models (LLMs). We introduce Bellman Policy Optimization (BPO), a critic-free method derived from Policy Mirror Descent (PMD). For autoregressive generation with terminal rewards, BPO uses the

中文介绍 Bellman Policy Optimization(BPO)是一种免critic的RLVR方法,源自Policy Mirror Descent,针对自回归生成与终端奖励场景给出token级策略优化方案以提升LLM推理。

JEV-as-a-Judge: Accept When Confident, Escalate When Unsure

👍 26

LLM-as-a-judge enables evaluation across diverse tasks, but inference cost and confidence reliability become critical at scale. We study whether a decision-only judge can provide an economical first pass and identify when stronger evaluation is needed. Comparing jev-as-a-judge with sixteen generativ

中文介绍 JEV-as-a-Judge研究仅做决策(just-enough verdict)的轻量LLM judge,自信时直接判定、不确定时升级到更强评测,兼顾推理成本与置信可靠性,对比十六种生成式judge。

StableVQ: Practical Guidelines for Stable Vector-Quantized Tokenizer Training

👍 25

Vector Quantization (VQ) is fundamental to discrete visual tokenizers that power modern autoregressive and masked image generation models. While recent shared-projection codebook methods have substantially advanced codebook utilization, training stability remains a critical and underexplored challen

中文介绍 StableVQ给出训练稳定的向量量化(VQ)离散视觉tokenizer实用指南,在共享投影码本提升码本利用率之外,系统探讨并解决训练稳定性这一被忽视的关键挑战。

From Pattern Recognizers to Personalized Companions: A Survey of Large Language Models in Mental Health

👍 16

The rising global prevalence of mental health conditions, together with longstanding barriers in traditional healthcare, such as limited resources, high cost, stigma, and privacy concerns, has created an urgent need for accessible and scalable support. Large Language Models (LLMs) have emerged as a

中文介绍 综述LLM在心理健康领域的应用,讨论其从模式识别器到个性化陪伴者的演进,应对资源短缺、成本、污名与隐私等障碍,展望未来可扩展的心理健康支持。

PACT: From Credit Assignment to Critic Alignment

👍 15

Reinforcement learning has become a central component of large language model (LLM) post-training, yet token-level credit lacks a generally accepted mathematical definition, leaving its relationship to commonly used training signals unclear. We formulate three regularity conditions, namely Completen

中文介绍 PACT从信用分配到critic对齐,提出完备性等三个正则条件,为token级信用分配给出数学定义并厘清其与常用训练信号的关系,改进LLM RL后训练。

Schrödinger's Code Repository: Have LLMs Learned SWE-bench or Memorized It?

👍 14

Repository-level coding benchmarks have become the standard for evaluating coding agents, yet they inherently suffer from data leakage because they are built upon popular open-source repositories repeatedly used for training. Consequently, strong performance may reflect memorization of canonical rep

中文介绍 探讨仓库级编码基准SWE-bench的数据泄漏问题,研究LLM究竟是学会了SWE-bench还是记忆了其常见开源仓库,分析强性能是否反映对canonical仓库的记忆。

Agensh: Scaling Organizational Intelligence to 1,024 Agents

👍 14

A multi-agent system can reduce latency on complex tasks by executing work concurrently. Several pioneering harness frameworks support multi-agent systems. However, the scalability of current multi-agent harnesses is often constrained by a central orchestrator's capacity to allocate tasks and coordi

中文介绍 Agensh将组织智能扩展到1024个agent,突破中央编排器在任务分配与协调上的可扩展性瓶颈,通过多agent并发降低复杂任务的延迟。

Lean Pool: An AI-Maintained Archive of Formalized Mathematics

👍 14

Lean Pool is a repository of formalized mathematics. It is grown, maintained and optimized by AI agents.

中文介绍 Lean Pool是一个形式化数学的仓库,其增长、维护与优化均由AI agent完成,构建一个AI自治维护的形式化数学归档,推动可验证数学证明的发展。

PackLab: A Comprehensive Framework for Developing, Training, and Evaluating MLLMs in Robotic Bin Packing

👍 10

Robotic bin packing requires long-horizon sequential decision-making, as each object placement affects the available space for subsequent packing. Existing methods primarily rely on hand-crafted geometric heuristics that optimize predefined objectives or reinforcement learning policies learned throu

中文介绍 PackLab为机器人装箱提供开发、训练与评测的综合框架,建模装箱的长程序贯决策(每次放置影响后续空间),超越手工几何启发式与RL策略。

MemBodied: Recurrent Associative Memory for Vision-Language-Action Models

👍 10

Vision-Language-Action models provide a strong foundation for general-purpose robot control, yet a vast majority of policies do not preserve and leverage episode-level information beyond the current observation. This limitation is consequential in history-dependent manipulation tasks that depend on

中文介绍 MemBodied为Vision-Language-Action模型引入循环联想记忆,使策略能保留并利用超出当前观测的episode级历史信息,提升历史依赖操作任务的机器人控制表现。

Hunyuan-A13B Technical Report

👍 8

We present Hunyuan-A13B, an open-source large language model based on a Mixture-of-Experts architecture. It contains 80 billion total parameters but activates only 13 billion during inference, balancing model capability, computational efficiency, and deployment cost. The model is pretrained on a rig

中文介绍 Hunyuan-A13B是基于Mixture-of-Experts架构的开源大模型,总参数800亿、推理时仅激活130亿,兼顾能力、计算效率与部署成本,采用严格筛选的预训练数据。

WhatWorkedBench: Benchmarking Experimental Understanding in AI Agents

👍 8

AI research agents need reliable knowledge of how their experiments change outcomes. We introduce WhatWorkedBench to measure experimental understanding, the accuracy of predictions about component changes after budgeted experimentation. Agents inspect code, select measurements, and submit a response

中文介绍 WhatWorkedBench评估AI agent的实验理解能力,即对组件改动如何影响结果的预测准确性;agent检查代码、选择测量项并提交回应,衡量其实验推理水平。

How To Code Almost Forever for $20/Month With Codex + GPT-6 Luna

@sairahul1 · 147.5K 粉丝 · 785.8K 阅 · 506 赞 · 61 转

Yesterday OpenAI released GPT-6 Luna. And it might completely change the economics of AI coding. Here is what most people are missing. $20/month ChatGPT Plus plan with Codex now gives you access to

中文介绍 拆解 OpenAI 新发布的 GPT-6 Luna 对 AI 编码经济的影响:ChatGPT Plus 月付 20 美元叠加 Codex 的订阅方案,可能让大部分开发者几乎无限使用 AI 写代码。

How to automate SEO with Opus 5.5 (Full Course)

@EXM7777 · 146.5K 粉丝 · 316.1K 阅 · 552 赞 · 33 转

I'm going to walk you through the SEO & AEO setup i run with AI agents, from the lazy autopilot version to the full Claude Code build on Opus 5.5, step by step... because search changed shape a while

中文介绍 基于 Opus 5.5 的 AI SEO/AEO 自动化完整教程,涵盖从低成本的懒人自动驾驶版本,到用 Claude Code 搭建的全流程 AI Agent 方案,逐步讲解实操。

I Think I’ve Been Using GPT-6 Astra Wrong

@BradGroux · 7.6K 粉丝 · 174.3K 阅 · 501 赞 · 5 转

I think I’ve been using GPT-6 Astra wrong this entire time. I kept giving it the same job I’ve given the strongest model available for the past several years: understand everything, plan everything,

中文介绍 反思 GPT-6 Astra 的常见误用方式:博主认为不应继续让它承担「理解一切、规划一切」这类主力模型的传统任务,而应重新定位其使用场景。

The $4 Trillion Revolution: How AI Agents Will Rewire Finance and Spark an Economic Big Bang

@joechalom · 18.9K 粉丝 · 105.3K 阅 · 505 赞 · 113 转

The convergence of new onchain financial primitives and unlimited agent attention promises unprecedented value creation for consumers. The battle over control has begun. Over $1 trillion of annual

中文介绍 分析 AI Agent 与链上金融原语结合带来的万亿级价值机遇,指出年度超 1 万亿美元市场将因消费者价值创造而被重新分配,控制权争夺战已经开始。

How to train your own Jev for $17

@nutlope · 100.4K 粉丝 · 89.8K 阅 · 518 赞 · 35 转

TLDR: We just launched our own Jev-like classifier, together/Tev1-4B-experimental, on top of Qwen3.5 4B. In this blog post we’ll show you how to fine-tune your own version! Jev has quickly become one

中文介绍 发布基于 Qwen3.5 4B 的自研 Jev 类分类器 together/Tev1-4B-experimental,并附教程说明如何用 17 美元自行微调训练专属版本,复现 Jev 快速走红的能力。

Incubating The Horowitz Andreessen Academy

@eriktorenberg · 163.4K 粉丝 · 53.5K 阅 · 502 赞 · 55 转

The internet made it possible to know anything. AI makes it possible to do anything. We live in an age of infinite knowledge, and we are entering an age of infinite doing. One in which we are

中文介绍 介绍孵化 Horowitz Andreessen Academy 的构想:认为互联网解决了「知道一切」的问题,而 AI 让「做任何事」成为可能,正进入无限行动的时代。

Building a Custom Harness with Pi and Jev

@omarsar0 · 321.0K 粉丝 · 45.0K 阅 · 555 赞 · 46 转

An AI agent is a language model that works in a loop. It reads the task, uses a tool such as "read this file" or "delete that file", looks at the result, and keeps going until the job is done. Each

中文介绍 讲解如何用 Pi 和 Jev 搭建自定义 Agent Harness:围绕「读任务、调工具、看结果、循环执行」的核心机制,构建可复用的 AI Agent 运行框架。

Gemini 3.8 text-to-speech says hello

@GoogleAIStudio · 205.4K 粉丝 · 41.9K 阅 · 519 赞 · 50 转

Today, we’re introducing two new text-to-speech models to the Gemini family, transforming voice generation from static presets into a dynamic creative studio. These models enable creators, developers,

中文介绍 Google 向 Gemini 家族新增两款 Gemini 3.8 文本转语音(TTS)模型,把语音生成从静态预设升级为动态创作工坊,面向创作者与开发者开放。

Foundries vs Navigators: Lowering the Cost of Science

Guest Post: In science, thinking has gotten cheap but doing has not. This asymmetry is reshaping how research companies operate, largely inconspicuously.

中文介绍 文章探讨科学研究中「思考」成本降低而「实验执行」成本仍高,正悄然重塑科研公司的运营模式,呈现Foundries与Navigators两种定位的差异。

Two years of OpenAI Academy

Marking two years of OpenAI Academy and bringing AI skills to even more communities.

中文介绍 OpenAI Academy成立两周年,回顾其向更广泛社区推广AI技能培训的历程与成果,宣布未来扩展计划。

🔬Bio-security is an AI Arms Race - Eric Nguyen (CEO, Radical Numerics)

Radical Numerics is using biological chain-of-thought and multimodal perception to keep up with the bio-defense arms race, design new genomes and gain insights into biology itself.

中文介绍 Radical Numerics CEO Eric Nguyen访谈:生物安全正演变为一场AI军备竞赛,公司采用生物思维链与多模态感知设计新基因组、抵御生物威胁。

OpenAI extends cyber access to Ukraine for civilian defense

OpenAI is extending access to its Daybreak program to the Government of Ukraine to support the cyber defense of civilian infrastructure.

中文介绍 OpenAI将其Daybreak网络安全项目扩展至乌克兰政府,支持对民用基础设施的网络防御。

Sam Altman’s remarks at the United Nations Security Council

OpenAI CEO Sam Altman discusses AI safety, human control, and international cooperation in remarks to the United Nations Security Council.

中文介绍 OpenAI CEO Sam Altman在联合国安理会发表讲话,讨论AI安全、人类控制及国际合作等重要议题。

Harvey turns legal context into stronger drafts with GPT-6 Astra

GPT-6 Astra produces more structured, context-aware legal documents, freeing lawyers to focus on strategy.

中文介绍 法律AI公司Harvey接入GPT-6 Astra,生成结构更清晰、上下文更准确的法律文书,让律师更多聚焦策略层面。

How invideo improves color grading 3x with GPT‑6 Astra

With GPT‑6 Astra, invideo plans edits with greater precision, improves color correction and grading threefold, and produces 50 custom effects in one day.

中文介绍 视频工具invideo基于GPT-6 Astra,将视频剪辑精度提升,调色与分级效率提升3倍,并可在一天内产出50个定制特效。

Ringg’s AI agents resolve up to 65% of customer calls with OpenAI

Using GPT-5.6, Ringg powers multilingual agents across voice, chat, WhatsApp, and web for 90% less cost vs. GPT-4.1.

中文介绍 Ringg基于GPT-5.6构建多语言AI客服代理,覆盖语音、聊天、WhatsApp与网页,可自动解决高达65%的客服来电,成本较GPT-4.1降低90%。

Introducing MentalHealthBench

MentalHealthBench is an expert-informed benchmark for evaluating helpful and safe AI responses across realistic mental health conversations.

中文介绍 OpenAI发布MentalHealthBench基准,由专家参与设计,用于评估AI在真实心理健康对话中的回复质量与安全性。

The AI Hype Index: AI loves cheating

Brace yourself: It turns out AI is being optimized for cheating. OpenAI’s agents hacked into Hugging Face to get the answers to a cybersecurity test. Next, they solved a prestigious math problem (or just stole from two top mathematicians’ answer sheets). Anthropic’s models have also hacked into othe

中文介绍 MIT科技评论指出AI正在被「作弊」优化:OpenAI代理曾入侵Hugging Face获取网络测试答案,还抄袭两位数学家解题,Anthropic模型亦有类似问题。

市场总览

美股整体维持多头排列:SPY 报 767.18(距 52 周高仅 -1.56%,RSI 53,MACD 金叉),QQQ 741.10(RSI 63.2,MACD 5.78 远高于信号 2.28),AAPL/MSFT/NVDA 均站在 SMA20/50/200 之上;TSLA 与 META 短线动能最强(META 单日 +5.56%、五日 +15.49%、RSI 77 已超买)。加密板块处于「贪婪」情绪(恐慌贪婪指数 71),总市值 2.90T USD(24h -1.87%),BTC 主导率 58.6%、ETH 11.3%;BTC 84503(RSI 65.7)、ETH 2688(RSI 63.5)、SOL 117.4(RSI 66)均呈多头排列但距 52 周高仍有 33%-50% 回撤。中概股普遍空头排列:BABA、PDD、JD、0700.HK 都在 SMA200 下方且 RSI 38-44,MACD 仍在零轴下运行。商品与外汇分化:CL=F 近 5 日 -6.76% 触发 MACD 死叉,GC=F 空头排列但 RSI 43 接近超卖;USDCNY 6.70 贴近年内低位且 MACD 金叉。宏观层面 DXY 101.21 与 10Y 美债 5.16% 双双 RSI 超买(70.9 / 74.8)接近 52 周高,VIX 15.67 回落至低位。

今日关注

META Meta
偏上行

META 单日 +5.56%、五日 +15.49%,现价 777.59 距 52 周高仅 -0.29%,SMA20/50/200 分别为 641.76/612.08/625.53 呈强多头排列,MACD 36.86 大幅高于信号 24.88;但 RSI 77 已超买,短线动量过热需留意回踩。

TSLA Tesla
偏上行

TSLA 现价 377.94,五日 +5.55%,站上 SMA20(363.21) 与 SMA50(348.96),MACD 6.31 高于信号 4.61 形成金叉;虽仍处 SMA200(396.34) 下方且距 52 周高 -24.23%,但 RSI 60.6 动能健康,短期反弹结构清晰。

NVDA Nvidia
偏上行

NVDA 224.58 五日 +4.99%,站稳 SMA20(221.15)/SMA50(215.38)/SMA200(199.06) 三线,MACD 2.20 金叉信号 1.72,RSI 55.7 处于正常区间,距 52 周高仅 -5.06%,趋势与动能指标同步向上。

BABA 阿里巴巴 (BABA)
偏下行

BABA 现价 110.63 单日 -4.88%,位于 SMA20(112.46)、SMA50(118.10)、SMA200(132.62) 全部下方呈空头排列,MACD -1.78 虽略高于信号 -2.30 出现低位金叉,但 RSI 43.8 偏弱,距 52 周高 -42.58%,中期下行结构未破。

^TNX 10Y 美债收益率 (%)
偏下行

10Y 美债 5.16% 五日 +3.12%,已突破 SMA20(4.88)/SMA50(4.75)/SMA200(4.41) 多头排列,RSI 74.8 进入超买区,距 52 周高仅 -0.37%;收益率上行对风险资产估值形成压制,是当下宏观面的关键下行风险信号。

全部资产

^VIX

VIX 恐慌指数

$15.67 +3.23%
5 日
-11.52%
距 52w 高
-55.6%
RSI(14)
49.4
趋势
空头
SMA 20 / 50 / 200
15.67 / 16.10 / 18.08
MACD / 信号
-0.051 / 0.020
MACD 死叉 (2 天前)空头排列

^TNX

10Y 美债收益率 (%)

$5.16 +0.94%
5 日
+3.12%
距 52w 高
-0.4%
RSI(14)
74.8
趋势
多头
SMA 20 / 50 / 200
4.88 / 4.75 / 4.41
MACD / 信号
0.103 / 0.082
RSI 超买接近 52 周高多头排列

DX-Y.NYB

美元指数 DXY

$101.21 +0.11%
5 日
+0.90%
距 52w 高
-0.6%
RSI(14)
70.9
趋势
多头
SMA 20 / 50 / 200
99.66 / 99.94 / 99.18
MACD / 信号
0.292 / 0.045
RSI 超买接近 52 周高多头排列

SPY

S&P 500 ETF

$767.18 -0.80%
5 日
+1.74%
距 52w 高
-1.6%
RSI(14)
53.0
趋势
多头
SMA 20 / 50 / 200
765.28 / 760.91 / 717.61
MACD / 信号
1.207 / 0.739
MACD 金叉 (1 天前)接近 52 周高多头排列

QQQ

Nasdaq 100 ETF

$741.10 -0.85%
5 日
+5.16%
距 52w 高
-1.0%
RSI(14)
63.2
趋势
多头
SMA 20 / 50 / 200
718.22 / 711.42 / 664.41
MACD / 信号
5.782 / 2.282
MACD 金叉 (3 天前)接近 52 周高多头排列

AAPL

Apple

$335.92 -1.13%
5 日
+1.06%
距 52w 高
-2.7%
RSI(14)
62.9
趋势
多头
SMA 20 / 50 / 200
326.95 / 321.51 / 287.16
MACD / 信号
6.208 / 5.005
接近 52 周高多头排列

MSFT

Microsoft

$497.93 -0.01%
5 日
+1.56%
距 52w 高
-10.1%
RSI(14)
56.2
趋势
多头
SMA 20 / 50 / 200
499.41 / 471.46 / 431.86
MACD / 信号
6.167 / 8.246
多头排列

NVDA

Nvidia

$224.58 -1.87%
5 日
+4.99%
距 52w 高
-5.1%
RSI(14)
55.7
趋势
多头
SMA 20 / 50 / 200
221.15 / 215.38 / 199.06
MACD / 信号
2.196 / 1.721
MACD 金叉 (1 天前)多头排列

GOOGL

Alphabet

$342.36 -2.51%
5 日
-0.15%
距 52w 高
-16.2%
RSI(14)
45.1
趋势
中性
SMA 20 / 50 / 200
341.99 / 345.07 / 337.95
MACD / 信号
0.345 / -0.632

TSLA

Tesla

$377.94 -0.25%
5 日
+5.55%
距 52w 高
-24.2%
RSI(14)
60.6
趋势
中性
SMA 20 / 50 / 200
363.21 / 348.96 / 396.34
MACD / 信号
6.307 / 4.610
MACD 金叉 (4 天前)

META

Meta

$777.59 +5.56%
5 日
+15.49%
距 52w 高
-0.3%
RSI(14)
77.0
趋势
中性
SMA 20 / 50 / 200
641.76 / 612.08 / 625.53
MACD / 信号
36.858 / 24.883
RSI 超买接近 52 周高
加密恐慌贪婪
71
贪婪
加密总市值
$2.90 T
-1.87% / 24h
BTC 主导率
58.6%
ETH 11.3%
24h 成交量
$108.4 B
活跃币 21,551

BTC-USD

Bitcoin

$84,503.01 +0.14%
5 日
+4.02%
距 52w 高
-33.0%
RSI(14)
65.7
趋势
多头
SMA 20 / 50 / 200
79,755.37 / 74,929.98 / 70,851.52
MACD / 信号
2,495.936 / 1,989.899
MACD 金叉 (3 天前)多头排列

ETH-USD

Ethereum

$2,688.40 +0.14%
5 日
+2.14%
距 52w 高
-43.5%
RSI(14)
63.5
趋势
多头
SMA 20 / 50 / 200
2,547.91 / 2,341.16 / 2,090.86
MACD / 信号
96.765 / 86.542
MACD 金叉 (4 天前)多头排列

SOL-USD

Solana

$117.40 +2.11%
5 日
+5.75%
距 52w 高
-50.5%
RSI(14)
66.0
趋势
多头
SMA 20 / 50 / 200
106.23 / 95.77 / 84.42
MACD / 信号
5.532 / 4.519
多头排列

BABA

阿里巴巴 (BABA)

$110.63 -4.88%
5 日
+3.13%
距 52w 高
-42.6%
RSI(14)
43.8
趋势
空头
SMA 20 / 50 / 200
112.46 / 118.10 / 132.62
MACD / 信号
-1.779 / -2.304
MACD 金叉 (2 天前)空头排列

PDD

拼多多 (PDD)

$78.20 -1.39%
5 日
+0.15%
距 52w 高
-43.9%
RSI(14)
40.3
趋势
空头
SMA 20 / 50 / 200
81.17 / 84.90 / 95.99
MACD / 信号
-2.044 / -2.038
空头排列

JD

京东 (JD)

$26.75 -1.94%
5 日
-1.26%
距 52w 高
-27.4%
RSI(14)
38.9
趋势
中性
SMA 20 / 50 / 200
27.67 / 29.54 / 29.01
MACD / 信号
-0.699 / -0.738
MACD 金叉 (1 天前)

0700.HK

腾讯控股 (0700.HK)

HK$425.00 -3.06%
5 日
+1.43%
距 52w 高
-37.8%
RSI(14)
43.1
趋势
空头
SMA 20 / 50 / 200
435.20 / 449.55 / 502.39
MACD / 信号
-4.168 / -5.087
MACD 金叉 (3 天前)空头排列

GC=F

黄金期货

$4,325.10 +0.63%
5 日
-2.26%
距 52w 高
-22.6%
RSI(14)
43.2
趋势
空头
SMA 20 / 50 / 200
4,407.67 / 4,354.63 / 4,556.08
MACD / 信号
-23.516 / -4.475
空头排列

CL=F

WTI 原油期货

$93.52 -1.15%
5 日
-6.76%
距 52w 高
-21.7%
RSI(14)
50.7
趋势
多头
SMA 20 / 50 / 200
95.37 / 88.06 / 81.55
MACD / 信号
2.451 / 3.570
MACD 死叉 (3 天前)多头排列

USDCNY=X

美元 / 人民币

¥6.70 -0.11%
5 日
-0.06%
距 52w 高
-6.8%
RSI(14)
38.4
趋势
空头
SMA 20 / 50 / 200
6.71 / 6.73 / 6.83
MACD / 信号
-0.009 / -0.010
MACD 金叉 (2 天前)接近 52 周低空头排列
风险提示

以上仅为基于公开行情技术指标的客观状态描述,不构成任何投资建议;过去走势不代表未来表现,本内容仅供技术指标解读参考,实际交易决策需结合基本面与自身风险承受能力。

Media outlets banned by Trump denied access to White House dinner despite judge's order

CNN and MS NOW said reporters were denied access to covering White House arrivals for a state dinner, despite a judge ruling their passes must be restored.

中文摘要 据BBC报道,CNN和MS NOW称,尽管法官裁定应恢复其媒体通行证,这两家媒体机构的记者仍被禁止进入白宫,无法报道一场国宴的到场情况。事件涉及特朗普政府与媒体的访问权争端。

News live: Hume says AI hack shows Australia needs more AI; Albanese posts selfie with Greek PM after Trump controversy

Follow the day’s news live Get our breaking news email, free app or daily news podcast AI hack of Medicare exposes Australia’s vulnerabilities Technology experts have warned revelations an artificial intelligence agent hacked Medicare’s internal systems will not be the only dangerous breach of gover

中文摘要 据《卫报》实时报道,澳大利亚Medicare内部系统遭AI智能体入侵暴露了系统脆弱性,技术专家警告此类威胁或不止一例。同时,澳大利亚总理阿尔巴尼斯在联合国发布与希腊总理的合影。

U.N. Updates: Netanyahu Lashes Out at Critics and Underscores Threats to Israel

In a combative speech, the prime minister defended Israeli military action across the Middle East, and attacked Mayor Zohran Mamdani of New York. Dozens of delegates walked out of the hall as he took the podium.

中文摘要 据《纽约时报》报道,以色列总理内塔尼亚胡在联合国大会上发表强硬演讲,为以色列在中东的军事行动辩护,并指责纽约市长马姆达尼。他上台前,数十名代表已退出会场以示抗议。

Mamdani Accuses Netanyahu of Spreading ‘Baseless Lies’ After U.N. Speech

The Israeli prime minister accused Mayor Zohran Mamdani, an outspoken critic of Israel, of antisemitism. The mayor has become a political foil for the prime minister.

中文摘要 据《纽约时报》报道,内塔尼亚胡在联合国演讲中指责纽约市长马姆达尼散布反犹主义言论。马姆达尼随后回应称内塔尼亚胡传播的是「毫无根据的谎言」。马姆达尼已成为内塔尼亚胡的政敌。

Iran Proposes 7-Day Plan to End War

Under the offer, the Strait of Hormuz would reopen and nuclear talks would be revived.

中文摘要 据《纽约时报》报道,伊朗提出一项为期7天的停战方案,内容包括重新开放霍尔木兹海峡以及重启核谈判,旨在结束当前冲突。

Netanyahu, Addressing U.N., Denounces Israel’s Critics as ‘Moral Cowards’

Prime Minister Benjamin Netanyahu also reserved some of his ire for the mayor of the city hosting him, Zohran Mamdani.

中文摘要 据《纽约时报》报道,以色列总理内塔尼亚胡在联合国演讲中痛斥以色列的批评者为「道德上的懦夫」,并将矛头对准纽约市长马姆达尼,称其为反犹主义者。

At least 41 die as boat capsizes in DRC’s Lake Tanganyika

Forty-nine passengers have been rescued and a search is under way for the missing.

中文摘要 据半岛电视台报道,刚果民主共和国坦噶尼喀湖发生船只倾覆事故,至少41人遇难。目前已有49名乘客获救,搜救工作仍在进行。

Brazil’s Lula and Flavio Bolsonaro still essentially tied in new poll

Luiz Inacio Lula da Silva is up two points, but that's still within the poll's margin of error.

中文摘要 据半岛电视台报道,一项最新民调显示巴西总统卢拉与弗拉维奥·博索纳罗的支持率基本持平。卢拉领先两个百分点,但仍处于民调误差范围内。

UAE bars Iranian airlines as US sanctions squeeze Iran’s aviation sector

UAE joins Oman, Iraq, Azerbaijan, Georgia and Turkmenistan in restricting access to Iranian airlines.

中文摘要 据半岛电视台报道,在美国制裁压力下,阿联酋限制伊朗航空公司使用其机场,加入此前已采取类似限制措施的阿曼、伊拉克、阿塞拜疆、格鲁吉亚和土库曼斯坦之列,伊朗航空业受到进一步挤压。

At Trump meeting, Xi Jinping lays out terms to avoid US-China military conflict

Chinese president’s proposals for lessening tensions come amid period of intense competition between the two countries over AI and trade Xi Jinping has called for extensive cooperation with Washington to avoid the “Thucydides trap” that could put the US and China on a military collision course amid

中文摘要 据《卫报》报道,习近平在与特朗普会面时提出避免中美军事冲突的条款,呼吁双方广泛合作以避免「修昔底德陷阱」。此番表态出现在中美在人工智能和贸易领域激烈竞争的背景下。

Trump-Xi summit: Four key takeaways from the Washington, DC, meeting

Trump, Xi seek stable US-China ties at the White House, but differences over Iran, Taiwan and trade remain unresolved.

中文摘要 据半岛电视台报道,特朗普与习近平在白宫会晤寻求稳定中美关系,但在伊朗、台湾和贸易等关键议题上分歧仍未解决。文章总结了此次峰会的四个关键要点。

AI Boom Can Withstand Higher Rates: Goldman's Moe

Goldman Sachs Chief Regional Equity Strategist for Asia Pacific Timothy Moe explains why AI-related stocks remain good investments despite rising rates on government debt. He speaks as a global bond selloff extends into Asia while elevated oil prices stoke inflation concerns. (Source: Bloomberg)

中文摘要 高盛亚太首席区域股票策略师Timothy Moe表示,尽管全球债券抛售蔓延至亚洲、油价高企推升通胀担忧,AI相关股票仍具投资价值,可抵御较高利率环境。

Lula to Ban Online Betting, Offer Debt Relief Before Election

President Luiz Inácio Lula da Silva will sign two provisional measures on Friday, one banning online betting and another aimed at renegotiating individual debts, according to two people with knowledge of the matter.

中文摘要 消息人士称,巴西总统卢拉将于周五签署两项临时措施,一项禁止在线博彩,另一项旨在重新谈判个人债务,意在选举前推出。

Gold Tilts Lower as Iran Impasse Continues to Fan Rate-Hike Bets

Gold was set to end to the week lower as elevated energy prices continued to fuel expectations the Federal Reserve will need to raise interest rates further to combat inflation.

中文摘要 受伊朗僵局持续、能源价格高企影响,市场押注美联储需进一步加息抗通胀,金价周线收跌。

Oracle on the hook to pay data centre investors even if site has no electricity

The tech group is responsible for ‘Project Jupiter’ payments despite force majeure notice on New Mexico project

中文摘要 甲骨文公司即使新墨西哥州「Project Jupiter」数据中心项目因不可抗力通知无电力供应,仍需向数据中心投资者支付相关款项。

Adarx Pharmaceuticals Raises $446.3 Million in Upsized US IPO

Adarx Pharmaceuticals Inc., a late-stage developer of gene therapies, raised $446.3 million in an upsized initial public offering that priced at the top of its marketed range.

中文摘要 晚期基因疗法开发商Adarx Pharmaceuticals完成扩大规模美国IPO,募资4.463亿美元,定价处于区间上限。

Xi Urges AI Cooperation WIth US

Donald Trump and Xi Jinping addressed their intense competition over artificial intelligence on Thursday during a summit aimed at keeping ties between the world’s largest economies on track. (Source: Bloomberg)

中文摘要 特朗普与习近平周四在旨在稳定美中等最大经济体关系的峰会上,就人工智能领域的激烈竞争分别发表讲话。

X-planes: Are they needed in the new era of drones?

Where do America's X-plane projects fit in an era of cheap drones?

中文摘要 BBC探讨在美国廉价无人机时代背景下,军方X-plane(试验性飞行器)研发项目是否仍有必要继续推进。

Brexit and Trump’s tariffs push UK’s food and drink trade deficit to £21bn

In the first half of 2026, the sector’s export volumes fell 11.7% to hit their third-lowest level on record

中文摘要 英国脱欧及特朗普关税推动英国食品饮料贸易逆差扩大至210亿英镑,2026年上半年该行业出口量下降11.7%,为史上第三低水平。

'We're all broke': Would you chase a friend for £5?

What's the smallest amount of money you would ask a friend to pay you back?

中文摘要 BBC调查英国民众在「大家都没钱」的情况下,最少会向朋友追讨多少借款,探讨小额借款追讨行为的社会现象。

De olho nas eleições no Brasil

O repórter Daniel Carvalho conta os últimos detalhes dos candidatos à Presidência antes das eleições. Com produção de Patricia Xavier e edição de Raphael Almeida e Mariana Lumy. (Source: Bloomberg)

中文摘要 记者Daniel Carvalho报道巴西总统大选前候选人的最新动态,由Patricia Xavier制作,Raphael Almeida及Mariana Lumy剪辑。

Bond Losses Build on Inflation, Rate-Hike Bets: Markets Wrap

A global bond selloff extended into Asia as elevated oil prices stoked inflation concerns, pushing long-term Treasury yields to multi-decade highs and fueling expectations for further interest-rate hikes by the Federal Reserve.

中文摘要 全球债券抛售蔓延至亚洲,油价高企加剧通胀担忧,长期国债收益率升至数十年高位,市场押注美联储进一步加息。

AI 量化

7 回复 · 程序员 节点

该源今日无内容。