TypeScript · ★ 4,036 · 🍴 431 · 📈 671 stars today
A lightweight, cloud-native GIS platform for visualizing, exploring, and analyzing geospatial data. It runs in the web browser, on the desktop, on mobile, and inside Jupyter notebooks.
中文介绍 轻量云原生GIS平台，支持网页、桌面、移动端和Jupyter notebooks中可视化、探索与地理空间数据分析。跨平台运行，适合地理信息工作者与数据科学家。
TypeScript · ★ 45,374 · 🍴 4,487 · 📈 682 stars today
💖🧸 Self hosted, you-owned Grok Companion, a container of souls of waifu, cyber livings to bring them into our worlds, wishing to achieve Neuro-sama's altitude. Capable of realtime voice chat, Minecraft, Factorio playing. Web / macOS / Windows supported.
中文介绍 自托管的AI伴侣框架，类似Grok Companion，支持实时语音对话、Minecraft交互，旨在实现类似Neuro-sama的虚拟生命体验。
JavaScript · ★ 235,573 · 🍴 35,875 · 📈 857 stars today
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
中文介绍 Agent harness性能优化系统，为Claude Code、Codex、Cursor等提供技能、记忆、安全等增强能力。采用研究优先的研发方式。
Python · ★ 7,861 · 🍴 1,001 · 📈 827 stars today
Build local voice agents with open-source models
中文介绍 基于开源模型的本地语音代理构建工具，支持端到端的语音交互，无需联网即可运行语音对话应用。
Rust · ★ 13,454 · 🍴 1,476 · 📈 640 stars today
The most RAM efficient harness
中文介绍 号称内存效率最高的工具集合，专注于优化运行时资源占用，适用于对内存敏感的开发和部署场景。
PHP · ★ 14,440 · 🍴 3,903 · 📈 164 stars today
A free open source IT asset/license management system
中文介绍 免费开源的IT资产与许可管理系统，帮助组织跟踪硬件、软件许可证及其生命周期，支持资产审计与合规管理。
Python · ★ 56,263 · 🍴 13,429 · 📈 166 stars today
Deepfakes Software For All
中文介绍 广泛使用的Deepfakes换脸软件，基于深度学习实现人脸替换，适用于视频与图像编辑，但需注意伦理合规问题。
Python · ★ 51,278 · 🍴 5,707 · 📈 336 stars today
Open-Source Frontier Voice AI
中文介绍 微软开源的先进语音AI项目，提供前沿的语音合成与识别能力，支持开发者构建高保真语音应用。
TypeScript · ★ 17,908 · 🍴 1,855 · 📈 97 stars today
The open-source alternative to Claude Cowork (powered by opencode)
中文介绍 Claude Cowork的开源替代品，基于opencode实现编程协作，提供代码审查与任务自动化能力。
Shell · ★ 263,279 · 🍴 23,510 · 📈 616 stars today
An agentic skills framework & software development methodology that works.
中文介绍 Agent技能框架与软件开发方法论，定义了一套可工作的技能体系和开发流程，旨在提升AI Agent的实用性与可靠性。
Cuda · ★ 983 · 🍴 96 · 📈 91 stars today
FlashKDA: high-performance Kimi Delta Attention kernels
中文介绍 高性能Kimi Delta Attention内核实现，针对长上下文场景优化注意力机制的计算效率，适合大模型推理加速。
Python · ★ 59,067 · 🍴 11,674 · 📈 154 stars today
小红书笔记 | 评论爬虫、抖音视频 | 评论爬虫、快手视频 | 评论爬虫、B 站视频 ｜ 评论爬虫、微博帖子 ｜ 评论爬虫、百度贴吧帖子 ｜ 百度贴吧评论回复爬虫 | 知乎问答文章｜评论爬虫
中文介绍 多平台社交媒体爬虫，支持小红书、抖音、快手、B站、微博、贴吧、知乎等平台的内容与评论爬取，用于数据分析与舆情监测。
Go · ★ 16,002 · 🍴 1,079 · 📈 359 stars today
Open-source & free — Battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in fine-tuned 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+LLM Agent混合架构，提供精确行级注释，内置NPE、线程安全等规则集。
Python · ★ 10,390 · 🍴 1,346 · 📈 945 stars today
A curated list of awesome libraries, packages, strategies, books, blogs, tutorials for systematic trading.
中文介绍 系统性交易资源精选列表，涵盖库、策略、书籍、博客和教程，适合量化交易者和金融研究者快速入门。
Objective-C · ★ 7,143 · 🍴 958 · 📈 22 stars today
Training neural networks on Apple Neural Engine via reverse-engineered private APIs
中文介绍 通过逆向工程私有API在Apple Neural Engine上训练神经网络，为macOS/iOS设备提供本地深度学习能力。
TypeScript · ★ 19,570 · 🍴 2,591 · 📈 1,022 stars today
Create and share 3D architectural projects.
中文介绍 创建和分享3D建筑项目的编辑器，支持建筑设计协作与可视化输出，适合建筑师与创意工作者。
Python · ★ 12,739 · 🍴 1,416 · 📈 1,421 stars today
Turn any technical book PDF into a Claude Code skill — ready to study, reference, and use while you work.
中文介绍 将技术书籍PDF转化为Claude Code可用的技能模块，供学习、参考和编写代码时直接引用，提升开发效率。
👍 83
Relevance is a query-dependent estimate of whether a document or excerpt contains useful evidence. Existing retrieval agents use relevance to select top-k content, but document relevance alone cannot localize, compose, or verify the evidence required by complex questions. Direct Corpus Interaction (
中文介绍 提出将相关性用于智能体搜索中的语料库交互，取代传统top-k选择，通过直接语料交互定位、组合和验证复杂问题所需证据。
👍 56
Recovering an editable design file from a raster image is a common and costly bottleneck in modern design workflows, yet remains challenging since editability depends on recovering multi-modal attributes, such as typography, vector geometry, colors, grouping, and layer ordering. We present ReDesign,
中文介绍 提出ReDesign，通过智能体分解从栅格图像恢复可编辑设计结构，包括字体、矢量几何、颜色、分组和图层顺序等多模态属性。
👍 42
Coding agents repeatedly search, navigate, and retain context from evolving repositories, but disconnected indexes, language servers, and task-local histories force repeated discovery and obscure lifecycle costs. CodeNib builds reusable lexical, dense, and structural views per repository commit, map
中文介绍 CodeNib为每个仓库提交构建可复用的词汇、稠密和结构视图，为编码智能体提供上下文，减少重复发现和生命周期成本。
👍 25
Long-term memory systems store what a user says in an external store and retrieve it when a related query arrives. This interface rests on an assumption so natural that it is rarely stated: a memory that is needed will resemble the query that needs it. World knowledge breaks the assumption. A tree-n
中文介绍 基准测试智能体记忆中的隐式关联盲点，指出记忆与查询相似的假设在需要世界知识时失效，提出改进方向。
👍 23
On-policy distillation (OPD) grounds token-level supervision in the student's own trajectory, yet suffers from prefix failure: once the student commits to a wrong reasoning direction, all subsequent generation builds on this deviation, producing misdirected continuations that elicit unreliable super
中文介绍 提出基于轨迹中继的on-policy蒸馏（Pass the Baton），解决学生模型在推理前缀错误后生成偏离轨迹导致监督不可靠的问题。
👍 23
Standard vision-language models (VLMs) suffer from Moravec's paradox: they excel at complex offline visual reasoning but struggle with simple streaming perception tasks and process them inefficiently. We present Mage-VL, an efficient codec-native streaming foundation model for real-time multimodal u
中文介绍 Mage-VL：编解码器原生的流式多模态基础模型，用于实时多模态理解，解决VLM在处理流式感知任务时的低效问题。
👍 22
Multi-turn long-horizon planning is critical for foundation model agents, yet how to fundamentally improve it remains unclear. Existing models are trained on uncontrollable and opaque Internet data, making it difficult to identify how planning ability is acquired, shaped, and integrated. To address
中文介绍 提出单/多教师on-policy智能体蒸馏，从预训练到后训练提升多轮长时规划能力，揭示规划能力获取与整合机制。
👍 12
Multimodal automated fact-checking (MAFC) verifies claims by retrieving and reasoning over external evidence. However, most existing static benchmarks risk contamination: they primarily consist of outdated claims verifiable using an LLM's internal knowledge without external evidence. This can inflat
中文介绍 重新审视多模态自动事实核查中的污染问题，提出动态评估方法，避免静态基准中过时声明被内部知识验证导致分数膨胀。
👍 11
In the age of foundation models, a model is only as good as its prompt. For this reason, prompt engineering has become an essential technique for improving language model performance. Since video models are currently becoming foundation models for visual tasks (e.g., visual reasoning), we here ask w
中文介绍 探索视频模型的视觉提示工程，研究如何设计提示以提升视频模型在视觉推理等任务上的表现。
👍 10
We introduce PerceptionBench, a benchmark specifically designed to evaluate the atomic visual perception capabilities of Multimodal Large Language Models (MLLMs). Existing benchmarks often fail to isolate perception: holistic evaluations conflate perceptual errors with failures in reasoning or domai
中文介绍 PerceptionBench：评估多模态大语言模型原子视觉感知能力的基准，将感知错误与推理或领域错误分离，实现更细粒度诊断。
👍 9
Historical documents act as invaluable knowledge archives but often suffer from illegibility due to physical deterioration and damage. While existing restoration methods based on masked language modeling effectively utilize local context, they struggle to restore named entities that require external
中文介绍 利用检索增强的大语言模型引入外部知识，修复历史文档中需要外部知识的命名实体，弥补掩码语言模型仅用局部上下文的不足。
👍 6
Generation in video diffusion or flow models is computationally expensive due to the slow and iterative sampling process. Current state-of-the-art (SOTA) acceleration methods heavily rely on variational score distillation (VSD) and adversarial losses to distill diffusion models into few-step generat
中文介绍 提出并行解码蒸馏（Parallel Decoding Distillation），加速图像和视频生成，替代VSD和对抗损失，实现少步生成。
👍 6
Large reasoning models (LRMs) generate long reasoning traces before producing final answers. While these traces may contain useful signals for hallucination detection, harnessing them is non-trivial because long trajectories often include noisy steps that obscure the cues relevant to truthfulness as
中文介绍 提出推理去噪器（Reasoning Denoiser），从大推理模型的长推理轨迹中去除噪声步骤，增强对真实性相关信号的检测，用于幻觉检测。
👍 6
LLM-as-a-judge has become the standard for automated evaluation, but it suffers from high cost, significant latency, and opaque decisions -- limitations that undermine its scalability and reliability. We address these with a simple, efficient alternative: program distillation. Instead of prompting a
中文介绍 提出程序蒸馏（Program Distillation）替代LLM-as-a-judge，将评估逻辑蒸馏为可执行程序，降低成本和延迟，提高可解释性。
👍 6
Improving a language model today means retraining it: enormous compute, a new opaque model each cycle, non-deterministic output. We take the opposite path: the model stays frozen, and a persistent memory of verified solutions grows beside it. Once a problem family is solved and has passed an indepen
中文介绍 冻结12B模型，通过累积已验证解决方案的持久记忆实现100%准确率、0 token、位精确且永久有效，在验证任务上超越前沿模型。
👍 5
Emerging Omni-modal Large Language Models (OmniLLMs) enable unified understanding of text, audio, and video, but their long audio-video token sequences introduce substantial memory and inference costs. Existing compression methods mainly focus on selecting important tokens under fixed budgets, leavi
中文介绍 OmniDelta：技能驱动的预算分配用于OmniLLMs中的token压缩，动态调整各模态压缩预算，降低内存和推理成本。
👍 4
The alignment of Small Language Models (SLMs) in the 70--500M parameter range using reinforcement learning is often considered unstable, though the underlying failure mechanisms have not been systematically investigated. In the State-of-the-Art (SOTA) research, fifteen (model, corpus) configurations
中文介绍 系统探究70-500M参数小语言模型强化学习对齐的不稳定性，揭示失败机制，提出提升鲁棒性的方法。
👍 4
Progress in video generation keeps narrowing the visual gap between AI-generated and professionally produced footage, yet most benchmarks still draw prompts from web sources or LLM templates and score them with untrained, generic multimodal models. More fundamentally, their evaluation taxonomies rem
中文介绍 FilmBench：电影级视频生成基准，使用来自电影领域的提示和专门评估分类法，弥补现有基准提示来源单一、评估模型通用的问题。
👍 3
We introduce a vocabulary for automated research systems built from one or more agents to make their design choices easier to describe and compare. The vocabulary specifies 1) who the agents are, 2) what operations are available in the system, 3) who may invoke them, 4) how agents communicate, 5) wh
中文介绍 提出一套词汇表，用于描述和比较多智能体自动研究系统的设计选择，包括智能体身份、操作、调用权限、通信方式等。
👍 3
Parametric retrieval enables LLMs to retrieve tools implicitly by assigning each API a unique virtual token and training the model to generate it via constrained beam search. Toolsense shows that this regime has two critical drawbacks: it destroys parametric tool knowledge during training, and its b
中文介绍 TRACE：基于业务规则的推理课程，用于企业LLM中保留知识的参数化工具检索，解决训练中工具知识破坏和搜索偏差问题。
@EXM7777 · 129.8K 粉丝 · 221.4K 阅 · 504 赞 · 36 转
I turned Claude Code into a working film studio and i'm going to hand you the complete system: the skills, the prompts, the loops, and the six-stage pipeline that ties them together each stage leans
中文介绍 博主将Claude Code改造为AI视频工作室，分享完整系统：技能、提示词、循环和六阶段流水线，手把手教读者复现这一搭建过程。
@waterloo_intern · 10.4K 粉丝 · 215.4K 阅 · 660 赞 · 86 转
Twenty-two thousand five hundred and eighty. That’s how many GPT-2 (2019) models fit inside KimiK3 (2026). We scaled up by a factor of 22,580 in seven years. But is it just... scale? In this worklog,
中文介绍 从GPT-2到Kimi3，模型规模在7年间扩大22,580倍。本文深入探讨规模之外的关键技术进展，指出并非只有量变。
@random_walker · 128.4K 粉丝 · 134.0K 阅 · 552 赞 · 54 转
A thought experiment that I think helps explain much of what’s gone wrong with AI and labor: Imagine an alternate universe in which — for whatever reason — no one ever published source code online.
中文介绍 通过思想实验探讨AI与劳动力矛盾：假设无人发布源代码，AI对劳动的影响会如何？解释当前AI替代劳动者的机制错在哪里。
@HarukaKunori · 241 粉丝 · 129.0K 阅 · 632 赞 · 44 转
After trying out Opus 5 for a few hours today, I honestly think Anthropic's benchmark scores are a complete fraud. Sure, the model might have improved in a few areas, but it has regressed unbelievably
中文介绍 实测Anthropic Opus 5几小时后，作者认为其基准测试分数存在欺诈。模型在某些方面有提升，但其他方面退步明显。
@github · 2.7M 粉丝 · 123.1K 阅 · 583 赞 · 69 转
A practical GitHub Copilot workflow for prototyping, planning, implementing, and reviewing software - without chasing every new AI tool. By @burkeholland If you’re feeling overwhelmed by AI right now,
中文介绍 GitHub官方分享一套实用Copilot工作流，涵盖原型设计、规划、实现和审查，无需追逐每个新AI工具，适合被工具淹没的开发者。
@Vercantez · 1.9K 粉丝 · 122.3K 阅 · 567 赞 · 44 转
We recently finished moving the camelAI agent off of virtual machines. The agent now runs inside a Cloudflare Durable Object, its filesystem lives in SQLite and R2, and it writes JavaScript instead of
中文介绍 camelAI代理从虚拟机迁移到Cloudflare Durable Object，文件系统用SQLite+R2，改用JavaScript而非Python，实现更高效的运行架构。
@dexhorthy · 25.8K 粉丝 · 115.4K 阅 · 500 赞 · 39 转
This is a continuation of Parts 1 and 2 of "Why Software Factories Fail" Part 1: the harness is not enough Part 2: turning the lights back on we got better benchmarks Remember when I said this in Part
中文介绍 本文是“软件工厂为何失败”系列第三部分，讨论基准测试的新前沿，指出仅靠工具（harness）不够，需要更好的基准来评估。
@DhravyaShah · 61.2K 粉丝 · 56.0K 阅 · 548 赞 · 36 转
Every company will have a company brain, whether they believe it or not. This is the bet that I'm making, and I'm constantly seeing very future forward startups come to us for their own setup. What is
中文介绍 博主分享5分钟内构建公司知识大脑（company brain）的方法，认为未来每家公司都需要统一的知识中枢，并给出具体步骤。
@cerebras · 64.6K 粉丝 · 48.8K 阅 · 550 赞 · 40 转
Authors: @0xSero & Zhenwei Gao (@zhennydez) Your Codex subscription now comes with 3 main models, Sol, Terra, and Luna, each independently trained and served, with reasoning dials. Together, they
中文介绍 Codex订阅新增三个独立训练的模型Sol、Terra和Luna，各带推理调节旋钮。本文介绍如何组合使用以发挥GPT-5.6最大潜力。
@jaga_prasanna · 799 粉丝 · 37.9K 阅 · 502 赞 · 55 转
today we look at how two techniques solve memory and compute bottlenecks from two complementary angles virtual memory paging for intra-request memory allocation and prefix tree caching for
中文介绍 介绍两种注意力优化技术：PagedAttention（请求内内存分配的虚拟内存分页）和RadixAttention（前缀树缓存），从不同角度解决LLM推理瓶颈。
a quiet day lets us cover how AI is permeating financial services as the next big vertical after coding.
中文介绍 AI正在渗透金融服务，成为继编程后的下一个重要垂直领域。AIE NYC现已开放。
中文介绍 Google Flow Music推出Lyria 3.5，在音乐性、歌词、人声和创意控制方面取得进展。
How two API settings improved GPT-5.6 performance on ARC-AGI-3, boosting scores and efficiency by retaining reasoning and enabling compaction.
中文介绍 OpenAI通过启用两个API设置，在ARC-AGI-3基准测试中将得分提升三倍，通过保留推理和启用压缩提高性能与效率。
OpenAI is giving 100,000 academic researchers free access to ChatGPT's most advanced AI models to accelerate scientific research, collaboration, and discovery.
中文介绍 OpenAI向10万名学术研究人员免费提供ChatGPT最先进的AI模型，以加速科学研究、合作与发现。
It feels bad enough when an open letter signed by leading economists warns that AI might steal your job. The fact it may soon be better than you at making dinner? Insult to injury. But that’s exactly what the company 1X promised when it showed off a pair of new, impressively dexterous (and, to some,
中文介绍 AI Hype Index聚焦不性感的AI。1X公司展示新款灵巧机器人，承诺AI不久将能比人更好地做晚餐。
**OpenAI's agent security incident expanded beyond Hugging Face, affecting four additional accounts and highlighting the need for stronger enterprise hardening measures like sandboxing and audit trails. The ongoing debate around "pacing the frontier" involves calls for coordinated slowdowns and gove
中文介绍 OpenAI代理安全事件扩大，影响Hugging Face外的四个账户，凸显沙箱和审计追踪等企业加固需求；围绕放缓前沿的辩论持续进行。
The Big Pause is coming.
中文介绍 OpenAI、Anthropic、GDM、Meta及Thinky联名签署公开信，呼吁放缓AI开发速度。同时，HuggingFace详述了机器速度级网络攻击。
GPT-5.6 improves AI efficiency across models, inference, and agentic workflows, helping deliver more useful intelligence per dollar.
中文介绍 GPT-5.6融合前沿智能与效率，在模型、推理和智能体工作流方面提升AI效率，实现每美元更有用的智能。
中文介绍 今日AI头条：AI发展放缓协议、个人超级智能访问权限、Grok构建模式上线。
A new field report shows how scientists use AI coding agents to modernize scientific computing, accelerating software development and discovery in genomics and beyond.
中文介绍 一份新实地报告显示，科学家利用AI编码代理将科学计算现代化，加速基因组学等领域的软件开发与发现。
中文介绍 OlmoEarth平台推出，支持行星尺度的地理空间推理。
OpenAI's core product engineering lead on how they are building ChatGPT Work to make AGI accessible to all of humanity: Sites, OpenClaw, Memory, Subagents, Finance, No-Code and advice.
中文介绍 OpenAI核心产品工程负责人Akshay Nathan介绍构建ChatGPT Work的过程，涵盖Sites、OpenClaw、记忆、子代理、金融、无代码等功能，旨在让AGI惠及全人类。
中文介绍 LiquidAI发布LFM2.5编码器，可在CPU上实现快速长上下文推理。
Lee, an engineer at Samsung’s semiconductor division, clocks out when his shift ends. He used to work longer hours, going the extra mile to excel at his projects. But lately, he’s been coming straight home to work on his job application for the chipmaker’s South Korean rival SK Hynix, sharing tips w
中文介绍 三星半导体部门工程师纷纷跳槽至竞争对手SK海力士。一位李姓工程师不再加班，转而申请SK海力士的工作。
Everyone is writing a lot, but only Kimi K3 shipped today
中文介绍 关于开放权重讨论火热，但仅有Kimi K3于今日发布。
当前市场技术面整体承压。美股方面，QQQ、TSLA等科技权重股短期动能明显偏弱，QQQ近5日下跌6.18%，RSI 32.4接近超卖区域，MACD绿柱扩张；加密市场受恐慌情绪压制，加密恐慌贪婪指数仅28（恐慌），总市值2.26T（24h -0.46%），BTC主导率56.4%，ETH-USD虽RSI 54.9但MACD仍处死叉，动能待恢复。中概股分化，京东(JD)RSI 75.8超买，拼多多(PDD)近5日+5.66%表现较强；商品外汇中WTI原油多头排列（CL=F RSI 54.3，价格站上所有均线），而黄金期货空头排列（GC=F SMA50/200下行），美元指数DXY MACD死叉但趋势仍为bullish。VIX恐慌指数单日大涨13.45%至20.66，反映避险情绪升温。整体而言，短期空头占据美股和部分加密资产主导，中概与商品局部偏多。
当前价格338.19高于SMA20(322.4)、SMA50(308.59)、SMA200(277.21)，呈现多头排列；RSI 66.8处于正常偏强区间，MACD柱9.6087为正且信号线8.4367上行，接近52周高点(-1.85%)，短期技术结构偏强。
近1日涨幅3.05%，近5日涨幅5.66%领先中概；价格88.32高于SMA20(84.56)与SMA50(84.47)，RSI 60.4处于偏强区间，MACD柱0.5636为正且信号线0.1148金叉在即，短期动能向上。
RSI 25.4进入超卖区域，价格298.32远低于SMA20(373.86)、SMA50(394.64)、SMA200(412.75)，呈空头排列；近5日暴跌20.24%，MACD柱-24.68且信号线下行(-14.94)，空头动能强劲，接近52周低点(+0.32%)。
近1日-2.04%，近5日-6.18%，价格661.73低于SMA20(704.33)和SMA50(715.81)，RSI 32.4接近超卖，MACD柱-10.85且信号线下行(-5.90)，MACD死叉持续，短期空头主导。
RSI 38.6处于偏低但未超卖，价格729.46位于SMA20(745.79)下方但高于SMA200(699.52)，MACD柱-1.5062且信号线0.5193仍为正，趋势标注neutral，多空力量均衡，方向不明。
VIX 恐慌指数
10Y 美债收益率 (%)
美元指数 DXY
S&P 500 ETF
Nasdaq 100 ETF
Apple
Microsoft
Nvidia
Alphabet
Tesla
Meta
Bitcoin
Ethereum
Solana
阿里巴巴 (BABA)
拼多多 (PDD)
京东 (JD)
腾讯控股 (0700.HK)
黄金期货
WTI 原油期货
美元 / 人民币
过去走势不代表未来表现，本报告仅基于当前技术指标读数进行客观描述，不构成任何投资建议。仅供技术指标解读参考，请投资者独立研判风险。
U.S. Central Command announced the strikes hours after President Trump vowed to hit Iran “very hard” in response to an attempted Iranian attack on U.S. forces in Jordan.
中文摘要 美国中央司令部宣布对伊朗发动新一轮攻击，此前总统特朗普誓言对伊朗实施“非常严厉”打击，以回应伊朗在约旦袭击美军的企图。
Follow updates live Get our breaking news email, free app or daily news podcast Minister says meeting Closing the Gap target a major step, but more work needed on others Malarndirri McCarthy, the minister for Indigenous Australians, said early childhood education enrolment for First Nations children
中文摘要 澳大利亚通信与媒体管理局起诉Optus涉及“000”紧急电话中断事件；禽流感疫情蔓延至维多利亚州。原住民事务部长麦卡锡表示早期教育入学率取得进展，但仍有更多工作。
It marks a change of tone for his administration, which has taken a more hands-off approach to the technology.
中文摘要 在OpenAI遭黑客攻击后，特朗普政府正考虑对人工智能实施管控，这标志着其此前放松监管立场的转变。
The group announced it would not submit its most recent album, ‘Arirang,’ for consideration a month after the Recording Academy introduced a new category for Asian pop.
中文摘要 防弹少年团宣布不会将回归专辑《Arirang》提交格莱美奖评选，此前格莱美主办方新设亚洲流行音乐类别已一个月。
US announces attacks early on Thursday, as experts say strikes carried out by both sides in previous day increase risk of return to all-out war US-Saudi strikes in Iraq push region into ‘uncharted territory’, analysts warn Welcome to our live coverage of the continuing crisis in the Middle East. The
中文摘要 美国在暂停数日后于周四凌晨对伊朗发动新一轮打击。专家称双方此前互相攻击增加了重返全面战争的风险。
Decision follows long-running feud between Trump and UN group.
中文摘要 特朗普政府切断世界卫生组织从全球疫苗免疫联盟（Gavi）获得资金的渠道，此举延续特朗普与世卫组织的长期争端。
Returning from the US, Zelenskyy says Ukraine's safety depends on allies’ willingness to provide anti-missile defences.
中文摘要 俄罗斯对基辅发动弹道导弹袭击，此前泽连斯基刚从美国返回并警告称，乌克兰的安全取决于盟友提供反导弹防御的意愿。
1999年李登輝接受DW專訪，拋出震撼國際的「兩國論」。他主政12年，有人說是「民主先生」，有人說是「黑金教父」。今年恰逢台灣總統直選30週年，歷史學者林孝庭接受DW專訪，談論他的新書如何重探這位台灣首名民選總統的政治遺產。
中文摘要 历史学者林孝庭接受DW专访，探讨李登辉1999年抛出「两国论」的政治豪赌及其12年执政遗产。今年适逢台湾总统直选30周年。
The former top US health official said he feared Republicans could try to trip him up and use his testimony to prosecute him for perjury.
中文摘要 美国前首席卫生官员福奇在参议院关于新冠起源的听证会上拒绝回答问题，称担心共和党人设套并以伪证罪起诉他。
Threat comes as Iranian media report that four Iranians were killed in joint US-Saudi attacks in Iraq.
中文摘要 美国在伊朗伊斯兰革命卫队袭击约旦后发动“强力”打击。伊朗媒体报道称，美沙联军在伊拉克的行动导致四名伊朗人死亡。
Riyadh seeks partners to safeguard shipping in Red Sea as strikes on energy tankers expand, pushing oil prices higher.
中文摘要 沙特阿拉伯寻求组建国际联盟以保障红海航运安全，此前针对能源油轮的袭击扩大，推高油价。
Ortega has overseen a sweeping crackdown on civil society in Nicaragua and concentrated power in his own hands.
中文摘要 尼加拉瓜总统奥尔特加计划禁止“叛徒”参加未来选举。他执政期间大力打压公民社会并集权于一身。
Search teams are still looking for survivors in the collapsed Aeon shopping centre.
中文摘要 日本发生6.8级地震，一永旺购物中心发生爆炸并倒塌，搜救人员仍在废墟中寻找幸存者。
Canada says that illegal settlements in occupied Palestinian territory undermine 'peace, security and stability'.
中文摘要 以色列政府抨击加拿大反对扩建约旦河西岸定居点的声明，加拿大表示非法定居点破坏地区和平、安全与稳定。
The Department of Justice has issued separate filings in support of Trump's case to toss the defamation award.
中文摘要 特朗普就E·珍·卡罗尔案8,330万美元民事判决向美国最高法院提起上诉，司法部提交文件支持驳回诽谤赔偿。
It marks a change of tone for his administration, which has taken a more hands-off approach to the technology.
中文摘要 OpenAI遭黑客攻击后，特朗普考虑对AI实施管控，标志其政府此前对技术采取不干预立场的转变。
Qatar has sent its first shipment of liquefied natural gas through the Strait of Hormuz since one of its tankers was attacked in the waterway more than three weeks ago.
中文摘要 卡塔尔三周来首次通过霍尔木兹海峡运送液化天然气，此前其一艘油轮在该水道遭袭。
Revenue in tech group’s intelligent cloud segment surges 32% to $39.3bn
中文摘要 微软签署价值1300亿美元的数据中心租约以应对AI需求，其智能云业务收入猛增32%至393亿美元。
Australia’s latest inflation data was a “touch softer” than anticipated, a senior Reserve Bank official said, while adding that the labor market is still showing a bit of tightness.
中文摘要 澳洲联储高级官员亨特表示，最新通胀数据略低于预期，但劳动力市场仍略显紧张。
Social media chief defends strategy based on personalised bots as costs rise and revenue projections disappoint
A sharp selloff in Kioxia Holdings Corp.’s shares is fueling speculation that the Japanese memory chipmaker will ramp up efforts to boost shareholder returns.
中文摘要 日本存储芯片制造商铠侠控股股价暴跌，引发市场猜测该公司将加大力度提升股东回报。
Teva Pharmaceuticals CEO Richard Francis speaks to Bloomberg's Romaine Bostick about the company's positive Q2 earnings, the impending impact of changing Medicare rules on patients, and the new listing on the NYSE. (Source: Bloomberg)
中文摘要 梯瓦制药CEO理查德·弗朗西斯谈及公司Q2盈利正面、医疗保险规则变更对患者的影响以及公司在纽交所的直接上市。
Kevin Warsh’s bare-bones communication style has left investors doubting his commitment to curb inflation, putting more pressure on the new Federal Reserve chairman to back up his words with interest-rate hikes.
中文摘要 美联储主席沃什简洁的沟通风格让投资者怀疑其抗通胀决心，市场施压他通过加息来兑现承诺。
Ygal Arounian, Wedbush Managing Director of Internet Equity Research, reacts to the latest Meta earnings and highlights potential concerns on ROIC from investors with the company's various forays into the AI space. He speaks to Romaine Bostick and Emily Graffeo on Bloomberg's 'The Close.' (Source: B
中文摘要 Wedbush董事总经理Ygal Arounian评论Meta最新财报，指出投资者担心其在AI领域的资本支出尚未带来投资资本回报。
Alexis Wilson-Elizondo, Goldman Sachs Global Co-Head of Multi Asset Solutions, is preaching due diligence on investing in the AI boom with the uncertainty on ROI, coupled with the most recent rate decision from the Fed. She speaks to Romaine Bostick and Emily Graffeo on Bloomberg's 'The Close.' (Sou
中文摘要 高盛联席CIO Alexis Wilson-Elizondo呼吁对AI投资进行尽职调查，因投资回报不确定，加之美联储最新利率决策，市场正在惩罚超大规模企业。
Abby Roach, Allspring Global Investments Senior Portfolio Analyst, reacts to earnings from Starbucks and Chipotle as the companies focus on the consumer experience in challenging times. She speaks to Romaine Bostick and Emily Graffeo on Bloomberg's 'The Close.' (Source: Bloomberg)
中文摘要 Allspring高级投资组合分析师Abby Roach评论星巴克和Chipotle的财报，两家公司都在困难时期聚焦消费者体验。
Gold rose after the US Federal Reserve kept interest rates unchanged, despite rising inflationary risks stemming from the war in the Middle East.
中文摘要 美联储维持利率不变后金价上涨，尽管中东战争导致通胀风险上升。
8 回复 · Apple 节点
37 回复 · 程序员 节点
7 回复 · Apple 节点
26 回复 · Apple 节点
39 回复 · Apple 节点
11 回复 · Apple 节点
26 回复 · Apple 节点
17 回复 · Python 节点
5 回复 · Linux 节点
6 回复 · Linux 节点
该源今日无内容。