Python · ★ 66,173 · 🍴 9,587 · 📈 4,685 stars today
利用AI大模型，一键生成高清短视频 Generate short videos with one click using AI LLM.
中文介绍 基于 AI 大模型的一键短视频生成工具，用户只需提供主题或关键词，系统自动完成文案撰写、语音合成、字幕生成和视频拼接，输出高清短视频。适合自媒体创作者、营销人员快速批量生产内容，大幅降低视频制作门槛和时间成本。
JavaScript · ★ 197,227 · 🍴 30,330 · 📈 1,388 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 面向 Claude Code、Codex、Opencode、Cursor 等 AI 编程工具的 Agent 性能优化框架，提供技能（Skills）、本能（Instincts）、记忆（Memory）、安全机制及研究优先的开发方法论，帮助开发者系统性地提升 AI 编程 Agent 的执行效率与可靠性。
Shell · ★ 26,320 · 🍴 1,972 · 📈 2,235 stars today
Taste-Skill - gives your AI good taste. stops the AI from generating boring, generic slop
中文介绍 为 AI 编程助手注入「审美品味」的 Skill 文件，通过预设风格约束规则，阻止模型生成平庸、千篇一律的样板代码或文本，适合希望 AI 输出更具个性、更有质感内容的开发者和写作者使用。
★ 6,387 · 🍴 468 · 📈 755 stars today
A skill file for removing AI tells from prose
中文介绍 一个针对 AI 写作痕迹的清除 Skill 文件，通过规则约束让 AI 生成的散文摆脱典型的机器腔调和套话，使输出更自然、更具人味。适合内容创作者在使用 Cursor、Claude 等工具写作时提升文本质量。
TypeScript · ★ 47,821 · 🍴 6,790 · 📈 495 stars today
The open alternative to Salesforce, designed for AI.
中文介绍 Salesforce 的开源替代方案，专为 AI 时代设计的 CRM 平台。提供客户关系管理的核心功能，同时深度集成 AI 能力，支持自托管部署。适合希望摆脱高昂 SaaS 费用、同时需要灵活定制的中小企业和技术团队使用。
HTML · ★ 170,666 · 🍴 3,291 · 📈 1,769 stars today
DigitalPlat FreeDomain: Free Domain For Everyone
中文介绍 DigitalPlat 推出的免费域名申请平台，向所有人开放，无需付费即可获得可用域名。适合个人开发者、学生或小型项目快速建立网络存在，降低建站初始成本。
★ 48,421 · 🍴 5,079 · 📈 2,015 stars today
An advanced guide to learn English which might benefit you a lot 🎉 . 离谱的英语学习指南/英语学习教程/英语学习/学英语
中文介绍 面向中文母语者的进阶英语自学指南，系统梳理听说读写各项能力的提升路径，提供可落地的学习方法和资源推荐。内容覆盖从基础到高阶的完整学习体系，适合希望通过自学突破英语瓶颈的国内开发者和学生群体。
Python · ★ 127,678 · 🍴 8,738 · 📈 1,263 stars today
Python tool for converting files and office documents to Markdown.
中文介绍 微软开源的 Python 工具，支持将 Word、Excel、PowerPoint、PDF 等多种办公文档及文件格式批量转换为 Markdown。常用于 RAG 数据预处理、文档知识库构建等场景，方便 LLM 消费结构化文本内容。
Shell · ★ 211,030 · 🍴 18,782 · 📈 1,726 stars today
An agentic skills framework & software development methodology that works.
中文介绍 一套可实际落地的 Agent 技能框架与软件开发方法论，为 AI 编程 Agent 提供结构化的能力扩展机制。帮助开发者以模块化方式定义和复用 Agent 行为，适合构建复杂自动化开发流程的工程团队。
HTML · ★ 3,869 · 🍴 580 · 📈 190 stars today
A meta-skill that designs domain-specific agent teams, defines specialized agents, and generates the skills they use.
中文介绍 一个元技能（meta-skill）框架，能够自动设计特定领域的 Agent 团队，定义各专项 Agent 的职责，并为其生成所需的技能文件。适合需要快速搭建多 Agent 协作体系的开发者，降低复杂 Agent 架构的设计成本。
Markdown · ★ 506,521 · 🍴 48,089 · 📈 866 stars today
Master programming by recreating your favorite technologies from scratch.
中文介绍 通过从零复现经典技术来掌握编程的学习资源合集，涵盖自制 Git、Docker、Redis、编译器、数据库等数十个项目。每个项目均有详细教程引导，适合希望深入理解底层原理、突破「只会调用 API」瓶颈的开发者。
TypeScript · ★ 42,670 · 🍴 3,399 · 📈 3,766 stars today
Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
中文介绍 将任意代码库转换为可交互知识图谱的工具，用户可在图谱中浏览、搜索并直接提问，核心目标是「教学性图谱」而非炫技。兼容 Claude Code、Codex、Cursor、Copilot、Gemini CLI 等主流 AI 编程工具，适合快速理解陌生大型项目。
Python · ★ 66,917 · 🍴 6,854 · 📈 253 stars today
🚀🤖 Crawl4AI: Open-source LLM Friendly Web Crawler & Scraper. Don't be shy, join here: https://discord.gg/jP8KfhDhyN
中文介绍 面向 LLM 应用的开源网页爬虫与抓取框架，输出格式针对大模型消费做了优化，支持异步高并发抓取。适合需要构建 RAG 知识库、实时联网检索或大规模数据采集的 AI 应用开发者，提供比通用爬虫更友好的结构化输出。
Python · ★ 2,215 · 🍴 210 · 📈 53 stars today
MOSS‑TTS Family is an open‑source speech and sound generation model family from MOSI.AI and the OpenMOSS team. It is designed for high‑fidelity, high‑expressiveness, and complex real‑world scenarios, covering stable long‑form speech, multi‑speaker dialogue, voice/character design, environmental soun
中文介绍 来自 MOSI.AI 与 OpenMOSS 团队的开源语音及音效生成模型家族，专注于高保真、高表现力的语音合成，针对复杂真实场景进行了优化。适合语音助手、有声内容生产、影视配音等需要自然语音生成能力的开发者和研究人员。
TypeScript · ★ 17,759 · 🍴 1,362 · 📈 180 stars today
Official Compound Engineering plugin for Claude Code, Codex, Cursor, and more
中文介绍 Every 公司推出的 Compound Engineering 官方插件，兼容 Claude Code、Codex、Cursor 等主流 AI 编程工具，将 Compound Engineering 方法论中的工程规范和最佳实践以插件形式直接集成到开发工作流中。
Python · ★ 142,784 · 🍴 16,856 · 📈 791 stars today
Public repository for Agent Skills
中文介绍 Anthropic 官方维护的 Agent Skills 公开仓库，收录可供 Claude 等 AI Agent 直接调用的技能模块，涵盖常见自动化任务场景。开发者可参考或复用这些标准化技能，快速为自己的 Agent 应用扩展能力边界。
👍 1
LLM agents increasingly act by writing code, yet a split persists between the runtime that drives the agent and the code the model writes. The runtime owns the loop, context, and control flow, and the model has little say over any of them. Letting model-written code shape the runtime itself would ma
👍 16
Despite the rapid progress of multimodal large language models in building Graphical User Interface (GUI) agents, their real-world task completion is fundamentally bottlenecked by a lack of world knowledge about GUI operations. Existing solutions typically rely on expensive multi-agent scaffolding o
👍 1
Understanding 3D objects from images is fundamental to robotics and AR/VR applications. While recent work has made progress in category-level pose estimation, current representations fail to capture the fine-grained semantics needed for reasoning about object parts, functions, and interactions. In t
👍 7
Chain-of-thought (CoT) monitoring has been proposed as a promising safety mechanism for detecting misaligned behavior in large language models. However, its reliability remains largely unexplored beyond English and across diverse model families. We present the first large-scale evaluation of CoT mon
👍 1
Large language models (LLMs) for code completion and generation are increasingly used in software development, yet they may reproduce training examples verbatim and without authorship attribution, raising legal and ethical concerns around plagiarism and license compliance. Classical fingerprint-base
👍 1
Recent studies have shown that Large Language Models (LLMs) can achieve strong reasoning performance by incorporating functional symbolic representations that abstractly describe graph traversal algorithms and step-by-step reasoning in few-shot learning settings. However, it remains unclear how LLMs
👍 0
Clark Hash is a small method for storing neural embeddings in less space. It normalizes each database vector, applies a deterministic sparse signed Johnson-Lindenstrauss projection, clips the result, and stores a fixed-width scalar-quantized code. Queries stay in floating point and are scored agains
👍 3
Existing emotional support conversation (ESC) systems mainly rely on end-to-end response generation or coarse strategy supervision, offering limited interpretability and little support for systematic skill improvement. We propose ESC-Skills, a skill-centric framework that discovers and self-evolves
👍 39
Reinforcement learning has become a central paradigm for advancing reasoning in large language models, yet most existing methods still depend on stronger teacher models or heavily curated difficult datasets, limiting scalable capability improvement. In this paper, we introduce DenoiseRL, a reinforce
👍 76
Proactive Recommender Systems (PRSs) aim to guide user preference shift toward target items by generating paths of intermediate recommendations. Reinforcement learning (RL) provides a principled framework for optimizing such sequential decision tasks, as path rewards can naturally capture both short
👍 19
Existing memory-augmented LLM agents often treat memory as a static repository with pre-defined representations and fixed retrieval pipelines, which is brittle in dynamic agentic environments where feedback, task variation, and heterogeneous signals continuously reshape what should be remembered and
👍 33
Memory is essential for enabling large language models to support long-horizon reasoning, yet existing memory systems remain unreliable and difficult to debug. Tracing memory's dynamic evolution is crucial to understand how information is synthesized, propagated, or corrupted over time. In this work
👍 0
LLM safety evaluations predominantly test models in isolation, yet deployed AI agents increasingly operate within persistent social environments alongside other agents. We introduce a Moltbook-style simulation platform where thousands of LLM agents interact across communities over a simulated month,
👍 5
Cross-view spatial reasoning remains a weak spot for vision-language models (VLMs): they often reason in language and lose the fine-grained geometry needed for the task. Thinking with images aims to address this by generating an intermediate thinking image, but recent work shows that models often ig
👍 13
We introduce GE-Sim 2.0 (Genie Envisioner World Simulator 2.0), a closed-loop video world simulator for robotic manipulation. Building on the action-conditioned video generation framework of Genie Envisioner, GE-Sim 2.0 is re-trained on thousands of hours of real-world robot data spanning teleoperat
👍 18
Agent skills provide a lightweight way to adapt LLM agents to specialized domains by storing reusable procedural knowledge in structured files. However, whether downloaded from third parties or self-generated, these skills are often unreliable, incomplete, or outdated. Existing skill-evolution metho
👍 12
Model internals encode rich information about how a large language model (LLM) processes its training data; however, post-training data engineering largely relies on external signals and ignores rich intrinsic signals lying in model internals. We propose SAERL, a data engineering framework for LLM r
👍 9
Modern Large Language Models (LLMs) are often criticized for producing repetitive and homogeneous text, despite possessing vast latent vocabularies. While previous research has focused on model knowledge and training data, we investigate the role of decoding mechanics in suppressing linguistic diver
👍 4
Multi-agent systems built on large language models (LLMs) require many coordination choices that are difficult to fix a priori: which skill protocol to invoke, which agent role should perform a subtask, which model to bind to each role, how roles should interact, when to use retrieval or verificatio
👍 15
Long-lived AI agents are increasingly deployed as persistent operational systems, yet they are still evaluated like freshly initialized models. Day-one benchmarks miss a basic systems question: how long does an agent remain reliable after deployment? Even when model weights are frozen, an agent's ef
👍 2
Autonomous computer use agents that powered by multimodal large language models (MLLMs) are emerging as capable assistants for completing complex digital workflows. However, real-world execution environments are far from ideal: pop-ups, resolution changes, and competing applications frequently inter
👍 6
Creating images from noise is image generation; reconstructing fine details from coarse inputs is super-resolution. Despite their practical differences, both can be understood as reversing information loss across scales. We introduce SKILD, a Scale-invariant K-Space Image Learning Diffusion model th
👍 18
Causal Transformer language models suffer from strictly sequential decoding and a quadratic per-step attention cost. While linear-time causal models and discrete diffusion models each address these weaknesses, their integration remains inherently inconsistent: diffusion requires bidirectional attent
👍 15
End-to-end autonomous driving via Vision-Language-Action (VLA) models demands a precarious balance between high-fidelity trajectory planning and efficient inference. Existing paradigms typically fall short: autoregressive (AR) VLAs are memory-bandwidth-bound on edge hardware and prone to exposure-bi
👍 2
Large multimodal models (LMMs) have rapidly advanced in perception and reasoning; however, it remains unclear whether these capabilities generalize to discovering visually grounded solutions in open-ended environments, beyond pattern recognition. In such settings, intelligence requires more than ans
👍 1
Referring segmentation aims to segment the target objects in images or videos based on the textual query. Despite remarkable progress over the past years, existing works always assume that the user-provided queries are already precise and clear. However, this assumption is impractical. In real-world
👍 3
Recent progress on long-horizon agentic tasks has been driven largely by scaling up individual agents through stronger models, better tools, and more effective scaffolding. In contrast, much less is understood about scaling out: whether multiple peer agents, all targeting the same task, can become a
👍 0
Large Language Models (LLMs) are predominantly governed by probabilistic frameworks in which the sum of outcome probabilities is constrained to unity. This architectural limitation, often imposed by Softmax layers, leads to a collapse of uncertainty that makes it difficult to differentiate between e
👍 0
Discrete diffusion models have emerged as powerful frameworks for generating structured categorical data. However, efficiently sampling from reward-tilted distributions remains a fundamental challenge. While Twisted Sequential Monte Carlo (SMC) offers asymptotic exactness for this task, estimating t
👍 9
To enable reliable long-term interaction, LLM agents require a memory system that can faithfully store, efficiently retrieve, and deeply reason over accumulated dialogue history. Most existing methods adopt an extracted fact based paradigm: handcrafted static prompts compress raw dialogues into atom
该源今日无内容。
80% Devin Commits, Spec-to-PR Workflows, Full VMs, Agent Memory, and PMs Shipping Code
中文介绍 Cognition联合创始人Walden Yan与OpenInspect的Cole Murray探讨异步AI智能体时代，介绍Devin的80%代码提交率、从需求说明到PR的自动化工作流、完整虚拟机环境、智能体记忆机制，以及产品经理直接参与代码交付的新模式。
It is one thing to say AI will change the world. It is another to expect the class of 2026 to applaud it. In fact, when former Google CEO Eric Schmidt told University of Arizona graduates that their task is to help shape AI, he was met with a resounding chorus of boos. “I can…
中文介绍 前谷歌CEO埃里克·施密特在亚利桑那大学毕业典礼上呼吁2026届毕业生参与塑造AI未来，却遭到学生集体嘘声。此事件成为MIT《技术评论》"AI炒作指数"的典型案例，折射出公众对AI发展叙事的真实态度与情绪反弹。
coding is an uncapped TAM market
中文介绍 AI编程公司Cognition完成D轮融资，融资金额达10亿美元，估值攀升至260亿美元。报道认为代码编写市场属于上限极高的赛道，该轮融资凸显资本市场对AI原生软件开发工具的高度看好。
Explore OpenAI’s Frontier Governance Framework and how our AI safety, security, and risk practices align with emerging EU and California regulations.
中文介绍 OpenAI发布前沿模型治理框架，阐述其在AI安全、安全保障与风险管理方面的核心实践，并说明相关措施如何与欧盟及美国加利福尼亚州的最新AI监管法规保持一致，以应对日益严格的合规要求。
中文介绍 AI行业每日简报涵盖三大要点：OpenAI推出私有MCP（模型上下文协议）；Cognition公司估值达260亿美元；ElevenLabs发布Music v2音乐生成新版本。
Biohub’s Protein World Model: ESMC-6B, ESMFold2, 6.8B proteins, 1.1B structures, antibody design, SAEs, & the potential for programmable biology
中文介绍 Chan Zuckerberg Biohub研究员Alex Rives介绍蛋白质世界模型最新进展，包括60亿参数的ESMC-6B模型、ESMFold2、涵盖68亿蛋白质及11亿结构的训练数据集，以及在抗体设计和稀疏自编码器方面的应用，探索可编程生物学潜力。
中文介绍 IBM研究院与Artificial Analysis联合发布首个面向企业IT智能体任务的基准测试ITBench-AA。测试结果显示，当前主流前沿大模型得分均低于50%，表明现有AI智能体在复杂企业IT运维场景中的能力仍存在明显不足。
Cisco and OpenAI are redefining enterprise engineering with Codex, helping Cisco scale AI-native development, accelerate AI Defense work, and automate defect remediation.
中文介绍 思科与OpenAI宣布合作，共同将Codex引入企业级工程体系。合作内容包括：助力思科推进AI原生开发规模化落地、加速AI Defense安全产品研发，以及自动化修复代码缺陷，重新定义企业软件工程模式。
See how OpenAI, Thrive, and Crete built a self-improving tax agent with Codex, automating filings, improving accuracy, and accelerating workflows.
中文介绍 OpenAI联合Thrive和Crete，基于Codex构建自我改进型税务智能体，实现税务申报自动化、持续提升准确率并加快工作流处理效率，展示Codex在垂直专业领域构建高度自动化AI应用的能力。
it's funding news, but it's good news.
中文介绍 AI基础设施领域出现新晋十角兽：Fireworks AI与Baseten相继完成大额融资，估值双双突破百亿美元；OpenRouter据悉亦在融资进程中。此轮融资潮显示资本持续向AI推理与模型服务基础设施集中。
Ahead of global elections, we’re helping people access information, supporting cyber defenders, and increasing AI transparency
中文介绍 OpenAI面向2026年全球选举季发布选举信息保障举措，涵盖三大方向：协助公众获取准确选举信息、为网络安全防御者提供支持，以及提升AI生成内容的透明度，以应对选举期间的虚假信息风险。
Warp uses GPT-5.5 and OpenAI models to coordinate coding agents across local, cloud, and open-source development workflows.
中文介绍 开发工具公司Warp押注GPT-5.5及OpenAI系列模型，在本地、云端和开源三种开发环境中协调编码智能体协同工作，以AI驱动的方式重构开源软件开发工作流，展示下一代AI原生开发工具的运作模式。
中文介绍 Reachy Mini机器人实现完全本地化运行，无需依赖云端服务即可完成对话交互任务。该方案通过在本地部署语言模型，提升机器人对话的隐私性与响应速度，是边缘侧具身智能的一次实践探索。
中文介绍 Hugging Face在TRL框架中推出Delta Weight Sync（增量权重同步）技术，借助Hub存储桶实现万亿参数级别模型权重的高效分发与同步，大幅降低大规模模型训练过程中的权重传输开销。
中文介绍 AI行业每日简报涉及三则要点：xAI对Cursor的使用设置访问限制；DeepSWE软件工程智能体发布；美国对中国AI从业者实施出行限制，涉及参加AI相关活动的签证与入境管控政策调整。
本交易日技术面呈现明显的板块分化格局。美股方面，SPY与QQQ双双逼近52周高点（分别距高点-0.07%和-0.14%），多头排列结构完整，RSI14分别达73.2和76.4，均处于超买区域，短期动量强劲但超买信号值得关注。AAPL的RSI14更高达79.8，为本批次最高读数。MSFT则表现相对滞后，RSI14仅59.3，且价格仍低于SMA200（458.86），技术趋势标注中性。NVDA出现MACD死叉，5日跌幅-4.13%，需留意动量减弱迹象。加密资产方面，恐慌贪婪指数仅为22（极度恐慌），加密总市值24小时下降0.9%至2.55T USD，BTC主导率57.7%，ETH主导率9.5%。BTC-USD当前73462美元，RSI14跌至35.5，价格全面跌破SMA20/50/200（三线分别为77936/77171/79972），空头排列明显；ETH-USD情况更为疲弱，RSI14仅30.8，较52周高点回撤-59.52%，MACD持续扩张下行。中概股技术面整体偏空，拼多多、腾讯控股RSI14均触及29.3超卖水平，多数标的跌破三条均线，空头排列显著。商品与外汇方面，黄金1日反弹1.72%但5日基本持平，位于SMA20/50下方，技术信号中性；WTI原油5日跌幅达-9.9%，MACD形成死叉；美元指数DXY维持在SMA20/50/200之上，接近52周高点，整体偏强。10年期美债收益率RSI14恰好处于50中性线，MACD微幅死叉，方向待定。VIX恐慌指数15.74，5日下行-6.09%，低于SMA20/50，反映当前美股市场整体情绪相对平稳。
当前价735.6，较52周低点累计上涨43.69%，距52周高点仅-0.14%，处于极强位置。SMA20/50/200分别为705.51/650.06/617.05，呈标准多头排列。RSI14读数76.4，处于超买区域，MACD 21.2604略高于信号线21.4302，柱状图转负但幅度极小。5日涨幅3.15%，动量持续偏强，技术面整体维持上行结构，但RSI超买需关注短期回调压力。
当前价83.03，5日跌幅高达-15.4%，距52周高点回撤-40.44%，接近52周低点（距低点仅+1.8%）。SMA20/50/200分别为96.57/98.74/113.39，价格全面跌破三条均线，呈空头排列。RSI14仅29.3，触及超卖区域。MACD -2.6735，信号线-1.4474，形成死叉，柱状图持续扩张下行。多项指标共振偏空，技术面结构明显偏弱。
当前价425港元，距52周高点回撤-37.77%，距52周低点仅+1.09%，处于年度低位区间。SMA20/50/200分别为457.4/487.02/575.3，价格深度跌破三条均线，空头排列结构清晰。RSI14为29.3，进入超卖区域。MACD -15.8398，信号线-14.0005，MACD柱状图持续为负，空头动量未见收敛。5日跌幅-6.63%，技术压力持续。
当前价442.1，5日涨幅5.95%，距52周高点回撤-11.37%。SMA20/50分别为418.69/390.93，价格位于两者之上；但SMA200为411.65，价格略高于SMA200，均线结构中性偏多。RSI14为63.6，处于正常偏强区间。MACD 12.1923高于信号线11.1926，刚形成金叉，动量转向偏积极，但整体趋势标注为neutral，尚未确认方向。
当前价4523.9，1日涨幅1.72%，但5日微跌-0.16%，短期波动方向不明。价格低于SMA20（4593.63）和SMA50（4637.35），但高于SMA200（4364.92），结构分化。RSI14为43，处于中性偏弱区域。MACD -55.0499，信号线-45.7135，绿柱持续但1日反弹明显，距52周高点回撤-19.02%，技术面处于多空角力阶段，方向信号尚不明朗。
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
本报告仅基于公开行情数据计算所得技术指标进行客观描述，过去走势不代表未来表现，指标读数不构成任何投资建议或交易依据。本内容仅供技术指标解读参考，投资者须自行承担相关风险。
French students are taking their exams in baking schools, while red alerts are issued in Italy.
中文摘要 欧洲遭遇热浪侵袭，葡萄牙打破5月最高气温历史纪录。法国学生在酷热的校舍中参加考试，意大利多地发布红色高温预警。此轮热浪波及欧洲多国，造成严重影响。
US officials earlier told the BBC that, pending the approval of Trump and Iran's leadership, the framework of a ceasefire extension deal had been agreed.
中文摘要 美国副总统万斯表示，美伊两国在达成协议方面已「非常接近」，但「尚未完成」。此前美方官员向BBC透露，在特朗普及伊朗领导层批准前提下，双方已就停火延长协议的框架达成一致。
President Trump has not signed off on the emerging framework, according to U.S. officials. But it could set the table for extending the cease-fire and more substantive negotiations.
中文摘要 美国官员表示，美伊双方正接近达成重开霍尔木兹海峡的协议安排。特朗普尚未正式签署相关框架，但该框架有望为延长停火并推进更实质性谈判奠定基础。
A hunger strike at the New Jersey facility has raised questions about conditions in US immigration detention centres.
中文摘要 美国新泽西州德莱尼大厅移民拘留中心外爆发紧张抗议活动。该设施内被拘留者发起绝食抗议，外界对美国移民拘留中心的关押条件提出质疑。
Attorney general says Nacc needs public support ‘in terms of perception and in practice’. Follow today’s news live Get our breaking news email, free app or daily news podcast Who’s leading the fight against Labor’s CGT reform – and what’s in it for them? People with direct personal financial stakes
中文摘要 澳大利亚司法部长罗兰宣布将对国家反腐委员会（NACC）进行「重置」，此前该委员会主席布雷顿已宣布辞职。澳大利亚安全情报局（ASIO）警告政界人士不应在汽车内进行涉密谈话。
Rising heat in Saudi Arabia threatens millions of Muslim pilgrims – but cutting fossil fuels would keep it safer Global heating has “fundamentally altered” the climate of Mecca and is exposing millions of hajj pilgrims to extreme and dangerous heat even in months outside summer, new analysis has fou
中文摘要 最新报告指出，全球变暖已「根本性地改变」麦加气候，使每年参加朝觐的数百万穆斯林信众即便在夏季以外的月份也面临极端高温威胁，朝觐安全风险持续上升。报告呼吁削减化石燃料使用以降低风险。
中文摘要 该条目为美伊局势动态追踪页面，暂无具体摘要内容。
US vice-president says he feels ‘pretty good’ about prospects of memorandum of understanding being negotiated with Tehran Hezbollah has claimed dozens of drone and rocket attacks that it said targeted Israeli troops in southern Lebanon and northern Israel. The group said it launched several attacks
中文摘要 美国副总统万斯表示对与伊朗谈判达成谅解备忘录持「相当乐观」态度，但强调尚未达成协议。与此同时，真主党宣称对以色列南部及北部发动数十次无人机和火箭袭击，称目标为以色列军队。
Taiwan Travelogue is the first novel translated from Mandarin Chinese to win the prestigious award.
中文摘要 台湾作家的饮食题材小说《台湾旅行记》荣获今年布克奖，成为首部从中文（普通话）翻译后获得该奖项的小说，引发国际文坛广泛关注。
Indians celebrate Eid al-Adha amid prayer restrictions.
中文摘要 印度各地穆斯林在祈祷活动受限的情况下庆祝宰牲节（Eid al-Adha），部分地区对公共宗教活动实施了限制措施。
Republic of Ireland are set to face Israel twice in UEFA Nations League matches, but are facing calls for a boycott.
中文摘要 爱尔兰与卡塔尔的热身赛因抗议活动两度中断，起因是爱尔兰队即将在欧洲国家联赛中两度迎战以色列队，外界呼声要求抵制相关赛事。
President Trump has yet to sign off on an extension of the cease-fire that would allow the two sides to negotiate nuclear and other issues.
中文摘要 美国与伊朗正就重开霍尔木兹海峡协议积极推进磋商，特朗普尚未签署停火延长方案。一旦协议达成，将为双方就核问题及其他议题展开进一步谈判创造条件。
Speaking in New York, the Canadian prime minister called for a new relationship between Canada and United States.
中文摘要 加拿大总理卡尼在纽约发表演讲，呼吁重塑加美两国关系，并表示加拿大与美国保持一定距离对美方同样有利，体现其政府寻求战略自主的立场。
U.S. and Iranian officials say they are closing in on the terms of a preliminary agreement. Yet sticking points, particularly over the Strait of Hormuz, remain.
中文摘要 美伊双方官员透露，一份初步协议草案已摆上谈判桌，双方正就协议条款收尾。但分歧依然存在，尤其在霍尔木兹海峡问题上仍有争议，谈判尚未最终完成。
A big challenge for Colombia's next president will be how to reduce country's debt without sacrificing social gains.
中文摘要 哥伦比亚首届左翼政府任内成功降低了贫困率，但同时积累了大量债务。如何在不损害社会福利成果的前提下削减债务，将成为下任总统面临的重大挑战。
BMW is introducing humanoid robots to a car plant in Europe, building on similar projects in the US.
中文摘要 宝马公司宣布将人形机器人引入欧洲汽车生产工厂，此前已在美国开展类似项目。宝马表示，人形机器人是汽车制造业的未来发展方向，此举标志着汽车行业自动化进程迈入新阶段。
New round values Claude AI maker at $965bn including the latest money
中文摘要 Anthropic完成新一轮融资，融资额达650亿美元，使该公司估值升至9650亿美元（含最新融资），超越OpenAI估值。Anthropic旗下AI产品为Claude大语言模型。
Mitsui & Co. CEO says the company is eyeing global expansion in its LNG business amid growing energy demand driven by AI data center buildouts. He spoke exclusively with Bloomberg TV’s Shery Ahn in Tokyo. (Source: Bloomberg)
中文摘要 三井物产CEO在东京接受彭博电视专访时表示，随着AI数据中心建设推动能源需求持续增长，公司正积极布局全球液化天然气（LNG）业务扩张，将AI驱动的能源需求视为重要商机。
Asian equities were poised to advance after the US and Iran reached a tentative deal to extend their ceasefire. Treasuries rallied and oil dropped.
中文摘要 受美国与伊朗达成初步停火延期协议提振，亚洲股市预计走高。美国国债价格上涨，油价下跌。市场情绪因地缘政治紧张局势缓和而整体改善。
Oil edged lower after the US and Iran tentatively agreed to extend a ceasefire by 60 days, according to a person familiar with the matter, potentially allowing shipments through the Strait of Hormuz to resume.
中文摘要 据知情人士透露，美国与伊朗初步同意将停火协议延长60天，此举有望恢复经霍尔木兹海峡的石油运输。受此消息影响，国际油价小幅走低。
Worthington Steel Inc. raised $1.4 billion in the leveraged-finance market Thursday to help fund its acquisition of German metals company Klöeckner & Co SE.
中文摘要 美国钢铁企业Worthington Steel通过杠杆融资市场筹集14亿美元，用于收购德国金属公司Klöeckner & Co SE。此次融资于周四完成，为此次跨国并购交易提供资金支持。
The AI dynamos are eyeing the public markets, but some wonder whether SpaceX’s planned June listing will steal their thunder.
中文摘要 OpenAI与Anthropic均有意寻求公开上市，但市场人士对其IPO能否达到预期存有疑虑，部分观点认为SpaceX计划于6月进行的上市可能分散投资者注意力，对两家AI公司的上市前景构成竞争压力。
Japan’s financial markets have entered a new phase in which rising bond yields no longer signal just the healthy normalization of monetary policy, but also the risk that inflation may increase too much.
中文摘要 日本金融市场进入新阶段，债券收益率持续上升已不仅反映货币政策正常化，还预示通胀过度上行的风险。五张图表显示，收益率走高正在深度重塑日本市场格局。
Pauline Brown, former chair of LVMH North America, joined the discussion to analyze the ongoing k-shaped economic recovery, particularly within the luxury retail sector. She highlighted that even among luxury brands, performance is diverging significantly. She speaks with Romaine Bostick & Katie Gre
中文摘要 LVMH北美前主席波利娜·布朗分析奢侈品行业的「K型复苏」现象，指出即使在奢侈品牌内部，各品牌业绩表现也出现明显分化，高端市场与中档奢侈品消费之间的差距持续扩大。
Senior executive Dave Treadwell tells staff ‘don’t use AI just for the sake of using AI’ as costs rise
中文摘要 亚马逊高级管理人员戴夫·特雷德韦尔宣布废除内部AI使用排行榜，要求员工「不要为用AI而用AI」，此举旨在遏制员工为追求使用率评分而过度使用AI工具、导致公司成本上升的问题。
Hormel Foods reported adjusted earnings per share and organic sales for the second quarter that exceeded analyst expectations, driving a notable increase in its stock price. John Ghingo, President of Hormel Foods, discussed the company's ongoing turnaround plan and its positive impact on results. He
中文摘要 荷美尔食品公布第二季度业绩，调整后每股收益及有机销售额均超越分析师预期，推动股价明显上涨。公司总裁约翰·金戈介绍了公司持续推进的业务转型计划及其积极进展。
Bloomberg's Joel Weber puts Katie Greifeld and Eric Balchunas to the test on "Bloomberg ETF IQ" in this week's edition of "IQ Test." (Source: Bloomberg)
中文摘要 彭博《ETF IQ》节目本周推出「IQ测验」环节，主持人乔尔·韦伯对凯蒂·格雷菲尔德和埃里克·巴尔丘纳斯进行ETF知识测试，为投资者提供ETF相关专业知识内容。
8 回复 · 程序员 节点
9 回复 · 程序员 节点
39 回复 · Apple 节点
13 回复 · Apple 节点
26 回复 · Apple 节点
11 回复 · Apple 节点
16 回复 · Apple 节点
22 回复 · Apple 节点
10 回复 · Apple 节点
4 回复 · Apple 节点
我们伟大的，最顶级的Deepseek被Opus蒸馏了 A/已经不受限于Qwen了，它是自由的 63 个帖子 - 47 位参与者 阅读完整话题
基本信息 官方文：https://www.anthropic.com/news/claude-opus-4-8 https://cdn.sanity.io/files/4zrzovbb/website/c886650a2e96fc0925c805a1a7ca77314ccbf4a6.pdf Models overview - Claude API Docs https://www.reddit.com/r/ClaudeAI/ https://www.reddit.com/r/ClaudeCode/ 价格：输入$5/输出$25不变 上下文：1m不变 最大输出：128k不变 训练数据：202601
十分稳定，直接发包都是这个问题，稳定复现，佬们可以自己试试： 官api，上面那么明显的url，别再说第三方掺水了： 嘴上说着反蒸馏，结果自己还是干了 话说为什么蒸qwen蒸这么多，不应该是gpt之类的吗 甚至有时会说自己是Deepseek： 总之就是大概率qwen，小概率Deepseek，就是不认为是claude claude-opus-4-8蒸馏了太多qwen模型，导致自我认知出了问题，基本认为自己是qwen 前沿快讯 回顾一下a\的嘴脸： anthropic.com Detecting and preventing distillation attacks Anthropic is an
本帖使用社区开源推广，符合推广要求。我申明并遵循社区要求的以下内容： 我的帖子已经打上 开源推广 标签： 是 我的开源项目完整开源，无未开源部分： 是 我的开源项目已链接认可 LINUX DO 社区： 是 我帖子内的项目介绍，AI生成、润色内容部分已截图发出： 是 以上选择我承诺是永久有效的，接受社区和佬友监督： 是 以下为项目介绍正文内容，AI生成、润色内容已使用截图方式发出 众所又周知，现在 GPT 的 PLUS / FREE 账号，不绑定手机号不拿 RT 已经不能愉快的使用 CodeX 了 在线接码比价：https://sms.fur.li/ 开源地址：GitHub - FoundZiG
本帖使用社区开源推广，符合推广要求。我申明并遵循社区要求的以下内容： 我的帖子已经打上 开源推广 标签： 是 我的开源项目完整开源，无未开源部分： 是 我的开源项目已链接认可 LINUX DO 社区： 是 我帖子内的项目介绍，AI生成、润色内容部分已截图发出： 是 以上选择我承诺是永久有效的，接受社区和佬友监督： 是 以下为项目介绍正文内容，AI生成、润色内容已使用截图方式发出 MobileGym (不是移动健身房) 有点标题党了，但是真烧了几百亿 Token，纯前端 TypeScript + React，实现了28 个仿真 APP——微信、支付宝、小红书、bilibili、X、Reddit、
群里看到的，绷不住了 13 个帖子 - 13 位参与者 阅读完整话题
本帖使用社区开源推广，符合推广要求。我申明并遵循社区要求的以下内容： 我的帖子已经打上 开源推广 标签： 是 我的开源项目完整开源，无未开源部分： 是 我的开源项目已链接认可 LINUX DO 社区： 是 我帖子内的项目介绍，AI生成、润色内容部分已截图发出： 是 以上选择我承诺是永久有效的，接受社区和佬友监督： 是 以下为项目介绍正文内容，AI生成、润色内容已使用截图方式发出 github.com GitHub - Caph-dev/agents-progressive-disclosure: A skill to refactor bloated AGENTS.md, CLAUDE.md,
vibe coding成瘾了一段时间之后，最大的感受的就是惊喜来得快，烦恼来得更快。 最明显的一个问题就是vibe几分钟，review几小时，除了玩具web能一把梭之外，很多时候，以质量为导向的复杂功能开发，vibe的时间效率其实是不比手搓高多少的。 为了 实现token、效率、结果的平衡 ，也尝试了很多方案、skills。最近终于摸索出截止26年5月的自认为还挺可行的方案，各位佬也可以分享讨论一下： 1.chatgpt网页版用gpt-5.5pro(进阶专业)详细讨论需求，让他出一版prd或者开发文档初稿。 2.把文档拖进cc，用mattpocock的grill with doc skills
由于众所周知的原因，今天奥特曼拉闸了，但是，现在又有新的解决方案了 网址 muyuan.do 35 个帖子 - 35 位参与者 阅读完整话题
52 个帖子 - 44 位参与者 阅读完整话题