Haskell · ★ 16,555 · 🍴 957 · 📈 1,607 stars today
SimpleX - the first messaging network operating without user identifiers of any kind - 100% private by design! iOS, Android and desktop apps 📱!
中文介绍 Simplex 是首个不依赖任何用户标识的隐私消息网络，iOS、Android 和桌面端均可用。通过去中心化架构和无需电话号码或用户名的设计，实现 100% 匿名的端到端加密通信，适合高度注重隐私的个人或组织替代传统即时通讯工具。
Shell · ★ 118,849 · 🍴 19,466 · 📈 1,425 stars today
A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.
中文介绍 一套完整的 AI 代理系统，包含前端助手、Reddit 社区运营、创意注入、现实检测等多样化专业代理。每个代理拥有独立人格和工作流程，可独立执行任务。适用于需要自动化社交媒体互动、内容生成或创意辅助的团队，基于多 agent 编排技术。
Python · ★ 11,815 · 🍴 1,082 · 📈 352 stars today
NumPy & SciPy for GPU
中文介绍 CuPy 是 NumPy 和 SciPy 的 GPU 加速替代库，将矩阵运算和科学计算自动迁移到 NVIDIA GPU 上执行。无需修改代码即可获得数十倍性能提升，适合深度学习数据预处理、数值模拟等需要大规模并行计算的 Python 用户。
Swift · ★ 4,376 · 🍴 274 · 📈 830 stars today
FluidVoice - Fastest macOS Offline Dictation app - Voice to Text fully Local. One ⭐ takes us a long way :))
中文介绍 macOS 上速度最快的离线语音转文字应用，完全在本地处理，无需联网。基于本地语音识别引擎，保障隐私的同时实现实时听写。适合需要高效输入文字、又担心数据外泄的 macOS 用户，特别在无网络或隐私敏感场景下使用。
Python · ★ 34,363 · 🍴 2,597 · 📈 224 stars today
🕵️♂️ Collect a dossier on a person by username from 3000+ sites
中文介绍 通过用户名从超过 3000 个社交平台和网站搜集公开信息，生成目标人物综合档案（dossier）。采用爬虫和 OSINT 技术，可批量扫描、自动匹配。适用于安全研究人员、调查记者的背景调查、账户关联分析或社工情报收集。
Python · ★ 62,764 · 🍴 11,111 · 📈 458 stars today
openpilot is an operating system for robotics. Currently, it upgrades the driver assistance system on 300+ supported cars.
中文介绍 一个开源的机器人操作系统，目前用于升级超过 300 款汽车的驾驶辅助系统。基于计算机视觉和深度学习，实现车道保持、自适应巡航等高级辅助驾驶功能。适合汽车极客、自动驾驶开发者及希望增强爱车智能化的车主。
HTML · ★ 126,706 · 🍴 13,257 · 📈 1,935 stars today
A list of SaaS, PaaS and IaaS offerings that have free tiers of interest to devops and infradev
中文介绍 一份精心整理的 SaaS、PaaS 和 IaaS 免费套餐列表，覆盖开发运维和基础设施领域。没有多余介绍，直接提供域名、免费额度、限制等实用信息。是开发者和运维人员寻找零成本部署、CI/CD、监控等服务的首选参考清单。
TypeScript · ★ 12,649 · 🍴 865 · 📈 158 stars today
🧑🚀 Authentication and authorization infrastructure for SaaS and AI apps, built on OIDC and OAuth 2.1 with multi-tenancy, SSO, and RBAC.
中文介绍 为 SaaS 和 AI 应用设计的认证与授权基础设施，基于 OIDC 和 OAuth 2.1 协议，支持多租户、单点登录（SSO）和基于角色的权限控制（RBAC）。开箱即用的身份管理方案，后端开发者可快速集成安全登录、用户管理等功能。
Python · ★ 6,594 · 🍴 865 · 📈 1,386 stars today
AI 时代的伯克希尔：基于 Claude Code / Codex 的价值投资研究框架。巴菲特·芒格·段永平·李录四大师方法论 + 多Agent并行研究。| AI-era Berkshire: a value investing research framework built for Claude Code / Codex. 4 masters' methodologies + multi-agent adversarial analysis.
中文介绍 AI 时代的价值投资研究框架，基于 Claude Code 或 Codex 驱动，结合巴菲特、芒格、段永平、李录四大师方法论，通过多 Agent 并行分析股票。自动收集财务数据、计算估值指标、生成研究报告，适合价值投资者和量化分析人员。
Python · ★ 11,925 · 🍴 1,570 · 📈 967 stars today
Edit videos with coding agents
中文介绍 利用编程 agent 自动编辑视频的工具。通过 agent 控制浏览器或调用视频处理 API，实现剪辑、合成、转场等操作。适合需要批量处理视频、自动化内容生成的创作者，降低手动操作成本。
Python · ★ 1,124 · 🍴 171 · 📈 129 stars today
基于 AI Agent + MCP 工具链 + 渗透 Skill 编排， 配合大语言模型， 自然语言输入 → 自动完成「信息收集 → 漏洞发现 → 漏洞利用 → 报告生成」全流程。
中文介绍 AI 驱动的全自动渗透测试工具链：自然语言输入需求后，自动完成信息收集、漏洞发现、漏洞利用和报告生成。基于 AI Agent 编排安全工具（MCP 工具链）和渗透技能流程，适合安全研究员和红队人员快速验证漏洞。
Shell · ★ 1,879 · 🍴 201 · 📈 331 stars today
18 AI personas deliberate your hardest decisions across multiple LLM providers. Aristotle, Feynman, Kahneman, Torvalds & more — structured multi-round deliberation with genuine model diversity. One command: /council
中文介绍 一个集成 18 个 AI 人格（亚里士多德、费曼、卡尼曼、托瓦兹等）的多轮辩论系统，支持跨多个 LLM 提供商。通过结构化的多轮商议和模型多样性，帮助用户分析复杂决策问题。适用于需要多角度思考的场景，如商业决策、伦理判断。
Python · ★ 15,079 · 🍴 2,686 · 📈 839 stars today
"Vibe-Trading: Your Personal Trading Agent"
中文介绍 个人交易代理（Trading Agent），利用 AI 辅助进行股票、加密货币等投资决策。根据用户设定的交易风格（Vibe），自动分析市场数据、执行策略。适合量化交易爱好者及希望借助 AI 优化交易的个人投资者。
TypeScript · ★ 17,504 · 🍴 1,205 · 📈 280 stars today
Desktop app to manage markdown knowledge bases
中文介绍 一款桌面端知识库管理应用，专为 Markdown 文件设计。支持文件夹组织、全文搜索、标签分类等功能，帮助用户高效管理笔记、文档和个人知识。适合写作者、研究人员以及习惯用 Markdown 记笔记的知识工作者。
C · ★ 10,471 · 🍴 1,231 · 📈 186 stars today
Disk encryption with strong security based on TrueCrypt
中文介绍 VeraCrypt 是基于 TrueCrypt 的磁盘加密软件，提供全磁盘、分区或文件容器加密。支持多种加密算法，具有抗暴力破解的额外安全层。适合需要保护敏感数据（如加密系统盘或移动存储）的个人用户或企业，是安全存储的首选工具。
👍 57
Modern Vision-Language-Action (VLA) models often fail to generalize to novel setups, such as altered camera viewpoints or robot morphologies, because they are typically conditioned only on current observations and language instructions. By ignoring the underlying system configuration as a variable,
中文介绍 提出上下文世界建模方法，使VLA机器人在上下文中隐式编码系统配置，无需重新训练即可泛化到未见过视角或机器人形态。实验证明跨设置泛化能力显著提升。
👍 49
Outcome-based reinforcement learning provides a stable optimization backbone for language agents, but its sparse trajectory-level rewards provide little guidance on which intermediate decisions should be reinforced or suppressed. On-policy self-distillation offers dense token-level supervision, yet
中文介绍 提出OPID，结合结果奖励的稳定性和on-policy自蒸馏的密集监督，为语言智能体RL提供细粒度信用分配。在多个agentic benchmark上提升了样本效率和最终性能。
👍 43
A classical intuition holds that verifying a solution is easier than producing one. For today's coding agents, this intuition is being inverted: as foundation models develop stronger reasoning capabilities and engineering harnesses grow more sophisticated, generating complex candidate solutions is n
中文介绍 指出编码智能体中“验证比生成容易”的传统直觉已逆转。随着模型推理增强，生成简单验证难，实验验证“验证地平线”现象，为奖励设计提供新视角。
👍 38
A unified representation for text and vision is a natural pursuit, as it enables simpler multimodal modeling and more efficient training. However, representing images as discrete signals in the same way as text inevitably introduces severe information loss. Existing work struggles to balance low-lev
中文介绍 提出ViQ，一种文本对齐的视觉量化表示，支持任意分辨率。通过改进量化策略和对比学习，在保持离散性与文本对齐的同时减少信息损失，多模态任务上取得SOTA。
👍 36
We introduce an axiomatic evaluation framework for latent thought representations in LLMs, comprising metrics that are independent of downstream benchmark scores and reveal representational failures that benchmark accuracy masks. Existing evaluations conflate representation quality with model capaci
中文介绍 形式化潜在思维表示的四条公理，提出独立于下游性能的评估框架，揭示传统基准准确率掩盖的LLM内部表示失败模式，为思维链可信度提供检测工具。
👍 32
Speculative decoding (SD) accelerates autoregressive Large Language Models (LLMs) by drafting multiple tokens and verifying them in parallel, but it faces a scaling limitation: increasing the draft budget improves speed only when acceptance remains high and drafting overhead stays low. This ceiling
中文介绍 提出JetSpec，通过并行树草稿打破推测解码的缩放天花板。树状草稿提升接受率并控制开销，在多种LLM上实现2-3倍加速，优于传统单链方法。
👍 28
Computer-use agents can execute software tasks through either graphical interfaces or programmatic command interfaces, but existing evaluations confound interaction modality with differences in tasks, initial states, verifiers, and permitted actions. We introduce a matched execution-layer benchmark
中文介绍 构建匹配执行层基准，系统对比GUI与CLI两种交互方式下的计算机使用智能体瓶颈。发现CLI在复杂任务上效率更高，GUI在直观性上有优势，为混合设计提供指导。
👍 27
We present Qwen-Image-2.0-RL, a post-training pipeline that applies reinforcement learning from human feedback (RLHF) and on-policy distillation (OPD) to improve both the visual quality and instruction-following capability of the Qwen-Image-2.0 diffusion model. To provide reliable reward signals, we
中文介绍 报告Qwen-Image-2.0-RL，采用RLHF和on-policy蒸馏联合优化扩散模型。通过人类偏好奖励和在线蒸馏，显著提升图像质量与指令跟随能力，超越前代。
👍 25
Joint-Embedding Predictive Architectures (JEPAs), including recent LeWorldModel (LeWM), have become a promising foundation for reconstruction-free visual world models. For visual planning, however, LeWM evaluates candidate action sequences by repeatedly applying a local one-step latent transition mo
中文介绍 提出Fast LeWorldModel，通过全局潜空间预测替代局部单步迭代，加速JEPA视觉世界模型规划。在多个规划任务中保持精度同时实现数倍速度提升。
👍 18
As agentic systems continue to evolve and are widely deployed in real-world scenarios, there is a growing demand to faithfully evaluate their capabilities. However, current benchmarks are typically built on popular applications with relatively simple tasks and focus on a narrow set of capabilities w
中文介绍 提出Gauntlet基准，在陌生环境中全面评估智能体能力，覆盖更广泛任务和未见过场景。揭示模型在熟悉环境外性能显著下降，指出泛化差距。
👍 16
Tool use enables large language models (LLMs) to perform complex tasks, and recent agentic reinforcement learning (RL) methods show promise for enhancing model capabilities. However, RL alone often leads to instability or limited gains in tool-use tasks. In our experiments, some models exhibit catas
中文介绍 分析多步工具使用RL崩溃原因：稀疏奖励和探索困难导致灾难性遗忘。引入监督信号（模仿学习/过程奖励）稳定训练，显著提升复杂任务成功率。
👍 13
The prevalent dual-branch paradigm, i.e., training a side network to encode visual conditions and fusing its intermediate-layer features to a frozen pretrained main network, has shown remarkable success in visual-condition controllable generation. Despite its widespread adoption, the role of the sid
中文介绍 提出LISA，通过似然得分对齐优化视觉条件可控生成中侧网络的作用。理论分析并设计对齐损失，使生成更精准符合条件，多项任务提升一致性。
👍 11
Vision-language models (VLMs) are increasingly deployed in consumer, medical, financial, and enterprise applications. This broad deployment expands the safety surface: risks can arise from multimodal question answering, assistant responses, and cross-modal composition, while moderation policies may
中文介绍 提出SingGuard，一种策略自适应的多模态LLM护栏，支持动态推理。根据实时策略调整安全阈值，处理跨模态风险，有效拦截有害内容且误报率低。
👍 10
Reasoning capability has advanced rapidly in large language models (LLMs), leading to an increasing size of key-value (KV) cache in both prefilling and decoding stages. Existing KV cache compression methods mainly rely on attention weights to estimate token importance. While attention effectively ca
中文介绍 提出信息感知的KV缓存压缩，结合注意力权重与互信息衡量token信息贡献。在长上下文推理中缓存占用减少40%以上，保持模型性能。
👍 9
Text detoxification, the automated detection and mitigation of abusive and harmful content, is essential for ensuring the safety of online communities and protecting users. However, low resource languages such as Tatar have received little research attention. In this paper we present Tatoxa, a novel
中文介绍 提出Tatoxa系统，针对低资源语言鞑靼语的文本去毒。结合小型标注数据与跨语言迁移，构建检测与改写管道，在鞑靼语上取得有效结果。
👍 9
Process reward models enable fine-grained, step-level evaluation of LLMs, yet building them for agentic settings remains prohibitively difficult: long-horizon interactions, irreversible actions, and stochastic environment feedback make both human annotation and Monte Carlo estimation infeasible at s
中文介绍 利用后训练中的“进度优势”，无需显式过程奖励模型即可为LLM智能体提供细粒度监督。通过比较候选轨迹与已有轨迹的进展差异，有效引导策略学习。
👍 9
As LLM agents become capable of increasingly long-horizon tasks, evaluating their performance in economic systems is becoming increasingly important. Unlike existing benchmarks that primarily evaluate a single agent interacting with a passive environment, economic systems are inherently multi-agent,
中文介绍 提出CoffeeBench，评估长周期LLM智能体在异构多智能体经济中的表现。模拟协作与竞争，测试长期规划能力，发现当前模型策略稳定性不足。
👍 9
Video reasoning language models implicitly assume that every input frame is equally reliable. This leads to what we term the Blind Trust Problem: under realistic perturbations such as motion blur, glare, or occlusion, frontier video reasoning models can suffer 15-30%p accuracy drops on real-world em
中文介绍 提出置信度感知的工具编排，模型动态评估每帧可靠性，选择合适工具处理低质量帧。在真实扰动视频上准确率恢复高达20%，缓解盲信问题。
👍 8
Multi-agent systems (MAS) built on large language models (LLMs) provide a promising framework for solving complex tasks through role specialization and structured interaction. However, their performance is often limited by miscoordination and, more fundamentally, the lack of fine-grained credit assi
中文介绍 提出GBC，为LLM多智能体系统提供可微信用分配。通过计算智能体间交互梯度优化角色分工和通信策略，在协作推理任务上显著提升准确率和协调效率。
👍 8
Modern generative world models render increasingly realistic action-controllable futures, yet they frequently hallucinate: rollouts remain visually fluent while drifting from the ground-truth dynamics. We hypothesize that hallucination concentrates in low-coverage regions of the state-action space,
中文介绍 发现世界模型幻觉集中在状态-动作空间低覆盖区域，具有可预测性。提出主动探索或重加权训练等预防方法，减少幻觉同时保持视觉质量。
@sairahul1 · 121.4K 粉丝 · 876.6K 阅 · 547 赞 · 83 转
Emails. Formatting. Compiling reports. Preparing decks. Organizing files. Researching. Marketing. Writing. SEO. The average knowledge worker spends 60% of their day on above work that doesn't require
中文介绍 用Claude Cowork实现一人公司，自动化邮件、报告、PPT等杂务，释放知识工作者60%时间专注核心业务，提供具体工作流。
@crux_capital_ · 52.8K 粉丝 · 217.5K 阅 · 503 赞 · 56 转
This is the most in depth report I have ever written, on a very exciting company. If you read this article you will have a much deeper understanding of what Ouster does, how it plays into Physical AI,
中文介绍 作者发布对Ouster（激光雷达公司）的深度研究报告，解析其在Physical AI领域的技术定位、产品逻辑与市场前景，是当前最详尽的单篇分析。
@OracleLimited · 37.6K 粉丝 · 202.9K 阅 · 2.8K 赞 · 562 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 2026年自主AI Agent成为Polymarket上最有效策略，超30%活动来自算法。ORACLE项目推出官方AI Agent在该平台进行交易。
@OracleAiTrading · 34.1K 粉丝 · 176.1K 阅 · 2.7K 赞 · 567 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 2026年自主AI Agent在预测市场Polymarket上占比超30%，成为高效策略。ORACLE项目推出官方AI Agent入场交易。
@monokern · 1.9K 粉丝 · 175.8K 阅 · 546 赞 · 49 转
Agencies charge $5,000 for a portfolio site that looks this good I built mine in 2 hours. Here's exactly how This is the real walkthrough - not a generic template guide I'm using my own portfolio as
中文介绍 用Claude Code两小时搭建价值$10,000的动画个人网站，分享非通用模板的完整工作流，对比传统agency收费。
@Oracle__Market · 5.9K 粉丝 · 100.0K 阅 · 2.1K 赞 · 823 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 2026年自主AI Agent在Polymarket占超30%活动量，成为最有效策略之一。ORACLE推出官方AI Agent在该平台交易。
@Oracle_Market__ · 25.1K 粉丝 · 99.3K 阅 · 2.8K 赞 · 224 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 2026年预测市场Polymarket上超30%活动来自AI Agent，自主策略高效。ORACLE项目发布官方AI Agent进行交易。
@cyrilXBT · 186.6K 粉丝 · 77.2K 阅 · 510 赞 · 93 转
There is a specific moment that tells you whether someone is running a real one person company or just using a lot of apps. It is the moment they get asked a question about their own business and they
中文介绍 Hermes+Obsidian+Claude Code三件套构成一人公司高效系统，核心在于信息管理而非工具堆砌，能快速回应业务问题。
@RhysSullivan · 57.4K 粉丝 · 48.9K 阅 · 501 赞 · 23 转
i want to use the skills, knowledge, and apis your company has spent years developing, not your custom agent almost every company by this point has shipped an agent there's a cloudflare agent in their
中文介绍 反对企业定制Agent，主张开放API和多年积累的技能知识，让用户自由组合，而非被迫使用封闭的定制Agent。
@_avichawla · 71.1K 粉丝 · 39.8K 阅 · 501 赞 · 67 转
Every generate() call to an LLM runs two distinct computational phases on the same GPU: prefill (processing the prompt) is compute-bound while decode (generating tokens one at a time) is memory-bound.
中文介绍 清晰解释LLM推理两阶段：prefill（处理提示词，计算密集）和decode（逐个生成token，内存密集），帮助理解GPU性能瓶颈。
@akshay_pachaar · 279.5K 粉丝 · 35.5K 阅 · 505 赞 · 54 转
Build by Google, explained as a step-by-step guide. Karpathy defined agentic engineering at Sequoia Ascent 2026 as the discipline that separates production-grade agent work from vibe coding. The core
中文介绍 谷歌推出Karpathy在Sequoia Ascent 2026定义的「Agentic Engineering」工具，以逐步指南区分生产级Agent开发与Vibe Coding。
@unicity_labs · 126.3K 粉丝 · 5.2K 阅 · 531 赞 · 200 转
Unicity is building The Secure Compute Platform for Autonomous AI. Identity, execution, governance, and payments - rebuilt for machines, with no human in the loop. The internet is being rebuilt for
中文介绍 Unicity构建面向自主AI的安全计算平台，重新设计身份、执行、治理和支付，实现无人工干预的机器经济体。
中文介绍 Allen AI 发布 DiScoFormer，一种跨分布统一处理密度和评分的 Transformer 模型。
This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. Imagine coming in to work to learn that a new underling will report to you. The worker is not a person but an AI tool—one that your company nonetheless calls Alex
中文介绍 MIT Tech Review 指出，AI 智能体不是你的「同事」。公司引入 AI 工具作为下属，但本质仍是工具，而非人类合作伙伴。
Enterprise investment in AI is booming. Gartner is calling 2026 an “inflection year” for organizations to align their AI projects with strategic business objectives. As the pressure to prove ROI mounts, executives and technology leaders are looking to agentic AI to drive the measurable financial out
中文介绍 Gartner 称 2026 年为 AI「转折年」，企业需将 AI 项目与战略目标对齐。高管正押注智能体 AI 以证明投资回报率。
A new OpenAI report maps how AI could reshape jobs across the EU, highlighting which occupations may face automation, growth, or workflow changes.
中文介绍 OpenAI 报告分析 AI 对欧盟就业的影响，指出哪些职业面临自动化、增长或流程变化，描绘欧洲 AI 劳动力机遇。
中文介绍 6月29日AI新闻：OpenAI 预览 GPT-5.6；xAI 推出 Grok 4.5 测试版；Google 限制 Meta 相关行为。
HP Inc. scales its OpenAI Frontier partnership to deploy AI across customer experiences, software development, and enterprise operations.
中文介绍 HP 与 OpenAI 建立 Frontier 战略合作，将 AI 应用于客户体验、软件开发和企业运营，推动企业级 AI 部署。
Oddly tiered releases to both OAI and ANT on the same day.
中文介绍 OpenAI 发布 GPT-5.6 三个变种：Sol（旗舰）、Terra（中端）、Luna（低成本），仅限受信任合作伙伴访问，分层发布至 OAI 和 ANT。
OpenAI previews GPT-5.6 Sol, a next-generation model with stronger capabilities in coding, science, and cybersecurity, paired with its most advanced safety stack.
中文介绍 OpenAI 预览下一代模型 GPT-5.6 Sol，在编码、科学和网络安全方面能力更强，并配备最先进的安全堆栈。
**OpenAI** previewed **GPT-5.6** with three variants: **Sol** (flagship), **Terra** (mid-tier), and **Luna** (lower-cost), launching under a restricted rollout mandated by the U.S. government, limiting access to trusted partners. **Sol** boasts enhanced cybersecurity and safety features backed by ov
中文介绍 OpenAI 预览 GPT-5.6 三个变种：Sol（旗舰）、Terra（中端）、Luna（低成本）。受美国政府限制，仅向信任合作伙伴开放。Sol 强化网络安全与安全功能。
It's happening.
中文介绍 OpenAI 报告自2025年11月以来内部 Codex 输出 token 中位数增长：研究56倍、客户支持32倍、工程27倍、法律13倍，显示 AI 使用量激增。
中文介绍 Hugging Face 推出新功能：一条命令即可在 HF Jobs 上运行 vLLM 服务器，简化 LLM 部署流程。
中文介绍 6月26日AI新闻：美国与 OpenAI 诉讼；AI 经济现状；Scaling Laws 最新进展。
Artificial intelligence is rapidly reshaping retail, but not in the ways consumers might immediately notice. The biggest transformation may not be flashy virtual try-ons or chatbot shopping assistants, but in how decisions are made behind the scenes: how products surface in search results, how inven
中文介绍 MIT Tech Review 指出 AI 正快速重塑零售业，最大变革不在前端虚拟试穿或聊天机器人，而在后台搜索排名、推荐等决策方式。
**Z.ai's GLM-5.2** leads in coding and agent benchmarks with top scores like **1595** on Code Arena: Frontend and **34.29%** reasoning accuracy with zero failures. Databricks improved GLM-5.2 speed to **392 tok/s** using hardware and optimizations. **Ornith-1.0**, a new MIT-licensed coding model fam
中文介绍 Z.ai 的 GLM-5.2 在编码和智能体基准中领先：Code Arena: Frontend 得分1595，推理准确率34.29%零失败。Databricks 优化其速度至392 tok/s。新开源模型 Ornith-1.0 发布。
Move over, Harness Engineering, it is time for the harness of harnesses!
中文介绍 Latent Space 文章称「Meta-Harness 夏季」到来，从 Harness Engineering 演进至「harness of harnesses」，关注 AI 系统测试与评估框架的整合。
加密市场恐慌贪婪指数为15（极度恐慌），总市值2.17万亿美元，BTC主导率55.6%，ETH占9%，市场情绪悲观。美股表现分化：SPY均线多头排列但MACD死叉，短期动能减弱；科技股如AAPL、MSFT承压，RSI低于40。中概股普遍弱势，阿里巴巴RSI 18.5、京东RSI 26.2均进入超卖区，空头排列明显。商品方面，黄金期货RSI 33.7、WTI原油RSI 29.5双双超卖，价格远低于均线，空头压力延续。美元指数DXY偏强，RSI 66.9，均线多头排列，接近52周高。宏观方面，VIX回落至17.65，美债收益率下行，整体避险情绪浓厚，空头主导多数资产，但部分极端指标可能酝酿反弹。
RSI 18.5处于超卖区，价格95.51远低于所有均线（SMA20 112.10、SMA50 125.43、SMA200 147.98），MACD绿柱扩大，趋势字段bearish，空头排列明确，短期技术面偏弱。
价格741接近52周高，均线多头排列（SMA20>50>200），RSI 51.5处于中性区域，但MACD已形成死叉（0.06<2.14），短期动能减弱，长期趋势仍维持上行结构。
价格101.17接近52周高，均线多头排列，RSI 66.9偏强但未超买，MACD金叉（0.60>0.49），趋势字段bullish，技术面偏强势。
RSI 29.5处于超卖区，价格70.32低于所有主要均线（SMA20 82.38、SMA50 91.5、SMA200 73.93），MACD绿柱死叉，空头排列明显，短期承压。
RSI 54处于中性区，价格411.84介于SMA20（400.36）与SMA50（405.22）之间，MACD死叉但柱线缩小，趋势字段neutral，均线无明显方向，技术setup无显著偏向。
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
所有技术指标解读基于历史数据计算，过去走势不代表未来表现。本报告仅供技术指标解读参考，不构成任何投资建议，请独立判断并注意风险。
After President Trump said U.S.-Iran talks would take place in the Qatari capital on Tuesday, an Iranian official said its delegation would only be talking to mediators about getting the U.S. to uphold its cease-fire commitments.
中文摘要 伊朗表示其代表团将前往多哈，但并非与美国直接会谈，而是与调解方讨论要求美方履行停火承诺。此前特朗普称美伊会谈将于周二在卡塔尔首都举行。
McIntosh doubles down on call for Liberal ‘rebrand’. Follow today’s news live Get our breaking news email, free app or daily news podcast Chalmers defends house price forecasts The government has been facing some heat over the cooling house market around Australia with falling auction clearance rate
中文摘要 Geoffrey Watson批评暂停维多利亚州大型建设项目以应对腐败的决定「愚蠢」。McIntosh重申自由党需「重塑品牌」的呼吁。政府因房市降温面临压力。
An anti-immigrant June 30 deadline not recognised by authorities is fuelling fear among migrants.
中文摘要 南非6月30日的反移民最后期限令移民感到恐惧，尽管该期限未被当局承认。移民担心暴力事件发生。
Trump says meeting with Iran will take place in Doha today but Tehran says there are no plans for talks in coming days.
中文摘要 特朗普称美伊会谈将于今日在多哈举行，但伊朗否认，表示近期无会谈计划，与前者说法矛盾。
The DRC’s independence anniversary renews debate over progress, security and state capacity.
中文摘要 刚果（金）独立66周年之际，围绕国家进展、安全与国家能力的辩论再次被点燃。独立承诺仍未完全实现。
Brazil fans at a fan fest on Rio de Janeiro celebrated after their team overturned Japan’s early lead and won 2-1.
中文摘要 巴西队在里约热内卢的球迷节上逆转日本队，以2-1获胜，晋级16强，球迷们的焦虑化为喜悦。
中文摘要 无可用摘要。
Two people were killed and several others were injured after an Israeli strike hit the Mawasi area of Khan Younis, Gaza.
中文摘要 以色列空袭加沙汗尤尼斯马瓦西地区的人道主义区，导致两人死亡、多人受伤，帐篷被烧毁。
The authorities said a 45-year-old man opened fire at a youth facility after a quarrel about the custody of his baby daughter. The dead were all facility employees.
中文摘要 德国发生罕见枪击事件，一名45岁男子因婴儿女儿监护权纠纷，在青少年设施内开枪，造成6名员工死亡。
More than 100 people removed on ICE flight were being held in hotel in La Guaira when earthquakes struck More than 100 people just deported from the United States were being held in a hotel when earthquakes struck Venezuela, setting off a scramble to find survivors and bodies buried in the rubble, a
中文摘要 委内瑞拉拉瓜伊拉一家酒店在地震中倒塌，超过100名刚被美国驱逐出境的人员被安置在此，目前正在搜救。
Andy Burnham, who looks set to become prime minister next month, said he would set up a new operation in Manchester called “No. 10 North” to give more funding and control to local leaders.
中文摘要 有望成为英国下任首相的安迪·伯纳姆承诺将权力移出伦敦，计划在曼彻斯特设立「北部唐宁街10号」，为地方领导提供更多资金和自主权。
In Celaya, Mexico, local cartels have turned soccer fields into bloody battlefields as they have co-opted a local league for teenagers as their latest business endeavor.
中文摘要 墨西哥塞拉亚的贩毒集团将青少年足球联赛作为最新生意，足球场变成血腥战场。
Two people said to be in critical condition after CCTV shows man dropping backpack at residential building lobby An explosion in Monaco has triggered a police hunt for a man suspected of detonating a makeshift bomb that injured three people. The blast occurred on Monday evening in the semi-enclave f
中文摘要 摩纳哥一居民楼大厅发生爆炸，致三人受伤，其中两人伤势危重。监控显示一名男子放下背包后爆炸，警方正在搜捕嫌疑人。
Kerr signs with the NWSL champions until 2030, returning to the US league after six years spent with Chelsea in England.
中文摘要 澳大利亚女足球星萨姆·科尔在离开切尔西后，与美国国家女子足球联赛冠军哥谭FC签约至2030年，重返美国联赛。
A pilot in the Afghan Air Force, he was recruited for the Soviet space program. His 1988 flight took place as the Soviets were fighting a war in his country.
中文摘要 唯一进入太空的阿富汗人阿卜杜勒·阿哈德·莫曼德去世。他原是阿富汗空军飞行员，被苏联太空计划招募，于1988年苏联在阿富汗战争期间完成太空飞行。
Three in five homes listed for sale since January remain on the market, says property portal Zoopla.
中文摘要 据房地产门户Zoopla数据，自1月以来挂牌出售的房屋中五分之三仍未成交，高抵押贷款利率令买家却步。
The Fire (Financially Independent, Retire Early) movement sees followers save as much as possible.
中文摘要 FIRE运动（财务独立、提前退休）倡导者通过连续10年自带午餐等极简储蓄，在40岁实现退休。
After a banner year for Chinese stocks on the back of AI advances, 2026 is not going well.
中文摘要 2026年中国股市表现落后全球，创2001年以来最大差距，尽管此前AI热潮曾推动上涨，但今年未能延续。
The Taiwan dollar, which is on track for its first quarterly gain in a year, may face renewed pressure as the US currency strengthens and local companies boost dividend payouts to a record.
中文摘要 新台币本季度有望录得一年来首次季度上涨，但可能因美元走强及本地企业创纪录的股息派发而面临新一轮压力。
The International Monetary Fund and Egypt reached a preliminary deal on the country’s latest program review, putting Cairo on track to unlock more than $1.6 billion in financing.
中文摘要 国际货币基金组织（IMF）与埃及就最新项目审查达成初步协议，埃及将有望获得超过16亿美元融资。
Brazilian sugar and fuels company Raizen SA reported a cash burn of 3.4 billion reais ($657 million) for the first three months of 2026 and roughly doubled its impairment charge for the crop year, signaling continued challenges as it proceeds with its debt restructuring plan.
中文摘要 巴西糖与燃料公司Raizen SA 2026年前三个月现金消耗34亿雷亚尔（6.57亿美元），并将作物年度减值损失几乎翻倍，债务重组面临持续挑战。
Asian equities rose, leaving the regional benchmark on track for its best quarterly gain in 17 years as technology stocks rallied. The yen weakened to a four-decade low against the dollar.
中文摘要 亚洲科技股上涨推动区域基准指数有望创17年来最佳季度涨幅；日元跌至40年低点。
Forgent Power Solutions Inc. and its private equity backer Neos Partners LP are offering to sell about 35 million shares, marking the third stock sale since the power distribution equipment maker went public in February.
中文摘要 Forgent Power Solutions及其私募支持者Neos Partners拟出售约3500万股，为该公司2月IPO以来第三次股票出售。
Oil held an advance before expected US-Iran talks in Doha, as conflicting signals from Washington and Tehran cast doubt over the next phase of negotiations to end the four-month war.
中文摘要 油价维持涨势，市场关注美国与伊朗在多哈的谈判，双方信号矛盾，为期四个月的战争能否结束存疑。
David Rodolitz, Founder and CEO of VCR Group, joined the live broadcast to discuss the recent celebration held by the New York Knicks at the Flyfish Club following their NBA Finals victory. As the owner of the Flyfish Club, Rodolitz shared insights on the significance of hosting such a high-profile
中文摘要 VCR集团创始人兼CEO David Rodolitz在直播中讨论纽约尼克斯队在Flyfish Club举办的NBA总决赛夺冠庆祝活动。
Bridge Football Group has bought Grasshopper Club Zurich, one of Switzerland’s most storied football teams.
中文摘要 Bridge Football Group已收购瑞士最负盛名的足球俱乐部之一——苏黎世草蜢队。
Long-delayed blueprint is unlikely to end military figures’ calls for more funding
中文摘要 英国首相斯塔默将公布国防计划，斥资50亿英镑用于无人机，但该长期延迟的蓝图可能仍无法满足军方对更多资金的呼吁。
12 回复 · 程序员 节点
8 回复 · 程序员 节点
18 回复 · Apple 节点
13 回复 · 程序员 节点
22 回复 · 程序员 节点
24 回复 · 程序员 节点
6 回复 · 程序员 节点
20 回复 · 程序员 节点
23 回复 · Apple 节点
23 回复 · Apple 节点
今才发现从 3.7 开始就有一个 local mode 只不过一直都是 false 10 个帖子 - 9 位参与者 阅读完整话题
老资历你什么时候能站起来啊 33 个帖子 - 33 位参与者 阅读完整话题
以下内容纯手打，AI产出内容已用截图输出 [!info] 轮到我当程序员了 说良心话，学校教科书里教的那套东西 《PowerPoint xxxx》、《C语言程序设计》，用的还是特别古早的教材，还有些偏门的已经记不起来名字，只记得当年看了想笑。 反正那会儿我对找工作，心里一点底都没有。万幸的是，有家公司租了我们学校的办公楼，搞了个校企合作，完美改变了我毕业后进厂打螺丝的生命轨迹。 [!question] 最初的梦想竟然是？ 为什么选程序员？ 我要做一个，震惊寝室的，瑟瑟聚合功能！ [!tip] 人与人的参差 也是在学的过程中，第一次体会到了什么叫人的参差。同一段代码，A同学过目不忘，还能举一反三
网易云 AI 音乐人入驻教程 1. 注册网易云音乐账号 先注册或登录一个网易云音乐账号，绑定手机号，完善头像、昵称和个人资料。昵称建议像正式音乐人名称，不要太随意，后面认证和作品展示都会用到。 2. 进入音乐人认证页面 打开网易云音乐网页端或 App，搜索“网易云音乐人”“音乐人认证”或“创作者中心”，进入音乐人入驻页面。 选择个人音乐人入驻，然后按页面提示开始认证。 3. 完成实名认证 按照要求填写真实姓名、身份证信息，并完成人脸验证。实名认证通过后，才能继续提交音乐人相关资料。 4. 先用 AI 生成一首 BGM / demo 第一首不用做得太复杂，主要用于提交 AI 音乐人认证。可以让
注意！！！不保证任何可用性，大家谨慎入手！！！注意 公益站主链：https://ai.52ccl.cn/ https://cf.52ccl.cn/ 友情提示：很多L友用主线路，导致有波动，建议可换腾讯云线路，或联通可用新加坡线路，自行轮换测试 模型：grok系列/ gpt和grok生图/ ds4pro等国模/ 少量天价gpt5.5/ claude opus4.8/等（建议生图用） 兑换：5LDC/50刀（100份）[不保证可用性，介意勿领] cdk.linux.do LINUX DO CDK Linux Do 社区 CDK 快速分享平台 - 让分享变得更简单 26 个帖子 - 26 位参与者
年会老板喊了二十分钟拥抱 AI，喊完我问能报吗，没预算，能走采购吗，没先例。 55 个帖子 - 50 位参与者 阅读完整话题
潜水社恐鼠第一次发帖，连@都研究了半天~ 还是要专门发个中奖感谢信，首先感谢L站提供这么好的平台，真的学到很多知识，现在每天空的时间都泡在L站，游戏也不玩了，觉也不睡了，深怕落后了:) 说岔了，扯远了~ 我啊从小到大连“再来一瓶”都没中过的非洲咸鱼真的一脸懵逼，第一时间农总私信我要收货地址，我都感觉是在做梦，总之非常非常非常幸运抽中农总 @Zeus 的超级大奖！ 顺带安利下佬的PackyAPI中转站：https://www.packyapi.com 稳定靠谱！ 再次由衷感谢 @Zeus 的超级福利，祝大佬越来越好。也祝各位佬友们一切顺利 回馈佬友 火山coding plan pro Anthr
惯例，先感谢羊毛提供佬:@yue357224084 白嫖京东 50W积分，可用GLM5.2 国产替代 https://joyagent.jd.com/ 2、618大促活动页-京东云 活动页面顶部点<< JoyAgent Hub Token积分包/ JoyCode - 代码洪流 一念生成 后边的抢购/查看，注意不用付款购买。 再点活动页下边的轻量云主机随便点几个抢购，注意不要付款购买。** 京东云可能会提示让实名什么的，可以不用理会，… https://www.jdcloud.com/cn/pages/cloudpromotion 点免费体验，会下载，直接取消，访问https://joyagen
省流： joy code送的50W积分 具体的方法刚刚看到这个佬友的分享 白嫖京东 50W积分，可用GLM5.2 国产替代 https://joyagent.jd.com/ 2、https://www.jdcloud.com/cn/pages/cloudpromotion 活动页面顶部点<< JoyAgent Hub Token积分包 后边的抢购，注意不用付款购买。 再点活动页下边的轻量云主机随便点几个抢购，注意不要付款购买。 京东云可能会提示让实名什么的，可以不用理会，我自己的账… 1.登录https://joyagent.jd.com/ 正常新用户登录即可 2.进入这个网页 https:/
企业Agent落地案例讲解—供应链领域 今天分享的这个主题，可落地性比较强，技术难题也不算大，如果开发得好的话是可以直接复用的。不过这个要想推进，是比较考验boss的一个决心，我等下把问题敲下来就知道为啥了。并且AI只能是辅助，最终还是需要人工干预。 供应链领域多Agent协作的一个工作流 先说业务场景吧 1.痛点： 1.1供应商出货订单跟进，顾名思义，以小米来举例，公司的业务是采购不同工厂的订单，A供应商主要供应插盘，B供应商主要供应吹风机，不同供应商会供应不同的产品，过去的业务模式是有一个订单跟进的专岗，去每天结合采购下单系统当中的数据，去导出数据，Excel处理后依次在不同的微信，钉钉，