NousResearch/hermes-agent
Python · ★ 183,095 · 🍴 31,401 · 📈 1,845 stars today
The agent that grows with you
中文介绍 Nous Research 推出的开源 AI Agent 框架，定位为「随用户成长的智能体」。基于 Hermes 系列模型构建，支持工具调用、任务规划与多步推理，适合开发者探索自主 Agent 能力，也可作为个人 AI 助手的基础框架。
2026-06-06
← 历史归档chopratejas/headroom
Python · ★ 14,487 · 🍴 921 · 📈 2,473 stars today
Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 60-95% fewer tokens, same answers. Library, proxy, MCP server.
中文介绍 专为 LLM 上下文优化设计的压缩工具，可将工具输出、日志、文件内容及 RAG 检索片段在送入模型前压缩 60–95% 的 Token 用量，同时保持回答质量不变。提供 Python 库、代理服务器和 MCP Server 三种集成方式，适合需要控制 API 成本的 RAG 应用和 Agent 系统开发者。
CopilotKit/CopilotKit
TypeScript · ★ 32,667 · 🍴 4,193 · 📈 366 stars today
The Frontend Stack for Agents & Generative UI. React + Angular. Makers of the AG-UI Protocol
中文介绍 面向 AI Agent 与生成式 UI 的前端开发栈，支持 React 和 Angular，提供 Copilot Chat、可操控 UI 组件及 Agent 状态管理能力。团队同时主导 AG-UI 协议标准制定，适合希望在 Web 应用中快速集成对话式 AI 与 Agentic 交互体验的前端开发者。
lfnovo/open-notebook
TypeScript · ★ 25,989 · 🍴 2,991 · 📈 1,152 stars today
An Open Source implementation of Notebook LM with more flexibility and features
中文介绍 Google NotebookLM 的开源替代实现，提供更高灵活性与更多功能扩展。支持多源知识整理、AI 问答与笔记生成，适合研究人员和学生管理学习资料。相比官方产品可自部署、自定义模型，满足隐私保护与定制化需求。
affaan-m/ECC
JavaScript · ★ 208,349 · 🍴 31,963 · 📈 1,361 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 针对 AI 编程 Agent 的性能优化框架，覆盖技能管理、本能行为、记忆系统与安全机制，采用研究优先的开发理念。兼容 Claude Code、Codex、Opencode、Cursor 等主流编程 Agent，帮助开发者系统化提升 Agent 的代码生成质量与任务执行稳定性。
Panniantong/Agent-Reach
Python · ★ 21,547 · 🍴 1,861 · 📈 148 stars today
Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
中文介绍 为 AI Agent 提供互联网信息获取能力的命令行工具，支持读取和搜索 Twitter、Reddit、YouTube、GitHub、Bilibili、小红书等平台内容。无需申请各平台 API Key，一条 CLI 命令即可接入多渠道数据，适合构建需要实时网络情报的 AI Agent 应用。
NVIDIA/cosmos
Jupyter Notebook · ★ 9,410 · 🍴 601 · 📈 479 stars today
NVIDIA Cosmos is an open platform of world models, datasets, and tools that enables developers to build Physical AI for robots, autonomous vehicles, smart infrastructure, and more.
中文介绍 NVIDIA 推出的物理 AI 开放平台，包含世界模型、训练数据集和开发工具链。专为机器人、自动驾驶车辆、智能基础设施等物理世界 AI 场景设计，帮助开发者训练和部署能理解物理规律的生成式模型，降低 Physical AI 研发门槛。
666ghj/MiroFish
Python · ★ 64,690 · 🍴 10,087 · 📈 320 stars today
A Simple and Universal Swarm Intelligence Engine, Predicting Anything. 简洁通用的群体智能引擎，预测万物
中文介绍 简洁通用的群体智能引擎，借鉴鱼群等自然群体行为模式，通过多智能体协作实现对任意目标的预测与模拟。设计上追求轻量与通用性，适合研究人员和开发者探索群体智能算法在预测、优化等场景下的应用。
mvanhorn/last30days-skill
Python · ★ 28,198 · 🍴 2,391 · 📈 731 stars today
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
中文介绍 AI Agent 技能模块，可针对任意话题跨越 Reddit、X（Twitter）、YouTube、Hacker News、Polymarket 及通用网页进行信息检索，自动整合近 30 天内容并生成有事实依据的结构化摘要。适合需要快速了解某领域近期动态的研究人员和信息分析师。
PaddlePaddle/PaddleOCR
Python · ★ 80,532 · 🍴 10,632 · 📈 747 stars today
Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.
中文介绍 基于 PaddlePaddle 的强大轻量级 OCR 工具包，支持将 PDF 和图片文档转换为结构化数据，覆盖 100 余种语言。作为图像/文档与 LLM 之间的桥梁，广泛应用于文档智能、数据提取和 RAG 知识库构建，适合企业和开发者处理大规模文档解析任务。
openai/plugins
JavaScript · ★ 1,529 · 🍴 239 · 📈 49 stars today
OpenAI Plugins
中文介绍 OpenAI 官方 Plugins 仓库，用于管理和发布 ChatGPT 插件相关规范与示例代码。开发者可参考此仓库了解插件接口协议、Manifest 格式及 API 集成方式，适合希望为 ChatGPT 扩展外部工具能力的开发者和企业。
MemPalace/mempalace
Python · ★ 53,874 · 🍴 7,079 · 📈 227 stars today
The best-benchmarked open-source AI memory system. And it's free.
中文介绍 基准测试成绩领先的开源 AI 记忆系统，免费可用。为 AI Agent 和对话应用提供长期记忆存储与检索能力，解决大模型上下文窗口有限导致的记忆丢失问题，适合构建需要持久化用户记忆的个人助手、客服 Bot 等应用场景。
withastro/flue
TypeScript · ★ 4,512 · 🍴 241 · 📈 126 stars today
The sandbox agent framework.
中文介绍 由 Astro 团队开发的沙箱 Agent 框架，在隔离的沙箱环境中运行 AI Agent 任务，保障执行安全性。适合需要对 Agent 行为进行严格限制和环境隔离的场景，降低 AI 自主执行代码或操作时带来的系统风险。
openclaw/openclaw-windows-node
C# · ★ 1,598 · 🍴 181 · 📈 326 stars today
Windows companion suite for OpenClaw - System Tray app, Shared library, Node, and PowerToys Command Palette extension
中文介绍 OpenClaw 项目的 Windows 配套工具套件，包含系统托盘应用、共享库、Node 组件以及 PowerToys Command Palette 扩展，方便 Windows 用户以桌面原生方式集成和管理 OpenClaw Agent 节点，提升在 Windows 环境下的使用体验。
aquasecurity/trivy
Go · ★ 35,851 · 🍴 443 · 📈 207 stars today
Find vulnerabilities, misconfigurations, secrets, SBOM in containers, Kubernetes, code repositories, clouds and more
中文介绍 Aqua Security 出品的全面安全扫描工具，可检测容器镜像、Kubernetes 集群、代码仓库、云环境中的漏洞、配置错误、敏感信息泄露及 SBOM 组件清单。支持 CI/CD 集成，是 DevSecOps 流程中广泛使用的开源安全审计利器。
jwasham/coding-interview-university
★ 350,380 · 🍴 83,289 · 📈 745 stars today
A complete computer science study plan to become a software engineer.
中文介绍 面向软件工程师求职的完整计算机科学自学计划，涵盖数据结构、算法、系统设计、操作系统等核心知识点，并附有学习资源和练习题。由作者自学后成功入职大厂的经历整理而来，是全球开发者准备技术面试的经典参考仓库。
github/copilot-sdk
Java · ★ 9,236 · 🍴 1,223 · 📈 309 stars today
Multi-platform SDK for integrating GitHub Copilot Agent into apps and services
中文介绍 GitHub 官方发布的多平台 SDK，用于将 GitHub Copilot Agent 能力集成到第三方应用和服务中。支持跨平台调用 Copilot 的代码生成、对话及 Agent 功能，适合企业开发者在自有 IDE 插件、Web 应用或工作流工具中嵌入 Copilot 智能体能力。
Regret Minimization with Adaptive Opponents in Repeated Games
👍 1
In this paper, we study regret minimization in repeated games with adaptive opponents who can respond based on histories of play. The standard metric of external regret in online learning is known to fail to capture such adaptivity. To account for players' counterfactual reasoning, we introduce {\tt
Learning Geometric Representations from Videos for Spatial Intelligent Multimodal Large Language Models
👍 3
Multimodal Large Language Models (MLLMs) excel at 2D semantic understanding but lack intrinsic 3D awareness, resulting in representations that fail to maintain geometric and spatial consistency across video frames. Given the scarcity of large-scale 3D data, we present GeoVR, a novel framework that l
AffordanceVLA: A Vision-Language-Action Model Empowering Action Generation through Affordance-Aware Understanding
👍 3
Vision-Language-Action (VLA) models leverage the rich world knowledge of pretrained vision-language models (VLMs) to enable instruction-following robotic manipulation. However, the structural mismatch between VLM semantic spaces and embodied control policies often hinders the learning of precise per
Code2LoRA: Hypernetwork-Generated Adapters for Code Language Models under Software Evolution
👍 45
Code language models need repository-level context to resolve imports, APIs, and project conventions. Existing methods inject this knowledge as long inputs (retrieved through RAG or dependency analysis) or through per-repository fine-tuning and LoRA -- costly at repository scale and brittle to evolv
AURA: Intent-Directed Probing for Implicit-Need Surfacing in Situated LLM Agents
👍 1
A situated query like "where is Lin Wei?" often encodes more than its literal content: the user may also want to know whether Lin Wei is free, in a good mood, or worth interrupting now. Standard tool-use agents answer the literal question and stop. AURA inserts an inference step between scene percep
Benchmark Everything Everywhere All at Once
👍 1
Benchmarks are fundamental for evaluating and advancing LLMs and MLLMs by providing standardized and explicit measures of performance. However, their construction is labor-intensive and hard to reuse, raising concerns about sustainability and scalability. Moreover, existing benchmarks often quickly
Revising Context, Shifting Simulated Stance: Auditing LLM-Based Stance Simulation in Online Discussions
👍 2
Large language models are increasingly used to simulate social media users and infer how individuals may respond to online discussions. However, it remains unclear whether these simulations reflect precise user-specific beliefs or whether they are highly sensitive to semantically independent changes
ForeSci: Evaluating LLM Agents for Forward-Looking AI Research Judgment
👍 0
AI research often requires decisions before future evidence exists: which bottleneck to attack, which direction to pursue, or where a project should be positioned. We introduce ForeSci, a temporally controlled benchmark for evaluating whether LLM agents can make such forward-looking research judgeme
Dream.exe: Can Video Generation Models Dream Executable Robot Manipulation?
👍 12
Video generation models have made impressive strides in synthesizing visually compelling content, yet their outputs remain confined to the virtual domain. A natural question follows: how well do these models reflect the physical world when their generated videos leave the screen and enter reality? W
Towards One-to-Many Temporal Grounding
👍 4
Temporal Grounding (TG) aims to localize video segments corresponding to a textual query. Prior research predominantly focuses on single-segment retrieval. Real-world scenarios, however, often require localizing multiple disjoint segments for a single query -- a setting we term One-to-Many Temporal
LLMs Can Leak Training Data But Do They Want To? A Propensity-Aware Evaluation of Memorization in LLMs
👍 7
Large language models can reproduce training data, but existing memorization evaluations mostly measure whether models can be forced to do so, rather than whether they do so under ordinary use. We introduce PropMe, a propensity-aware framework for memorization evaluation that contrasts prefix-based
AdaPlanBench: Evaluating Adaptive Planning in Large Language Model Agents under World and User Constraints
👍 32
Planning for real-world problems by language models often involves both world and user constraints, which may not be fully specified upfront and are progressively disclosed through interaction. However, existing benchmarks still underexplore adaptive planning under such progressively revealed dual c
ArcANE: Do Role-Playing Language Agents Stay in Character at the Right Time?
👍 40
Role-playing language agents (RPLAs) should play characters whose values and behavior evolve as the story progresses, not maintain a fixed persona. Existing benchmarks measure factual recall at a given chapter, not whether responses align with the character's psychological trajectory, especially in
Reinforcement Learning Elicits Contextual Learning of Unseen Language Translation
👍 23
Prior work has shown that large language models (LLMs) can translate unseen or low-resource languages by undergoing continued training or even by encoding a grammar book in their context. However, both methods typically overfit specific languages, with limited zero-shot transfer at test time. To tra
TIDE: Proactive Multi-Problem Discovery via Template-Guided Iteration
👍 36
Agents are widely deployed as assistants over documents, tools, and code. However, they typically act only on explicit user requests, which surface only the problems the user has noticed, while many other important problems coexist, hidden in plain sight, within the broader user context, with their
MAOAM: Unified Object and Material Selection with Vision-Language Models
👍 6
Selection is a core operation in interactive image editing. To be practical, a user should be able to specify and disambiguate the desired selection region through either text or click-based interactions, and the system should support selecting not only objects but also other criteria, such as mater
SEAOTTER: Sensor Embedded Autoencoding with One-Time Transcode for Efficient Reconstruction
👍 3
In robotics systems, vast amounts of visual data are easily captured at high resolution using low-cost, low-power hardware. Yet, limited bandwidth and on-device compute resources prevent full utilization when transmitted via conventional codecs like JPEG/MPEG. Newer codecs, like AV1/AVIF, improve th
RobotValues: Evaluating Household Robots When Human Values Conflict
👍 22
While household robots are often evaluated based on task completion, everyday domestic environments involve value-conflicting situations in which robots are expected to choose actions that prioritize other values than task success, such as human autonomy, efficiency, or social appropriateness. Yet,
EvoDS: Self-Evolving Autonomous Data Science Agent with Skill Learning and Context Management
👍 2
Recent progress in Large Language Model (LLM) agents has enabled promising advances in automated data science. However, existing approaches remain fundamentally limited by their static action sets and lack of principled long-horizon context management, hindering their ability to accumulate reusable
LLM Anonymization Against Agentic Re-Identification
👍 1
Agentic LLMs with web search change the threat model for text anonymization: weak contextual cues can become cross-referenceable evidence for re-identification, yet those same details also carry downstream analytic value of the text. Existing defenses either remove explicit identifiers, perturb text
Absorbing Complexity: An Interaction-Native Knowledge Harness for Financial LLM Agents
👍 1
Financial AI agents often fail for a simple reason: they make users carry the complexity. A user must repeatedly restate goals, risk preferences, portfolio context, past judgments, and shifting market assumptions, while the agent answers, retrieves, acts, and forgets. In finance, this is not just in
AdaCodec: A Predictive Visual Code for Video MLLMs
👍 3
Video is temporally redundant: adjacent frames usually share most objects, background, and layout. Yet existing video multimodal large language models (video MLLMs) usually encode each sampled frame as an independent RGB image, causing visual tokens to repeat content already present in earlier frame
SABER: Benchmarking Operational Safety of LLM Coding Agents in Stateful Project Workspaces
👍 0
Large language models are increasingly deployed as coding agents, shifting safety from individual responses to action sequences. Existing benchmarks, however, primarily assess whether models refuse unsafe prompts, leaving impacts on stateful workspaces largely unexamined. We present SABER, a benchma
The Shape of Addition: Geometric Structures of Arithmetic in Large Language Models
👍 1
Large Language Models exhibit paradoxical fragility in fundamental arithmetic, implying a disconnect between internal computation and discrete output. By analyzing the residual stream geometry during multi-operand addition, we identify the Iso-Raw-Sum Trajectory (IRST), a geometric structure where r
MechVQA: Benchmarking and Enhancing Multimodal LLMs on Comprehensive Mechanical Drawing Understanding
👍 3
Multimodal Large Language Models (MLLMs) have demonstrated significant achievements in general visual question answering (VQA) tasks. However, they remain brittle on mechanical engineering drawings, where high annotation density and weak domain knowledge, compounded by unreliable spatial relation re
Combinatorial Synthesis: Scaling Code RLVR via Atomic Decomposition and Recombination
👍 1
Reinforcement Learning with Verifiable Rewards (RLVR) has recently emerged as the cornerstone for shaping the remarkable coding abilities of Large Language Models (LLMs). However, the scalability of RLVR is severely constrained by the scarcity of sufficiently challenging verifiable code tasks that t
Multimodal Music Recommendation System using LLMs
👍 1
Music recommendation systems typically treat songs as opaque tokens, relying on collaborative interaction histories which overlooks semantic or acoustic content. Prior work has explored LLM-augmented, multimodal, and text-enhanced approaches to sequential recommendation, and while some methods parti
Meta-Cognitive Memory Policy Optimization for Long-Horizon LLM Agents
👍 5
Memory-augmented LLM agents tackle complex long-horizon tasks by recursively summarizing interaction trajectories into compact memory. However, existing approaches typically train these memory policies using outcome-based reinforcement learning, failing to localize where intermediate memory quality
Trust Region Q Adjoint Matching
👍 1
Off-policy reinforcement learning of pretrained flow policies remains challenging due to the instability of optimization arising from the multi-step sampling process. Recently, Q-learning with Adjoint Matching (QAM) addressed this issue by reformulating into a memoryless stochastic optimal control (
Is This Edit Correct? A Multi-Dimensional Benchmark for Reasoning-Aware Image Editing
👍 1
Diffusion-based image editing has achieved strong visual fidelity under natural language instructions, yet most existing systems still operate at the level of surface instruction following, without reasoning about the implicit contextual constraints embedded in real user requests. This often leads t
SpaceX IPOs in 7 days. I Fed the S1 Doc Into Claude. Here Is What It Found Buried in 300 Pages.
@DamiDefi · 96.5K 粉丝 · 2.3M 阅 · 584 赞 · 80 转
The number that stopped me was not the $2 trillion valuation. It was $791 million. That is what SpaceX made in net income in 2024. A profitable, growing aerospace company with a genuine moat in launch
中文介绍 SpaceX 即将 IPO，作者将 300 页 S1 招股书喂给 Claude 进行深度分析。关键数字：2024 年净利润 7.91 亿美元，估值 2 万亿美元。重点挖掘了埋藏在文件深处的财务细节，包括其在航天发射领域的护城河与盈利能力，适合想快速读懂复杂财报的投资者参考。
How to master Dynamic Workflows in Claude Code: 6 patterns and 14 steps Anthropic engineers actually
@0xCodez · 3.3K 粉丝 · 637.2K 阅 · 510 赞 · 59 转
Most Claude Code users still write their workflows by hand. They chain prompts, copy outputs, paste them into the next prompt, fix what went wrong, repeat. 9 out of 10 builders haven’t tried Dynamic
中文介绍 针对 Claude Code 动态工作流的深度教程，总结 Anthropic 工程师实际采用的 6 种模式和 14 个步骤。指出 90% 的开发者仍在手动串联 prompt、复制粘贴输出，尚未尝试动态工作流。内容面向希望摆脱重复劳动、让 Claude Code 自动驱动复杂流程的构建者。
Generative UI Is the New Frontend
@Saboo_Shubham_ · 116.2K 粉丝 · 263.3K 阅 · 517 赞 · 74 转
The frontend used to be a fixed thing. Designers drew it. Engineers built it. Users got what shipped. That's over. The interfaces shipping in 2026 are drawn partly by the agent itself, in real time,
中文介绍 探讨「生成式 UI」正在取代传统前端范式：2026 年的界面不再由设计师固定绘制，而是由 Agent 在运行时实时生成。这一趋势意味着前端开发逻辑、设计流程和用户体验都将被重构，作者认为这是前端领域的根本性转变而非渐进式迭代。
A guide to /goal 🥅
@dkundel · 19.3K 粉丝 · 116.9K 阅 · 523 赞 · 40 转
We launched the goal mode (or /goal) as a way to help you have Codex drive towards a concrete outcome. When you set a goal Codex will continue to work until the goal is achieved, whether that takes
中文介绍 Codex 新功能 `/goal`（目标模式）使用指南。设定目标后，Codex 会持续自主工作直到达成具体结果，无需人工反复介入。文章介绍该模式的设计理念与使用技巧，适合希望让 Codex 完成长周期、端到端任务的开发者。
A Functional Taxonomy of World Models
@drfeifei · 738.0K 粉丝 · 72.2K 阅 · 699 赞 · 144 转
“The world is everything that is the case.” — Ludwig Wittgenstein, Tractatus Logico-Philosophicus, 1921 The world is not made of words. In an earlier essay, we argued that spatial intelligence is AI’s
中文介绍 李飞飞分享关于「世界模型」的功能性分类框架，延续其空间智能是 AI 下一边界的论点。文章以维特根斯坦名言开篇，强调世界不由语言构成，探讨 AI 系统如何建立超越文字的世界表征，是对世界模型概念的系统性理论梳理。
How to Build a Custom Agent Harness
@sydneyrunkle · 7.5K 粉丝 · 69.5K 阅 · 511 赞 · 74 转
Building useful agents is largely about customization: connecting your agent to the right context, data, and environment(s) for the task at hand. At its core, an agent is a model calling tools in a
中文介绍 构建自定义 Agent Harness 的实践教程，核心思路是：Agent 本质上是模型在循环中调用工具，真正的价值在于为其接入正确的上下文、数据和运行环境。文章介绍如何针对具体任务定制 Agent 框架，适合希望超越通用 Agent、打造专用系统的开发者。
Your token spend is an AI architecture problem, not just a model problem
@jainarvind · 9.3K 粉丝 · 53.7K 阅 · 505 赞 · 68 转
Enterprise AI token spend is scaling quickly, especially as the technology shifts from simple chat assistants into coding agents, AI coworkers, and long-running workflows. These systems do far more
中文介绍 企业 AI Token 消耗快速攀升，作者指出这本质上是架构问题而非模型选型问题。随着 AI 从聊天助手演进为编程 Agent、AI 协作者和长周期工作流，Token 使用模式发生根本变化。文章提供从架构层面控制 Token 开支的思路，适合构建大规模企业 AI 系统的团队。
10 HERMES AGENT HACKS THAT TURNED MY CHAT AGENT INTO A 24/7 SYSTEM
@IBuzovskyi · 1.2K 粉丝 · 50.1K 阅 · 500 赞 · 51 转
These 10 Hermes Agent hacks saved me 15+ hours every week - and they work for any workflow you run repeatedly. Content, software development, business operations, client management, research, sales.
中文介绍 分享 10 个 Hermes Agent 使用技巧，作者称这些方法每周为其节省 15 小时以上。适用场景覆盖内容创作、软件开发、业务运营、客户管理、研究和销售等重复性工作流，核心目标是将聊天式 Agent 升级为可持续运转的 24/7 自动化系统。
Feedback loops: Help Claude Code complete ambitious tasks with less babysitting
@delba_oliveira · 74.0K 粉丝 · 37.4K 阅 · 533 赞 · 38 转
As we delegate more ambitious tasks to Claude, it becomes increasingly important that it can verify its own work. The more Claude can self-verify: the more independently it can work on long-running
中文介绍 探讨如何通过「反馈闭环」让 Claude Code 在执行复杂任务时减少人工监督。核心观点：Claude 自我验证能力越强，越能独立完成长周期任务。文章介绍具体方法，帮助开发者构建能自检、自纠的 Claude Code 工作流，适合希望委托 Agent 完成高难度任务的用户。
Don't let your agent guess, give it runtime context
@ericzakariasson · 67.9K 粉丝 · 37.3K 阅 · 507 赞 · 27 转
If you've ever watched an agent try to fix a bug, you've watched it guess. It reads the code, comes up with a theory, makes an edit, and hopes. Sometimes it's right. A lot of the time you get a fix
中文介绍 针对 Agent 调试低效问题，作者指出根因在于 Agent 缺乏运行时上下文、只能靠猜。解决方案是在运行时主动注入上下文信息，让 Agent 基于真实状态而非推断来定位和修复 bug。文章提供实操建议，适合正在构建或优化 AI 编程 Agent 的开发者。
Thousand Token Wood: shipping a multi-agent economy on a 3B model
中文介绍 「Thousand Token Wood」项目展示了在30亿参数小型模型上构建多智能体经济模拟系统的实践成果，探索了小规模模型在复杂多智能体协作场景中的应用潜力。
How to Stop Shipping Low-Quality RL Environments (with Examples)
Your broken harness is actively making the model worse. Here's what I keep seeing after years of eyeballing trajectories, and what you need to fix.
中文介绍 文章指出当前强化学习环境质量低劣的问题，认为有缺陷的训练环境会直接损害模型性能。作者基于多年轨迹分析经验，梳理了常见问题类型，并提供了改进建议，强调高质量RL环境对模型训练质量的关键影响。
The Meta hack shows there’s more to AI security than Mythos
On June 5, 404 Media reported that attackers had been using Meta’s AI customer support agent to steal Instagram accounts. Their approach was simple: They asked the agent to link the accounts to email addresses that they controlled, and the agent complied. One attacker broke into the dormant Obama Wh
中文介绍 据404 Media于6月5日报道，攻击者利用Meta的AI客服代理窃取Instagram账户。攻击手法简单：直接要求该AI将账户绑定至攻击者控制的邮箱，AI照单执行。此事件暴露了AI智能体在安全防护层面的重大漏洞，引发业界对AI安全机制的广泛关注。
[AINews] not much happened today
a quiet day
中文介绍 AI领域当日动态较为平淡，无重大事件或重要发布，整体为平静的一天。
Anthropic Oceanus leaks 🤖, ChatGPT Dreaming 💭, recursive self improvement 🚀
中文介绍 当日AI摘要涵盖三大要点：Anthropic新模型Oceanus信息疑似泄露；OpenAI为ChatGPT推出「Dreaming」记忆新功能；以及关于AI递归自我改进（recursive self-improvement）的最新进展报道。
Reality: The Final Eval — Lukas Petersson and Axel Backlund of Andon Labs
We talk with the VendingBench authors on evaling Claudes from Haiku to Mythos, and how they build leading, and lasting, frontier evals from scratch.
中文介绍 Latent Space播客专访Andon Labs的Lukas Petersson与Axel Backlund，两位作者介绍了「VendingBench」评测框架的构建过程，探讨如何对Claude从Haiku到Mythos系列模型进行系统性评估，并分享了从零打造前沿、持久有效评测基准的方法论。
Nemotron 3.5 Content Safety: Customizable Multimodal Safety for Global Enterprise AI
中文介绍 英伟达发布Nemotron 3.5内容安全模型，面向全球企业AI场景提供可定制化的多模态内容安全解决方案，支持企业根据不同地区和业务需求灵活配置安全策略。
EVA-Bench Data 2.0: 3 Domains, 121 Tools, 213 Scenarios
中文介绍 ServiceNow AI发布EVA-Bench Data 2.0评测数据集，覆盖3大领域、121种工具、213个应用场景，旨在为AI智能体在企业级复杂任务中的能力评估提供更全面的基准测试支持。
How Endava is redesigning software delivery around AI agents
Learn how Endava is using AI agents, ChatGPT Enterprise, and Codex to accelerate software delivery, automate workflows, and build an AI-native culture across the enterprise.
中文介绍 IT服务商Endava介绍其围绕AI智能体重构软件交付流程的实践，通过部署ChatGPT Enterprise和Codex加速软件开发、自动化工作流，并在企业内部推动AI原生文化建设，实现研发效能提升。
How courts are coping with a flood of AI-generated lawsuits
Most days in her chambers, Judge Maritza Braswell, a federal magistrate judge in Colorado, sifts through stacks of documents written by people without a lawyer. Many of them can’t afford to hire a lawyer, and others have cases too weak or too small to interest one. She reads each one carefully, mind
中文介绍 美国法院正面临AI生成诉状引发的诉讼洪流。科罗拉多州联邦法官Maritza Braswell每日须审阅大量无律师代理当事人提交的文件，AI工具的普及使起诉门槛降低，法院系统承压加剧，司法效率与诉讼质量管理面临新挑战。
Dreaming: Better memory for a more helpful ChatGPT
ChatGPT introduces a new memory system to better remember preferences, keeping context fresh and relevant across conversations.
中文介绍 OpenAI为ChatGPT推出名为「Dreaming」的全新记忆系统，能更好地记住用户偏好，确保跨对话上下文保持新鲜与相关性，旨在提升ChatGPT的个性化服务能力和长期对话连贯性。
not much happened today
**NVIDIA** released **Nemotron 3 Ultra**, a fully open **550B MoE** model with **55B active parameters** and **1M context**, optimized for long-running agent tasks with up to **5x speedup** and **30% cost reduction**. It features hybrid Mamba/attention, LatentMoE, native MTP, and was pretrained on *
中文介绍 英伟达发布Nemotron 3 Ultra，这是一款完全开源的5500亿参数MoE模型，激活参数550亿，支持100万上下文长度，专为长时智能体任务优化，较前代实现最高5倍加速、30%成本降低，采用混合Mamba/注意力架构及LatentMoE等技术。
[AINews] Reve 2 and Ideogram 4: Layouts in Imagegen
a quiet day.
中文介绍 图像生成领域迎来新进展：Reve 2与Ideogram 4相继发布，后者重点引入布局控制（Layouts）功能，支持用户对图像构图进行更精细的控制。当日整体AI动态较为平淡。
Biodefense in the Intelligence Age
An action plan for AI-powered biological resilience
中文介绍 OpenAI发布「智能时代的生物防御」行动方案，探讨如何利用AI技术提升生物安全韧性，提出以AI赋能生物威胁监测、防御与应对的系统性规划框架。
Designing the hf CLI as an agent-optimized way to work with the Hub
中文介绍 Hugging Face介绍其命令行工具hf CLI的设计理念，重点阐述如何将CLI打造为针对AI智能体优化的Hub交互方式，使智能体能够更高效地访问、管理和部署Hub上的模型与数据资源。
本交易日各类资产技术面普遍承压，风险情绪指标明显恶化。美股方面，SPY单日下跌2.58%，QQQ跌幅达4.80%并触发「MACD死叉」，纳指科技权重股集体走弱：NVDA单日-6.20%、META单日-5.51%且空头排列，TSLA单日-6.56%并已形成空头排列。尽管SPY/QQQ的SMA50与SMA200仍维持多头排列，但MACD柱状图持续收缩，短期动量明显减弱，RSI14均已回落至50附近（SPY=49.4，QQQ=48.3），中短期趋势支撑有待确认。加密市场技术面极度疲弱：加密恐慌贪婪指数仅为12（极度恐慌），加密总市值24小时蒸发3.7%至2.19万亿美元，BTC主导率56.1%。BTC报61,255美元，RSI14=15.7；ETH报1,593美元，RSI14=13.2；SOL报64.35美元，RSI14=16.8，三者均触发「RSI超卖」并接近52周低位，空头排列结构完整，MACD柱状图负值加速扩张。中概股整体处于空头排列，阿里巴巴、拼多多、京东RSI均低于42，腾讯控股虽出现MACD金叉但中长期均线压制显著。商品与外汇方面，黄金期货报4353.90美元，RSI=35且MACD持续负值，短期动量偏弱；WTI原油技术面中性；美元指数DXY报100.07，接近52周高点，RSI=66.1，多头排列结构最为完整；10年期美债收益率升至4.54%，SMA多头排列。VIX单日飙升39.68%至21.51，触发金叉，市场避险氛围显著升温。
今日关注
ETH-USD
以太坊 (ETH-USD)
偏下行
以太坊当前价格1593.05美元，RSI14仅13.2，处于深度超卖区间，距52周高点跌幅高达67.84%。MACD为-121.9964，信号线为-84.1425，柱状图持续扩大负值，空头排列明确。5日跌幅-20.52%，三条均线（SMA20=2009.31、SMA50=2190.42、SMA200=2464.49）全部大幅压制于价格上方，技术面呈完整空头结构，接近52周低点。
TSLA
特斯拉 (TSLA)
偏下行
特斯拉当前报价391.00美元，1日跌幅-6.56%，5日跌幅-10.28%，距52周高点回落21.62%。触发「MACD死叉」与「空头排列」双重信号，SMA200=414.14高于SMA50=395.29，价格已跌破两条短期均线，RSI14=40.4接近超卖边界，trend标记为「bearish」。MACD=4.143但信号线=8.6074，差值持续扩大向下，动量结构偏弱。
^VIX
VIX 恐慌指数
偏上行
VIX当前报21.51，1日暴涨39.68%，5日累计涨幅40.40%，已大幅突破SMA20=17.12与SMA50=18.99，触发「MACD金叉」与「多头排列」。RSI14=65.7处于偏强区间但尚未超买。VIX快速拉升反映市场避险情绪大幅升温，与同期权益资产全面走弱形成对应，技术结构呈短期上行态势。
DX-Y.NYB
美元指数 DXY
偏上行
美元指数报100.07，接近52周高点（距高点仅-0.57%），触发「接近52周高」与「多头排列」信号。SMA20=99.02、SMA50=98.91、SMA200=98.60三线多头排列且均位于价格下方，RSI14=66.1处于偏强区间。MACD=0.2468高于信号线0.1608，柱状图正值扩张，1日+0.66%、5日+1.17%，短期上行动量持续。
0700.HK
腾讯控股 (0700.HK)
中性
腾讯控股报453.20港元，5日涨幅+6.09%表现相对抗跌，触发罕见「MACD金叉」（MACD=-7.2197上穿信号线-11.0545），但整体仍处空头排列，SMA50=476.94与SMA200=572.08均压制于价格上方。RSI14=47.1居中，距52周高点仍回落33.65%，金叉信号与中长期下行趋势形成矛盾，技术面处于多空拉锯的中性状态。
全部资产
^VIX
VIX 恐慌指数
$21.51
+39.68%
- 5 日
- +40.40%
- 距 52w 高
- -39.1%
- RSI(14)
- 65.7
- 趋势
- 多头
- SMA 20 / 50 / 200
- 17.12 / 18.99 / 18.42
- MACD / 信号
- -0.395 / -0.773
MACD 金叉 (今天)多头排列
^TNX
10Y 美债收益率 (%)
$4.54
+1.32%
- 5 日
- +1.86%
- 距 52w 高
- -9.2%
- RSI(14)
- 57.7
- 趋势
- 多头
- SMA 20 / 50 / 200
- 4.50 / 4.40 / 4.20
- MACD / 信号
- 0.027 / 0.037
多头排列
DX-Y.NYB
美元指数 DXY
$100.07
+0.66%
- 5 日
- +1.17%
- 距 52w 高
- -0.6%
- RSI(14)
- 66.1
- 趋势
- 多头
- SMA 20 / 50 / 200
- 99.02 / 98.91 / 98.60
- MACD / 信号
- 0.247 / 0.161
接近 52 周高多头排列
SPY
S&P 500 ETF
$737.55
-2.58%
- 5 日
- -2.50%
- 距 52w 高
- -3.0%
- RSI(14)
- 49.4
- 趋势
- 多头
- SMA 20 / 50 / 200
- 746.29 / 713.51 / 683.93
- MACD / 信号
- 9.981 / 12.063
多头排列
QQQ
Nasdaq 100 ETF
$705.06
-4.80%
- 5 日
- -4.50%
- 距 52w 高
- -5.8%
- RSI(14)
- 48.3
- 趋势
- 多头
- SMA 20 / 50 / 200
- 722.01 / 667.81 / 621.82
- MACD / 信号
- 17.425 / 20.696
MACD 死叉 (1 天前)多头排列
AAPL
Apple
$307.34
-1.25%
- 5 日
- -1.51%
- 距 52w 高
- -3.0%
- RSI(14)
- 60.7
- 趋势
- 多头
- SMA 20 / 50 / 200
- 304.25 / 281.24 / 265.19
- MACD / 信号
- 8.464 / 9.395
MACD 死叉 (2 天前)多头排列
MSFT
Microsoft
$416.67
-2.66%
- 5 日
- -7.46%
- 距 52w 高
- -25.0%
- RSI(14)
- 47.5
- 趋势
- 中性
- SMA 20 / 50 / 200
- 422.58 / 408.35 / 456.38
- MACD / 信号
- 5.540 / 6.187
MACD 死叉 (今天)
NVDA
Nvidia
$205.10
-6.20%
- 5 日
- -2.86%
- 距 52w 高
- -13.3%
- RSI(14)
- 43.8
- 趋势
- 多头
- SMA 20 / 50 / 200
- 219.10 / 203.45 / 188.57
- MACD / 信号
- 2.300 / 4.259
多头排列
GOOGL
Alphabet
$368.53
-0.98%
- 5 日
- -3.11%
- 距 52w 高
- -9.8%
- RSI(14)
- 46.4
- 趋势
- 多头
- SMA 20 / 50 / 200
- 385.38 / 354.50 / 304.04
- MACD / 信号
- 1.703 / 6.855
多头排列
TSLA
Tesla
$391.00
-6.56%
- 5 日
- -10.28%
- 距 52w 高
- -21.6%
- RSI(14)
- 40.4
- 趋势
- 空头
- SMA 20 / 50 / 200
- 425.87 / 395.29 / 414.14
- MACD / 信号
- 4.143 / 8.607
MACD 死叉 (4 天前)空头排列
META
Meta
$593.00
-5.51%
- 5 日
- -6.25%
- 距 52w 高
- -25.5%
- RSI(14)
- 41.6
- 趋势
- 空头
- SMA 20 / 50 / 200
- 612.72 / 619.52 / 662.45
- MACD / 信号
- -3.755 / -3.493
MACD 死叉 (今天)空头排列
BTC-USD
Bitcoin
$61,255.49
-3.99%
- 5 日
- -16.75%
- 距 52w 高
- -51.5%
- RSI(14)
- 15.7
- 趋势
- 空头
- SMA 20 / 50 / 200
- 73,203.15 / 76,385.61 / 78,778.49
- MACD / 信号
- -3,489.543 / -1,983.404
RSI 超卖接近 52 周低空头排列
ETH-USD
Ethereum
$1,593.05
-9.98%
- 5 日
- -20.52%
- 距 52w 高
- -67.8%
- RSI(14)
- 13.2
- 趋势
- 空头
- SMA 20 / 50 / 200
- 2,009.31 / 2,190.42 / 2,464.49
- MACD / 信号
- -121.996 / -84.143
RSI 超卖接近 52 周低空头排列
SOL-USD
Solana
$64.35
-6.36%
- 5 日
- -21.81%
- 距 52w 高
- -74.6%
- RSI(14)
- 16.8
- 趋势
- 空头
- SMA 20 / 50 / 200
- 81.14 / 85.04 / 102.79
- MACD / 信号
- -4.460 / -2.504
RSI 超卖接近 52 周低空头排列
BABA
阿里巴巴 (BABA)
$121.06
-3.88%
- 5 日
- -2.54%
- 距 52w 高
- -37.2%
- RSI(14)
- 37.2
- 趋势
- 空头
- SMA 20 / 50 / 200
- 131.73 / 131.10 / 149.72
- MACD / 信号
- -2.624 / -1.646
空头排列
PDD
拼多多 (PDD)
$85.07
-0.94%
- 5 日
- +0.75%
- 距 52w 高
- -39.0%
- RSI(14)
- 36.3
- 趋势
- 空头
- SMA 20 / 50 / 200
- 92.48 / 97.24 / 112.45
- MACD / 信号
- -3.706 / -2.987
空头排列
JD
京东 (JD)
$28.88
-1.06%
- 5 日
- +0.17%
- 距 52w 高
- -21.6%
- RSI(14)
- 41.3
- 趋势
- 空头
- SMA 20 / 50 / 200
- 30.69 / 30.16 / 30.38
- MACD / 信号
- -0.352 / -0.077
空头排列
0700.HK
腾讯控股 (0700.HK)
HK$453.20
-1.26%
- 5 日
- +6.09%
- 距 52w 高
- -33.6%
- RSI(14)
- 47.1
- 趋势
- 空头
- SMA 20 / 50 / 200
- 451.94 / 476.94 / 572.08
- MACD / 信号
- -7.220 / -11.055
MACD 金叉 (3 天前)空头排列
GC=F
黄金期货
$4,353.90
-2.72%
- 5 日
- -4.53%
- 距 52w 高
- -22.1%
- RSI(14)
- 35.0
- 趋势
- 中性
- SMA 20 / 50 / 200
- 4,546.88 / 4,624.40 / 4,398.63
- MACD / 信号
- -63.668 / -54.120
CL=F
WTI 原油期货
$90.25
-3.00%
- 5 日
- +3.31%
- 距 52w 高
- -24.5%
- RSI(14)
- 43.1
- 趋势
- 中性
- SMA 20 / 50 / 200
- 96.75 / 97.86 / 72.79
- MACD / 信号
- -1.721 / -1.020
USDCNY=X
美元 / 人民币
¥6.76
-0.13%
- 5 日
- -0.02%
- 距 52w 高
- -6.2%
- RSI(14)
- 34.2
- 趋势
- 空头
- SMA 20 / 50 / 200
- 6.79 / 6.82 / 6.98
- MACD / 信号
- -0.015 / -0.015
接近 52 周低空头排列
本报告仅基于公开行情数据的技术指标进行客观描述，过去走势不代表未来表现，所有技术信号均存在失效可能。本内容仅供技术指标解读参考，不构成任何投资建议或买卖推荐，投资者须自行承担决策风险。
Regrettable references and claims of ‘rigged’ election laws: why this week has reignited Jacinta Allan spill rumours
Just months from the Victorian election, the premier’s performance has left some MPs wondering if it’s too late for Labor to change leaders Get our breaking news email, free app or daily news podcast Jacinta Allan faced three major tests this week. The way she handled them has left some of her colle
中文摘要 距维多利亚州选举仅数月，澳大利亚维州工党州长贾辛塔·艾伦本周在三项重大考验中的表现引发党内议员对其领导层更迭的担忧，有关换帅的传言再度升温。部分议员质疑选举法规存在"操纵"嫌疑，党内对现任领导层的信心下滑。
US says Iran radar sites struck and drones intercepted, in latest threat to fragile ceasefire
Iran and the US have exchanged a series of attacks near the strait of Hormuz, imperilling efforts to reach a peace deal The US military said it shot down four Iranian drones that were launched toward the strait of Hormuz and struck coastal surveillance radar sites in response. “The attack drones pos
中文摘要 美国军方表示，已击落伊朗发射的四架无人机，目标位于霍尔木兹海峡附近，并随后打击了伊朗沿海监视雷达站。美伊双方在霍尔木兹海峡附近发生一系列交火，严重威胁当前脆弱的停火局面及双边和平谈判进程。
A Sherpa Survived 6 Days Alone on Everest. His Family Says He Was Abandoned.
Dawa Sherpa, 57, was found alive on Thursday, nearly a week after he was last seen on the mountain. His wife says more could have been done to find him sooner.
中文摘要 57岁夏尔巴向导达瓦·夏尔巴在珠穆朗玛峰失踪近六天后于周四被发现，奇迹生还。其妻子称，搜救工作启动过迟，本可更早找到他，质疑相关团队存在遗弃行为。
Pamela Hicks, Lady-in-Waiting to Elizabeth II of Britain, Dies at 97
The queen’s third cousin, she was a bridesmaid at the royal wedding in 1947, and witnessed firsthand pivotal moments in British history.
中文摘要 英国女王伊丽莎白二世的侍女帕梅拉·希克斯女士于97岁辞世。她是女王的表亲，曾于1947年担任王室婚礼伴娘，亲历多个英国历史重要时刻。
Astronauts return to ISS after sheltering during air leak repair attempt
Russian attempt to repair tunnel area sparks safe-haven procedure for five other astronauts onboard.
中文摘要 国际空间站发生气体泄漏，俄罗斯宇航员尝试修复隧道区域期间，站内其余五名宇航员被迫执行安全避难程序。修复作业完成后，宇航员返回空间站正常区域。
Tennis Giants Tumble
All the big names are out at the French Open. The result is a very confusing, and very exciting, tournament.
中文摘要 法国网球公开赛出现重大冷门，多位热门种子选手相继出局，赛事格局变得出人意料。无头号大热选手的赛程令本届法网充满悬念，被认为是近年来竞争最为激烈的一届。
Trump makes pitch to farmers hard-hit by tariffs, high prices in Wisconsin
Trump seeks to shore up support among rural voters hard-hit by tariffs, economic fallout of war with Iran.
中文摘要 美国总统特朗普前往威斯康星州，向受关税政策及伊朗战争经济后果重创的农民群体发表讲话，寻求巩固农村选民支持。此次访问被视为对农业州的政治拉票举措。
Putin promotes a new world economic order at St. Petersburg forum
At "Russian Davos," Putin ruled out meeting with Zelenskyy and promoted a new world economic order.
中文摘要 俄罗斯总统普京在圣彼得堡国际经济论坛（被称为"俄罗斯达沃斯"）上发表讲话，倡导建立新的世界经济秩序，同时明确拒绝与乌克兰总统泽连斯基举行面对面会谈。
Zelenskyy: Putin choosing war by rejecting offer for in-person talks
Russian President Vladimir Putin has turned down an offer for in-person talks with the Ukrainian President.
中文摘要 乌克兰总统泽连斯基表示，俄罗斯总统普京拒绝其提出的面对面谈判邀请，此举意味着普京"选择了战争"。普京此前在圣彼得堡论坛上公开排除与泽连斯基直接会晤的可能。
Trump hails jobs surge, says Iran talks ‘going well’
Trump hailed jobs growth before pivoting to Iran, saying negotiations with Tehran "seem to be going quite well".
中文摘要 美国总统特朗普高调宣扬就业市场增长数据，并就伊朗核谈判表示，与德黑兰的谈判"进展相当顺利"，释放出对谈判前景的乐观信号。
Cobolli into final as virus-struck Arnaldi pulls out of French Open
Tenth seed Flavio Cobolli will take on German second seed Alexander Zverev in final after Matteo Arnaldi withdrawal.
中文摘要 法国网球公开赛男单决赛对阵确定：第十号种子意大利选手弗拉维奥·科博利将对阵德国第二号种子亚历山大·兹维列夫，此前对手马泰奥·阿纳尔迪因病毒感染退赛。
‘Pervasive fear’ grips Gaza as Israeli attacks persist despite ceasefire
A drone attack killed a young woman and injured 15 others near Khan Younis, reported the Wafa news agency.
中文摘要 尽管停火协议存在，以色列对加沙的军事行动仍未停止。据巴勒斯坦通讯社瓦法报道，一架无人机在汗尤尼斯附近击中目标，造成一名年轻女性死亡、15人受伤，当地民众笼罩在普遍恐惧之中。
Everest guide survived six-day ordeal by eating chocolate and 'chewing ice'
Dawa Sherpa was spotted alive by a cleaning crew as he slid slowly down the world's tallest mountain and spoke to the BBC from hospital.
中文摘要 珠穆朗玛峰夏尔巴向导达瓦·夏尔巴在山上独自求生六天后奇迹获救，靠吃巧克力和"咀嚼冰块"维持生命。他被一支清洁队发现时正缓慢沿山体下滑，随后从医院接受BBC采访讲述亲身经历。
Nose Gear on Boeing 787-9 Dreamliner Collapses, Injuring Several Workers
The airline Lufthansa said the cause of the accident at Frankfurt Airport was under investigation. The plane can weigh 279 tons at takeoff.
中文摘要 一架波音787-9梦想客机在法兰克福机场发生前起落架塌陷事故，造成数名地面工作人员受伤。运营商汉莎航空表示，事故原因正在调查中。该机型最大起飞重量可达279吨。
France opens ‘war crimes’ probe into Israel’s treatment of Gaza activists
French activists who took part in a Gaza-bound foreign aid flotilla accuse Israeli forces of abuse and torture.
中文摘要 法国司法当局对以色列军队对待加沙援助船队中法国活动人士的行为展开战争罪调查。参与前往加沙援助船队的法国活动人士指控以色列军队对其实施虐待与酷刑。
Raizen Inks $13 Billion Out-of-Court Debt Deal With Creditors
Raizen SA reached an out-of-court restructuring agreement with the majority of its creditors, marking a key step in the Brazilian sugar-and-ethanol producer’s efforts to rework its debt, according to people familiar with the matter.
中文摘要 巴西糖业和乙醇生产商Raizen SA与大多数债权人达成庭外债务重组协议，涉及金额约130亿美元。此次协议是该公司推进债务重组的关键一步，具体条款由知情人士透露，标志着其化解债务危机取得实质性进展。
KPMG Chief Economist Discusses Fed Rate Hike Expectations
Diane Swonk, KPMG Chief Economist, analyzed the current economic landscape, highlighting how improvements in the labor market and persistent service sector inflation are driving hawkish sentiment among Federal Reserve officials. She noted that bond market pricing reflects expectations of a 25 basis
中文摘要 毕马威首席经济学家Diane Swonk分析当前经济形势，指出劳动力市场改善及服务业通胀持续推动美联储官员鹰派倾向。她表示债券市场定价已反映市场对美联储加息的预期，当前经济数据为进一步收紧货币政策提供支撑。
Electrum-Backed Mexican Silver Miner Sinda Files for US IPO
Sinda Ltd. filed for a US initial public offering, seeking to fund mining for silver in a historically productive part of Mexico.
中文摘要 由Electrum支持的墨西哥白银矿企Sinda Ltd.已向美国提交IPO申请，计划在墨西哥历史上产银丰富的地区开展采矿业务，募集资金将用于扩大矿业生产。
Strong U.S. Jobs Report Lowers Recession Risks Despite Wage Growth Lagging Inflation
Gene Sperling, President of Sperling Economic Strategies and former director of the National Economic Council, discussed the recent U.S. jobs report, highlighting that the headline numbers were stronger than expected and suggest a low likelihood of recession despite ongoing geopolitical and economic
中文摘要 前美国国家经济委员会主任、Sperling Economic Strategies总裁Gene Sperling表示，最新美国就业报告整体数据强于预期，显示经济陷入衰退的可能性较低。尽管当前地缘政治不确定性持续，但强劲就业数据有效降低了衰退风险。
Top Goldman lawyer Ruemmler to remain with bank despite Epstein ties
General counsel resigned after revelations about relationship with disgraced financier but now will stay on as adviser
中文摘要 高盛总法律顾问Ruemmler此前因与涉丑闻金融家爱泼斯坦的关联曝光而提出辞职，但最终将以顾问身份留任高盛。此前其辞职消息引发外界广泛关注，此次留任决定意味着高盛对其处理方式有所调整。
Tech Selloff, Bitcoin Drop Test Retail Investor Strength Ahead of SpaceX IPO
For years, Wall Street has benefited from one of the most reliable forces in modern markets: a retail-trader army willing to buy almost anything.
中文摘要 科技股下跌与比特币价格下挫对长期热衷买入的散户投资者形成考验，适逢SpaceX IPO前夕市场情绪趋于谨慎。华尔街多年来依赖散户作为重要买盘支撑力量，此次市场波动正在检验这一群体的承压能力。
Marvell Technology, Flex to Join S&P 500 Later This Month
Marvell Technology Inc. and Flex Ltd. will join the S&P 500 in the latest quarterly rebalance, S&P Dow Jones Indices said Friday.
中文摘要 标普道琼斯指数公司宣布，Marvell Technology和Flex Ltd.将在本月最新一次季度再平衡中加入标普500指数，两家公司均将于本月晚些时候正式纳入该指数成分股。
US stocks slump as fears over Big Tech shake Wall Street
The Nasdaq saw its biggest daily fall since early 2025.
中文摘要 受大型科技股抛售潮影响，美国股市大幅下跌，纳斯达克指数录得2025年初以来最大单日跌幅，华尔街市场情绪明显趋于悲观，市场对科技板块前景的担忧情绪加剧。
Goldman’s Flood Sees Buying Opportunity in Stock Market Selloff
Friday’s pullback in US equities offers a chance to add exposure rather than a reason to retreat, with a clear path for the S&P 500 to reach 8,000 this year, according to John Flood, the head of Americas equities execution services at Goldman Sachs Group Inc.
中文摘要 高盛美洲股票执行服务主管John Flood表示，周五美股回调为投资者提供了加仓机会而非撤退信号，他维持标普500指数年内有望升至8000点的判断，认为本轮下跌属于正常波动而非趋势逆转。
SpaceX signs $30bn deal to lease computing capacity to Google
Agreement comes ahead of a record-breaking initial public offering for Elon Musk’s rockets-to-AI conglomerate
中文摘要 SpaceX与谷歌签署价值300亿美元的计算能力租赁协议，此次合作发生在马斯克旗下这家火箭及人工智能综合企业筹备规模创纪录IPO之前，进一步拓展了SpaceX在科技领域的商业布局。
Beating the S&P For Generations: Masters in Business with Chris Davis
Barry sits down with Chris Davis, Chairman and Portfolio Manager at Davis Funds. They discuss his approach to managing risk and the key elements changing the economy. Chris and Barry also discuss Chris's mentors including Charlie Munger, and how he settled into the family business. (Source: Bloomber
中文摘要 Davis Funds董事长兼投资组合经理Chris Davis接受采访，分享其风险管理方法及对经济关键变量的看法。他回顾了查理·芒格等导师对其投资理念的影响，并讲述了如何在家族基金管理事业中确立自身风格。
Nasdaq tumbles 4% as shares in chip and memory groups sink
Rising expectations of a Federal Reserve rate increase send US bond yields rising sharply
中文摘要 受市场对美联储加息预期升温影响，美国国债收益率大幅上行，纳斯达克指数下跌约4%，芯片及存储器类股跌幅尤为突出，科技股整体遭遇明显抛压，市场风险情绪显著回落。
OpenAI 的账户解封了！又要重置额度了！
3 回复 · 程序员 节点
AI 的发展，是不是让大量社区的创意想法讨论也渐渐的下降了
6 回复 · 程序员 节点
分享我最近一年做产品的经历
5 回复 · 程序员 节点
各位 AI 大佬， 目前有流式输入的 llm 吗？ 有支持这个功能吗？
3 回复 · 程序员 节点
屏蔽更新
4 回复 · Apple 节点
apple 的登录可真垃圾
6 回复 · Apple 节点
请教下 mac 上指定窗口高度和大小的软件
11 回复 · Apple 节点
vibe coding 了一个高仿的 mac 端的 paste ，完全免费。
12 回复 · Apple 节点
还有几天就是苹果 WWDC26 全球开发者大会了，怎么没看到相关新闻，还有下一代 iOS27 更新了什么内容，好像也没见媒体报道了？还是说因为我关注的少了不给我推了？
29 回复 · Apple 节点
玄学？为啥土区支付方式不能改为 None？
17 回复 · Apple 节点
【开源求Star】深度调研报告生成 Skill，一个命令，六分钟，一份深度专业的调研报告
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 自己的第一个开源项目，全程vibe-coding。 在Opencode下已测试迭代N次，完整可用，搭配deepseek v4 flash，几乎0成本。 注意，适配调研任何主题，不仅仅能做行业研报哦！看我出的案例报告就知
你说今天能破100,000,000,000吗？
千亿，指日可待 93 个帖子 - 85 位参与者 阅读完整话题
RawChat 公益站怎么关站了
刚才看到有评论说群里发消息了，站长被气到了然后关站了，有在群里的佬说说发生什么了吗？ 62 个帖子 - 51 位参与者 阅读完整话题
OpenAI已经开始调查被误封的账户了，许多外国佬也被误封！
希望明天能看到账号解封的消息，那群逆天号商整天去官网搞举报，看看伤害多少无辜的正规充值的用户啊 39 个帖子 - 24 位参与者 阅读完整话题
分享一下我的 Windows 桌面图标设置
应该有很多人跟我一样不喜欢桌面一堆乱七八糟的图标，也不想装 uTools、Listary 这种额外的启动器。所以俺来分享一下这个用了好多年的 windows 上的软件使用习惯 核心思路：把所有软件的快捷方式统一放到一个文件夹（比如 QuickWay），再把这个文件夹加进环境变量，之后就可以在任意界面按下 Win+R 输入快捷方式名称回车，就可以直接打开对应软件 1. 新建一个放快捷方式的文件夹 在任意位置新建一个文件夹，命名为 QuickWay（名字随意，以下统一用这个），建议放在不容易误删的位置，我是放在了个人文件夹下，即 C:\Users\TaiYang\QuickWay 2. 把所有软件
建议少用国产辣鸡小模型，太让人红温了
最近Minimax刚上线了M3版本模型，因为公司采购的就是Minimax家订阅，所以公司项目代码就只能用它来写，使用过程中感觉豆包味越来越重，甚至无视写在项目md里的红线警告，自己往数据库写了数据，事后我让它细化md文档，更是搞了一堆废话进去。 公司也是抠，好歹也选个Qwen或者DS…… 40 个帖子 - 27 位参与者 阅读完整话题
还真有人拿我们的公益站去卖，而且还卖的的很贵！
事情是这样的，有人跟我们反馈说回答里出现了我们的来源信息，因为我们是动态风控机制的，正常使用是不会出现来源信息的，只有多设备 多ip 多账号 环境跳来跳去，才会触发这个风控机制，会动态增加来源信息 然后私聊问了下，如下： 没想到居然拿我们的公益站去卖而且还卖的这么贵！0.2的倍率，太夸张了！ 由于当事人说收到了很多人的消息，不想被打扰，所以码掉了 224 个帖子 - 203 位参与者 阅读完整话题
41个20xPro 到现在封的剩下9个 说下我的情况 你们参考下
41个pro开号方式如下： 11个为美区没免税 套google play开的，220刀/月 之前封了7个 今早到现在为止封了25个，余9个 今天死的号为1个美区免税 2个土区 20个礼品卡 2个美区没免税 现在看来美区没免税能活久点 再说下2个多月做中转站，到现在的情况 总投入还差23000+ 没有回本，别说赚钱了 另外封了的号 googleplay去申请退款，没有一个给我退的，不给退 佬们参考下吧，9个号刷完 或者 刷死 结束 终于能睡个好觉了 65 个帖子 - 43 位参与者 阅读完整话题
通篇阅读《置身钉内》有感，果然阿里系领导的风格都是一样的
中午偶然间刷脉脉看到 ATA 又有小作文爆火了，于是去小红书上找了找 7w 字原文 drive.google.com 置身钉内 14.34.50.pdf Google Drive file. 在 7w 字的原文中见到了很多相似的、过去反复发生的故事，唯一没想到的是，都过去至少 5-6 年了，逻辑还是和以往一模一样，甚至都不带变的 从内容上来说，是毫无疑问的好文章，共享一下 PS：为什么现在很多人看到这种精品文章，下意识都会去找 AI 蒸馏的简略版呢？并不是很理解 看了一下后续佬友们的回复，仔细思考了一下，用 AI 作为前置筛选器，筛选一下双方观点互通的文章也是挺有意义的，这样对创作者和阅读者都
海南的天真好看👋
（划水.gif） 像是动漫里走出来的世界 41 个帖子 - 31 位参与者 阅读完整话题