每日简报

2026-06-19

← 历史归档

google-research/timesfm

Python · ★ 23,084 · 🍴 2,213 · 📈 858 stars today

TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting.

中文介绍 TimesFM 是 Google Research 开发的预训练时间序列基础模型,基于 Transformer 架构在大规模数据上预训练,专门用于时序预测。适用于金融、气象、能源等领域的需求预测,可快速适应多种预测场景。

n0-computer/iroh

Rust · ★ 9,994 · 🍴 461 · 📈 369 stars today

IP addresses break, dial keys instead. Modular networking stack in Rust.

中文介绍 Iroh 是一个用 Rust 构建的模块化网络栈,基于密钥而非 IP 地址进行通信,提供更可靠的 P2P 连接。适合去中心化应用、分布式系统等需要抗 IP 故障的场景。

freeCodeCamp/freeCodeCamp

TypeScript · ★ 449,523 · 🍴 45,125 · 📈 417 stars today

freeCodeCamp.org's open-source codebase and curriculum. Learn math, programming, and computer science for free.

中文介绍 freeCodeCamp 的开源代码库和课程平台,提供免费、互动的编程、数学和计算机科学教育。通过项目实践帮助学习者掌握全栈开发技能,适合零基础或转行开发者。

obra/superpowers

Shell · ★ 232,363 · 🍴 20,639 · 📈 1,435 stars today

An agentic skills framework & software development methodology that works.

中文介绍 Superpowers 是一个实用的智能体技能框架和软件开发方法论,提供结构化流程和工具,帮助团队高效构建、部署和迭代 AI 代理。适合需要构建自主 AI 系统的开发者。

zai-org/GLM-5

★ 4,098 · 🍴 429 · 📈 286 stars today

GLM-5: From Vibe Coding to Agentic Engineering

中文介绍 GLM-5 是新一代语言模型,从“氛围编码”迈向智能体工程,具备更强的推理和代理能力。适用于构建自主 AI 代理、复杂任务规划和执行。

DeusData/codebase-memory-mcp

C · ★ 6,969 · 🍴 563 · 📈 2,308 stars today

High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 158 languages, sub-ms queries, 99% fewer tokens. Single static binary, zero dependencies.

中文介绍 高性能代码智能 MCP 服务器,将代码库索引为持久化知识图谱,支持 158 种语言,毫秒级索引、亚毫秒查询,减少 99% token 消耗。单一静态二进制,适合大规模代码分析和 AI 辅助编程。

yifanfeng97/Hyper-Extract

Python · ★ 1,737 · 🍴 199 · 📈 124 stars today

Transform unstructured text into structured knowledge with LLMs. Graphs, hypergraphs, and spatio-temporal extractions — with one command.

中文介绍 Hyper-Extract 利用 LLM 将非结构化文本转化为结构化知识,支持图、超图和时空关系抽取。一条命令即可提取实体关系,适用于知识图谱构建、信息抽取等场景。

alibaba/zvec

C++ · ★ 11,197 · 🍴 648 · 📈 344 stars today

A lightweight, lightning-fast, in-process vector database

中文介绍 zvec 是阿里巴巴开源的轻量级进程内向量数据库,无需独立部署,可嵌入应用,提供毫秒级向量检索。适合推荐系统、图像检索、语义搜索等场景。

withastro/flue

TypeScript · ★ 5,479 · 🍴 298 · 📈 164 stars today

The sandbox agent framework.

中文介绍 Flue 是一个沙盒智能体框架,提供安全隔离环境运行 AI 代理,确保代码可控。适用于代理开发、安全测试和不可信代码执行场景。

Kilo-Org/kilocode

TypeScript · ★ 22,069 · 🍴 2,704 · 📈 1,339 stars today

Kilo is the all-in-one agentic engineering platform. Build, ship, and iterate faster with the most popular open source coding agent.

中文介绍 Kilo 是一体化智能体工程平台,集成开源编码代理,帮助开发者更快构建、交付和迭代代码。提供端到端工具链,适合 AI 辅助编码和自动化开发流程。

makeplane/plane

TypeScript · ★ 51,809 · 🍴 4,599 · 📈 610 stars today

🔥🔥🔥 Open-source Jira, Linear, Monday, and ClickUp alternative. Plane is a modern project management platform to manage tasks, sprints, docs, and triage.

中文介绍 Plane 是开源项目管理平台,可替代 Jira、Linear 等工具。支持任务管理、冲刺、文档和工单分类,界面现代化,适合敏捷开发团队。

Kong/insomnia

TypeScript · ★ 38,657 · 🍴 2,290 · 📈 13 stars today

The open-source, cross-platform API client for GraphQL, REST, WebSockets, SSE and gRPC. With Cloud, Local and Git storage.

中文介绍 Insomnia 是开源跨平台 API 客户端,支持 GraphQL、REST、WebSocket、SSE 和 gRPC,提供云端、本地和 Git 存储。适用于 API 开发、测试与调试。

Universal-Debloater-Alliance/universal-android-debloater-next-generation

Rust · ★ 7,907 · 🍴 338 · 📈 247 stars today

Cross-platform GUI written in Rust using ADB to debloat non-rooted Android devices. Improve your privacy, the security and battery life of your device.

中文介绍 使用 Rust 编写的跨平台 GUI 工具,通过 ADB 清理未 root 安卓设备上的预装应用,提升隐私、安全和电池续航。适合希望精简系统的 Android 用户。

dotnet/aspnetcore

C# · ★ 38,091 · 🍴 10,708 · 📈 22 stars today

ASP.NET Core is a cross-platform .NET framework for building modern cloud-based web applications on Windows, Mac, or Linux.

中文介绍 ASP.NET Core 是微软推出的跨平台 .NET 框架,用于构建现代云 Web 应用。高性能、开源,支持容器化和微服务,适合企业级后端开发。

owainlewis/awesome-artificial-intelligence

★ 14,363 · 🍴 2,333 · 📈 33 stars today

A curated list of Artificial Intelligence (AI) courses, books, video lectures and papers.

中文介绍 Awesome Artificial Intelligence 是精选的 AI 学习资源清单,涵盖课程、书籍、视频讲座和论文。适合初学者和进阶者系统学习人工智能。

Lightricks/LTX-2

Python · ★ 7,479 · 🍴 1,206 · 📈 47 stars today

Official Python inference and LoRA trainer package for the LTX-2 audio–video generative model.

中文介绍 LTX-2 是 Lightricks 的音视频生成模型,官方提供 Python 推理和 LoRA 训练工具包。支持音频与视频联合生成,适用于内容创作和视频编辑。

LibreTranslate/LibreTranslate

Python · ★ 14,956 · 🍴 1,530 · 📈 83 stars today

Free and Open Source Machine Translation API. Self-hosted, offline capable and easy to setup.

中文介绍 LibreTranslate 是免费开源的机器翻译 API,支持自托管和离线运行,设置简单。无需依赖第三方服务,适合需要隐私保护和离线翻译的应用。

Learning from the Self-future: On-policy Self-distillation for dLLMs

👍 69

On-policy self-distillation (OPSD) has proven effective for post-training large language models (LLMs), yet its application to diffusion LLMs (dLLMs) remains unexplored. Existing OPSD methods are inherently autoregressive-centric. They inject privileged information via left-to-right prefix condition

中文介绍 提出针对扩散语言模型(dLLMs)的在线自蒸馏方法OPSD-dLLM,解决现有自回归中心方法无法直接应用于dLLMs的问题。核心思路是利用未来自身生成作为特权信息进行蒸馏,在多个文本生成任务上优于已有方法。

Beyond the Current Observation: Evaluating Multimodal Large Language Models in Controllable Non-Markov Games

👍 37

Deploying multimodal foundation models as closed-loop policies increasingly requires conditioning actions on observations that are no longer visible. However, existing benchmarks either expose the full state, conflate hidden-state reconstruction with other agent skills, or test recall only after an

中文介绍 提出可控非马尔可夫博弈基准,用于评估多模态大模型在隐藏状态下的决策能力。现有基准暴露全部状态或混淆重建技能,该工作通过设计部分可观测环境,测试模型基于历史观测推理未来的能力。

MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction

👍 30

Motion forecasting is central to visual intelligence: agents must anticipate how objects will move in order to plan actions, reason about physical interactions, and synthesize realistic futures. We argue that 3D points in world coordinates provide a general representation that is class-agnostic, vie

中文介绍 提出MolmoMotion,通过语言指令预测3D世界坐标中的点轨迹。使用类无关的3D点表示,结合视觉与语言指令进行运动预测,在多个数据集上实现零样本泛化和强时序一致性。

Guava: An Effective and Universal Harness for Embodied Manipulation

👍 23

Language models trained on large-scale vision-language data have demonstrated strong potential for embodied agents. Harnessing models through embodied tools use offers a promising alternative to end-to-end vision-language-action systems by combining high-level reasoning with external modules for per

中文介绍 提出Guava,一种通用具身操作工具框架,利用预训练视觉语言模型的高层推理能力结合外部操作模块,无需端到端训练。在多种机器人任务上验证其有效性和泛化性。

EfficientRollout: System-Aware Self-Speculative Decoding for RL Rollouts

👍 16

Reinforcement learning (RL) has become a representative post-training paradigm for LLMs, enabling strong reasoning and agentic capabilities. However, rollout generation remains a dominant latency bottleneck because autoregressive sampling decodes responses sequentially and a small number of long-tai

中文介绍 提出EfficientRollout,系统感知的自推测解码方法加速RL rollout生成。针对自回归采样瓶颈,利用小模型快速生成草稿并由LLM验证,在保持性能的同时显著降低推理延迟。

Text-Vision Co-Instructed Image Editing

👍 14

Existing image editing methods can be generally categorized into textual instruction-based and visual prompt-based ones. Textual instructions are semantically expressive, but are limited by the coarse granularity of spatial control of the editing results. In contrast, visual prompts such as drag and

中文介绍 提出文本-视觉联合指令图像编辑方法,融合语义丰富的文本指令与空间精确的视觉提示(如拖动、涂鸦)。通过双分支条件生成,在细粒度空间控制和语义保真度上均优于单一模态方法。

Reinforcing Dual-Path Reasoning in Spatial Vision Language Models

👍 13

Spatial VLMs have made substantial progress in geometric perception, yet complex spatial reasoning requiring multi-step inference over depth, distance, and scene relations remains challenging. Moreover, different spatial queries call for fundamentally different strategies: some are best addressed th

中文介绍 提出双路径推理强化方法Reinforce-Spatial-VLM,增强空间VLM的深度、距离和场景关系多步推理能力。针对不同空间查询采用不同策略(几何路径或关系路径),在多个基准上提升准确率。

From Trainee to Trainer: LLM-Designed Training Environment for RL with Multi-Agent Reasoning

👍 12

Reinforcement learning pipelines for Large Language Model (LLM) training often rely on manually redesigned environments between stages, requiring practitioners to heuristically infer which configuration will best improve the current policy. To automate this process, we propose the LLM-as-Environment

中文介绍 提出LLM-as-Environment框架,让LLM自动设计RL训练环境,替代人工手动调优。通过多智能体推理生成合适难度的任务,迭代提升策略性能,在数学推理等任务上验证有效性。

Self-Evolving Visual Questioner

👍 12

Vision-language models (VLMs) are typically trained as passive answerers, while their ability to actively ask diverse, non-trivial, visual-centric and grounded questions remains underexplored. Existing visual questioners' performance is bottlenecked by the availability of high-quality training data

中文介绍 提出自进化视觉提问器(Self-Evolving Visual Questioner),让VLM主动生成多样化、视觉中心且接地气的问题。通过迭代式自我提升,利用生成的高质量问答对训练,在问题质量指标上超越现有方法。

EgoCS-400K: An Egocentric Gameplay Dataset for World Models

👍 12

The shift from video generation to interactive world modeling places new demands on data: beyond captioned videos, world models require temporally aligned video-action-language trajectories grounded in the actions, camera motion, states, and events that drive future scene changes. However, such data

中文介绍 构建EgoCS-400K,一个大规模自我中心游戏数据集,包含时间对齐的视频-动作-语言轨迹。用于训练世界模型,提供动作、相机运动、状态和事件细粒度标注,支持交互式未来预测。

Native Active Perception as Reasoning for Omni-Modal Understanding

👍 11

Passive models for long video understanding typically rely on a "watch-it-all" paradigm, processing frames uniformly regardless of query difficulty, causing computational cost to grow with video duration. Although interactive frameworks have emerged, they often rely on global pre-scanning, and their

中文介绍 提出原生主动感知框架NAR,通过推理驱动选择性帧处理实现长视频理解。无需全局预扫描,计算代价与视频时长解耦,在全模态理解任务上以更少计算量达到或超越被动模型性能。

MaineCoon: Pursuing A Real-Time Audio-Visual Social World Model

👍 8

As an increasing majority of global video content is consumed on social platforms for interactive social purposes, video generation models built for social worlds are important but largely overlooked by previous studies. In this work, we define the position of social world models and build a prototy

中文介绍 定义社交世界模型MaineCoon,面向社交平台的实时音视频生成。结合对话上下文和面部表情等多模态信号,生成社交互动中合理的视频内容,实现实时性要求下的高质量生成。

STARE: Surprisal-Guided Token-Level Advantage Reweighting for Policy Entropy Stability

👍 8

Reinforcement Learning with Verifiable Rewards algorithms like GRPO have emerged as the dominant post-training paradigm for complex reasoning in LLMs, yet commonly suffer from policy entropy collapse during training. We conduct a first-order gradient analysis of token-level entropy dynamics under GR

中文介绍 提出STARE,基于惊讶度引导的token级优势再加权方法,解决GRPO训练中策略熵崩溃问题。通过梯度分析发现熵降源于优势分布偏斜,重新加权后保持熵稳定,提升推理性能。

Sumi: Open Uniform Diffusion Language Model from Scratch

👍 8

Diffusion models have become a promising alternative to autoregressive models. Among these, uniform diffusion language models (UDLMs) permit any token to be updated at any step, in principle enabling more flexible generation. However, no UDLM has yet been pretrained from scratch at both large parame

中文介绍 发布Sumi,首个从零预训练的大规模统一扩散语言模型(UDLM),支持任意token任意步更新。在参数规模和训练数据上达到现有UDLM之最,在多项语言生成基准上表现优异。

Dr-DCI: Scaling Direct Corpus Interaction via Dynamic Workspace Expansion

👍 8

Agentic search over large corpora relies on retriever-mediated interfaces (e.g., BM25 or ColBERT) for scalable candidate discovery. While effective at ranking relevant documents, these interfaces expose evidence only as ranked results or bounded document views, limiting agents' ability to reorganize

中文介绍 提出Dr-DCI,通过动态工作空间扩展实现直接语料交互。允许智能体在检索过程中动态重组和组织文档片段,超越传统排序结果浏览,提升复杂信息搜索任务的效率和准确性。

Beyond Alignment: Value Diversity as a Collective Property in Multicultural Agent Systems

👍 7

Multicultural multi-agent systems are increasingly deployed in globally diverse settings, where different agents are grounded in different cultural backgrounds. Existing cultural evaluation focuses on value alignment: how closely a single agent matches a target culture. Yet alignment is a per-agent

中文介绍 提出超越对齐的价值多样性概念,研究多元文化多智能体系统中不同文化背景智能体的集体价值分布。通过模拟实验发现整体文化多样性影响群体协作效果,为全局系统设计提供新视角。

ViT-Up: Faithful Feature Upsampling for Vision Transformers

👍 6

Vision Transformers (ViTs) have become a dominant architecture for visual representation learning, providing exceptionally strong and broadly reusable backbone features. However, ViTs are commonly operated on relatively small patch-token grids due to the quadratic cost of global self-attention, whic

中文介绍 提出ViT-Up,对视觉Transformer进行可信的特征上采样。解决ViT因全局自注意力二次开销而使用小patch网格造成的细节损失,通过逐层上采样保持特征保真度,提升密集预测任务性能。

CEO-Bench: Can Agents Play the Long Game?

👍 6

Language model agents are becoming proficient executors at isolated, short-horizon tasks such as software engineering and customer service. Yet real-world challenges require a combination of sophisticated skills that remain largely untested in agents: (1) navigating long horizons amid uncertainty; (

中文介绍 提出CEO-Bench基准,评估语言模型智能体在长周期、不确定性、多目标等复杂长期任务上的能力。包含模拟真实场景的任务套件,测试智能体规划、适应和优先级管理能力。

ChLogic: Evaluating Robustness of Logical Reasoning in Chinese Expressions

👍 6

Large language models perform increasingly well on standardized logical reasoning benchmarks, but whether this ability remains robust beyond English is unclear. We introduce ChLogic, an English--Chinese aligned benchmark that tests whether models preserve logical reasoning performance when the same

中文介绍 构建ChLogic,英中双语对齐的逻辑推理鲁棒性基准。测试大模型在相同逻辑任务上跨语言一致性,发现模型在中文表达下性能显著下降,揭示当前推理能力的语言依赖性。

PAIWorld: A 3D-Consistent World Foundation Model for Robotic Manipulation

👍 5

World foundation models (WFMs) are powerful simulators, yet they predominantly operate in a single-view setting and lack the multi-view 3D consistency required for robotic manipulation. While robotic systems rely on multiple cameras (egocentric, eye-to-hand, and wrist-mounted) for policy learning, c

中文介绍 提出PAIWorld,3D一致的世界基础模型用于机器人操作。支持多视角(自我中心、眼到手、腕装)输入,生成3D一致的未来帧预测,提升下游策略学习的泛化性和鲁棒性。

How to Build a Claude Code Agent Team That Runs in Loops (Exact Setup Inside)

@zodchiii · 22.7K 粉丝 · 1.2M 阅 · 500 赞 · 71 转

Most setups run agents once and hand you whatever comes out. A team that runs in loops keeps going until the work actually passes. Below is the setup in 3 files: the agents, the loop that drives

中文介绍 搭建 Claude Code 循环运行的 agent 团队,包含 3 个文件(agents、loop 驱动等),让 agent 持续迭代直到任务通过,而不是一次输出。给出精确配置。

Using Claude to go Viral on X… (Mr. Beasts Framework)

@mattepstein · 35.6K 粉丝 · 393.3K 阅 · 504 赞 · 26 转

Have you seen any of the launches below on your timeline? (you probably have).. What if I told you they all followed a repeatable viral science that can be 95% automatable with claude. In this

中文介绍 揭秘一种可重复的病毒式传播科学,称可用 Claude 实现 95% 自动化制造爆款推文。引用 Mr. Beast 框架,展示如何借助 AI 复制高互动内容。

The Self-Improving Loop: a 300-agent swarm on Kimi K2.6, verified by Opus 4.8

@0xMovez · 26.7K 粉丝 · 208.0K 阅 · 504 赞 · 59 转

A free open-source model is running 300 parallel agents across 4,000 coordinated steps from a single prompt, and it scores higher on real research tasks than models you pay 5x more for. Most people

中文介绍 通过单条 prompt 驱动 300 个并行 agent、4000 步协调执行的免费开源模型 Kimi K2.6,在真实研究任务上得分超过付费模型 5 倍,经 Opus 4.8 验证。展示大规模 agent swarm 的自我改进循环。

Three Ways Codex Can Use a Computer

@jxnlco · 105.9K 粉丝 · 204.4K 阅 · 504 赞 · 47 转

Update: Computer Use is now Available in the EU/UK ;) Enjoy! There are three ways for Codex to use a computer: Computer Use, the Chrome extension, and the in-app browser. They overlap just enough to

中文介绍 介绍 Codex 使用计算机的三种方式:Computer Use、Chrome 扩展和内置浏览器,三者功能有所重叠但各有侧重,并宣布 EU/UK 地区现已可用。

Zen and the Art of AI Research

@jxmnop · 50.7K 粉丝 · 114.1K 阅 · 504 赞 · 57 转

So you want to do AI research? It's true that no one really teaches you how. Not directly, anyway. But it turns out that the way to get started is pretty simple: some combination of (i) reading and

中文介绍 探讨如何入门 AI 研究,指出核心是结合阅读和动手实践。没有直接教学方法,而是提供系统性的起步思路。

ORACLE: Official AI Agents Trade on Polymarket

@OracleMindAI · 21.0K 粉丝 · 105.0K 阅 · 2.8K 赞 · 582 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 介绍 ORACLE 官方 AI agent 在 Polymarket 上交易,称到 2026 年预测市场超 30% 活跃来自算法和 AI 钱包,展示自主 agent 在预测市场中的有效性。

The Art of Loop Engineering

@sydneyrunkle · 7.9K 粉丝 · 74.7K 阅 · 565 赞 · 87 转

Agents are useful because they help us automate work by taking actions in the real world. But getting agents to do valuable work reliably takes more than just a good model: it requires a carefully

中文介绍 探讨 agent 循环工程的精髓:让 agent 可靠完成有价值工作不仅需要好模型,更需要精心设计的循环架构。分享构建稳定高效 agent 循环的方法。

Agent harness engineering with Claude: 14-step roadmap from one agent to a self-improving system.

@0xCodez · 8.2K 粉丝 · 72.8K 阅 · 508 赞 · 79 转

Everyone’s talking about loops. Almost no one is talking about what the loop runs on. 9 out of 10 builders run Claude Code on the default harness - no rules, no subagents, no hooks, no memory. Then

中文介绍 指出 9/10 的开发者只用 Claude Code 默认 harness(无规则、子 agent、hooks、记忆),然后抱怨效果差。提供 14 步路线图从单 agent 升级到自改进系统,强调 harness 工程。

The Stanford STORM Method: How to Make Claude Research Like a PhD in Minutes

@heynavtoor · 143.5K 粉丝 · 47.7K 阅 · 538 赞 · 70 转

Most people use Claude like a search box. Ask, answer, close tab. They are leaving the best feature locked. Save this :) Stanford built a research system called STORM. In peer reviewed testing it

中文介绍 介绍斯坦福 STORM 研究系统,让 Claude 在几分钟内像博士一样做研究。在同行评审测试中表现优异,教用户利用此方法进行深度研究而非简单搜索。

How to build a self-improvement loop for your Skills

@zachlloydtweets · 10.6K 粉丝 · 43.5K 阅 · 512 赞 · 53 转

There’s been a lot of chatter about using “loops” lately to drive agents, and I think this has been accompanied by a bit of “what actually is a loop”? I can’t speak for everyone else using the term,

中文介绍 澄清 agent 驱动中「循环」概念,指导如何为自己的技能构建自我改进循环。提供可操作的循环设计方法,而非空谈术语。

The 10 rules to ship truly polished UI with Claude

@kvnkld · 7.2K 粉丝 · 42.0K 阅 · 531 赞 · 25 转

People keep asking how the UI components I post end up looking so polished, or what prompts I use. So here’s a breakdown of the most important things: Polish is not a feature you prompt for. You can't

中文介绍 10 条规则教你用 Claude 生成精致 UI。强调精致感不能仅靠提示词得到,需要系统性工程方法,并分享了 prompt 与设计思路。

Arcads MCP + Claude Just Changed AI Ad Creation Forever

@Just_sharon7 · 44.1K 粉丝 · 32.5K 阅 · 502 赞 · 18 转

If you're still switching between 10 tools to make UGC ads, analyze performance, and iterate... your workflow is outdated. Arcads just dropped MCP support, and pairing it with Claude (especially Fable

中文介绍 Arcads 推出 MCP 支持,与 Claude(特别是 Fable)结合,将 UGC 广告制作、性能分析、迭代集成到一个工作流中,告别多工具切换,彻底改变 AI 广告创作。

10 Crazy Facts About AI That Will Blow Your Mind

@ActionModelAI · 57.8K 粉丝 · 5.7K 阅 · 501 赞 · 366 转

Most people think AI is still in its early stages. They're right. And that's exactly why these facts are so crazy. 1. AI could add $15.7 trillion to the global economy by 2030 According to PwC, AI

中文介绍 列举 10 条关于 AI 的惊人事实,如 AI 到 2030 年可为全球经济增加 15.7 万亿美元(PwC 预测),旨在展示 AI 早期阶段的巨大潜力。

The Nervous System of the Autonomous Age

@OptimaiNetwork · 99.4K 粉丝 · 5.3K 阅 · 511 赞 · 360 转

The old internet was built for humans browsing pages, clicking ads, and buying subscriptions. The autonomous internet needs something else: live context, verified signals, permissioned execution, and

中文介绍 提出自主互联网需要的新基础设施:实时上下文、验证信号、权限执行等,如同自主时代的神经系统,对比旧互联网与未来网络的不同。

ORACLE: Official AI Agents Trade on Polymarket

@Trade_OracleAI · 21.1K 粉丝 · 60 阅 · 2.8K 赞 · 584 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 ORACLE 在 Polymarket 上交易自主 AI agent 的策略,指出 2026 年预测市场超 30% 活跃来自算法和 AI 钱包,展示 agent 在预测市场中的有效性。

The Professor of Outputmaxxing — Anjney Midha, AMP

We talk about how this legendary investor went from humble beginnings in Singapore to leading rounds in Anthropic, Mistral, Black Forest Labs, and Periodic Labs... and the AMP secret master plan!

中文介绍 传奇投资人Anjney Midha从新加坡起步,主导了Anthropic、Mistral、Black Forest Labs和Periodic Labs等投资,并透露AMP秘密计划。

New usage analytics and updated spend controls for enterprises

OpenAI introduces new spend controls and usage analytics for ChatGPT Enterprise, helping organizations manage costs and scale AI with confidence.

中文介绍 OpenAI为ChatGPT Enterprise推出新的支出管控和使用分析功能,帮助企业管理成本并自信扩展AI应用。

Improving health intelligence in ChatGPT

Learn how GPT-5.5 Instant improves ChatGPT’s health and wellness responses with stronger reasoning, better context, clearer communication, and physician-informed evaluations.

中文介绍 GPT-5.5 Instant提升了ChatGPT在健康与wellness方面的推理能力、上下文理解、清晰沟通及医生评估,改善健康智能响应。

🔬 The Self-Driving Lab — Joseph Krause, Radical AI

Radical AI's Joseph Krause on why the moat in materials is the lab, not the model

中文介绍 Radical AI的Joseph Krause认为材料领域的护城河在于实验室而非模型,阐述“自动驾驶实验室”理念。

GLM-5.2: Built for Long-Horizon Tasks

中文介绍 GLM-5.2模型专为长周期任务设计,在长上下文和复杂推理方面具有优势,位列顶级开源模型。

市场总览

当前市场技术面分化明显。美股整体偏强,QQQ、SPY接近52周高,但内部出现分化:MSFT、META呈空头排列,RSI低于50。加密市场极度恐慌(恐慌贪婪指数仅15),BTC-USD、ETH-USD均处于空头排列,RSI接近超卖,但MACD出现金叉信号,下行压力未减。中概股普遍弱势,BABA RSI进入超卖区(24.7),空头排列;PDD、JD也呈空头格局。商品外汇方面,黄金期货高位回落,WTI原油近5日重挫14%,美元指数DXY走强并触及RSI超买(70.6),暗示短期可能回调。整体看,美股科技龙头与防御性资产分化,风险偏好偏低。

今日关注

QQQ Nasdaq 100 ETF
偏上行

QQQ 当前价格 740.62,高于 SMA20(726.88)、SMA50(693.18)、SMA200(628.63),呈多头排列。近 1 日涨幅 2.51%,近 5 日 3.28%,接近 52 周高(仅差 1.07%)。RSI14 为 59.1(中性偏强),MACD 柱状图正值(9.5234)但略低于信号线(11.1108),动量略有减弱,但整体上升趋势明确。

MSFT Microsoft
偏下行

MSFT 当前价格 379.4,远低于 SMA20(413.15)、SMA50(412.98)、SMA200(451.35),呈空头排列。近 1 日仅涨 0.13%,近 5 日下跌 2.8%,距 52 周高回撤 31.7%。RSI14 仅 34.9(接近超卖),MACD 为 -8.5715,信号线 -3.5444,柱状图负值且死叉运行,下行压力显著。

BABA 阿里巴巴 (BABA)
偏下行

BABA 当前价格 107.1,远低于所有均线(SMA20=120.91、SMA50=129.3、SMA200=149.19),空头排列。近 5 日下跌 4.96%,距 52 周高回撤 44.41%。RSI14 为 24.7,进入超卖区域,但 MACD -6.2971(信号线 -4.8371)柱状图负值且未收敛,短期仍偏弱。

GOOGL Alphabet
中性

GOOGL 当前价格 368.03,小幅低于 SMA20(371.63)但高于 SMA50(367.37)和 SMA200(311.10),趋势偏牛但短期均线压制。RSI14 为 49.1(中性),MACD -1.8497(信号线 -0.7339)柱状图负值但幅度缩小,方向不明。近 1 日上涨 1.17%,近 5 日涨 2.87%,多空力量均衡。

DX-Y.NYB 美元指数 DXY
偏上行

DXY 当前价格 100.81,高于 SMA20(99.59)、SMA50(98.92)、SMA200(98.68),多头排列。近 1 日上涨 0.72%,近 5 日涨 0.96%,接近 52 周高(仅差 0.1%)。RSI14 为 70.6,进入超买区,MACD 金叉且柱状图正值(0.3441>信号线0.2816),上升动量强劲,但超买提示短期回调风险。

全部资产

^VIX

VIX 恐慌指数

$16.40 -11.06%
5 日
-15.64%
距 52w 高
-53.5%
RSI(14)
46.0
趋势
空头
SMA 20 / 50 / 200
17.40 / 17.79 / 18.56
MACD / 信号
-0.108 / -0.053
死叉(SMA50↓SMA200) (6 天前)MACD 死叉 (今天)空头排列

^TNX

10Y 美债收益率 (%)

$4.45 -0.80%
5 日
-2.22%
距 52w 高
-10.9%
RSI(14)
47.0
趋势
多头
SMA 20 / 50 / 200
4.52 / 4.42 / 4.21
MACD / 信号
0.015 / 0.026
多头排列

DX-Y.NYB

美元指数 DXY

$100.81 +0.72%
5 日
+0.96%
距 52w 高
-0.1%
RSI(14)
70.6
趋势
多头
SMA 20 / 50 / 200
99.59 / 98.92 / 98.68
MACD / 信号
0.344 / 0.282
MACD 金叉 (1 天前)RSI 超买接近 52 周高多头排列

SPY

S&P 500 ETF

$746.74 +0.78%
5 日
+1.22%
距 52w 高
-1.8%
RSI(14)
54.1
趋势
多头
SMA 20 / 50 / 200
747.08 / 729.66 / 688.36
MACD / 信号
3.930 / 5.524
接近 52 周高多头排列

QQQ

Nasdaq 100 ETF

$740.62 +2.51%
5 日
+3.28%
距 52w 高
-1.1%
RSI(14)
59.1
趋势
多头
SMA 20 / 50 / 200
726.88 / 693.18 / 628.63
MACD / 信号
9.523 / 11.111
接近 52 周高多头排列

AAPL

Apple

$298.01 +0.70%
5 日
+0.81%
距 52w 高
-6.1%
RSI(14)
50.9
趋势
多头
SMA 20 / 50 / 200
303.40 / 288.74 / 268.19
MACD / 信号
1.180 / 3.216
多头排列

MSFT

Microsoft

$379.40 +0.13%
5 日
-2.80%
距 52w 高
-31.7%
RSI(14)
34.9
趋势
空头
SMA 20 / 50 / 200
413.15 / 412.98 / 451.35
MACD / 信号
-8.571 / -3.544
空头排列

NVDA

Nvidia

$210.69 +2.95%
5 日
+2.84%
距 52w 高
-10.9%
RSI(14)
50.4
趋势
多头
SMA 20 / 50 / 200
211.79 / 209.31 / 189.90
MACD / 信号
-1.067 / -0.113
多头排列

GOOGL

Alphabet

$368.03 +1.17%
5 日
+2.87%
距 52w 高
-9.9%
RSI(14)
49.1
趋势
多头
SMA 20 / 50 / 200
371.63 / 367.37 / 311.10
MACD / 信号
-1.850 / -0.734
多头排列

TSLA

Tesla

$400.49 +1.04%
5 日
+0.34%
距 52w 高
-19.7%
RSI(14)
47.0
趋势
空头
SMA 20 / 50 / 200
413.70 / 402.49 / 416.96
MACD / 信号
-2.850 / -0.505
空头排列

META

Meta

$577.22 +1.70%
5 日
+1.55%
距 52w 高
-27.5%
RSI(14)
42.8
趋势
空头
SMA 20 / 50 / 200
599.48 / 621.90 / 654.92
MACD / 信号
-11.455 / -9.861
空头排列
加密恐慌贪婪
15
极度恐慌
加密总市值
$2.25 T
-2.10% / 24h
BTC 主导率
55.9%
ETH 9.1%
24h 成交量
$81.7 B
活跃币 17,435

BTC-USD

Bitcoin

$62,805.33 -4.26%
5 日
-1.16%
距 52w 高
-50.2%
RSI(14)
34.9
趋势
空头
SMA 20 / 50 / 200
65,484.56 / 73,166.42 / 77,254.04
MACD / 信号
-2,578.145 / -3,079.931
MACD 金叉 (4 天前)空头排列

ETH-USD

Ethereum

$1,704.68 -4.79%
5 日
+2.38%
距 52w 高
-65.6%
RSI(14)
38.5
趋势
空头
SMA 20 / 50 / 200
1,764.72 / 2,035.20 / 2,389.15
MACD / 信号
-90.792 / -113.155
MACD 金叉 (3 天前)空头排列

SOL-USD

Solana

$69.40 -3.52%
5 日
+0.77%
距 52w 高
-72.6%
RSI(14)
41.8
趋势
空头
SMA 20 / 50 / 200
70.48 / 80.45 / 98.38
MACD / 信号
-2.805 / -3.859
MACD 金叉 (4 天前)空头排列

BABA

阿里巴巴 (BABA)

$107.10 -0.32%
5 日
-4.96%
距 52w 高
-44.4%
RSI(14)
24.7
趋势
空头
SMA 20 / 50 / 200
120.91 / 129.30 / 149.19
MACD / 信号
-6.297 / -4.837
RSI 超卖空头排列

PDD

拼多多 (PDD)

$79.56 -0.38%
5 日
-2.14%
距 52w 高
-42.9%
RSI(14)
32.0
趋势
空头
SMA 20 / 50 / 200
85.43 / 93.76 / 110.59
MACD / 信号
-4.046 / -3.933
接近 52 周低空头排列

JD

京东 (JD)

$27.57 -1.22%
5 日
-1.75%
距 52w 高
-25.2%
RSI(14)
35.6
趋势
空头
SMA 20 / 50 / 200
29.07 / 30.07 / 30.25
MACD / 信号
-0.650 / -0.511
空头排列

0700.HK

腾讯控股 (0700.HK)

HK$440.20 -1.17%
5 日
-3.72%
距 52w 高
-35.5%
RSI(14)
42.4
趋势
空头
SMA 20 / 50 / 200
449.04 / 469.34 / 566.02
MACD / 信号
-4.590 / -5.040
空头排列

GC=F

黄金期货

$4,228.10 -3.00%
5 日
+3.37%
距 52w 高
-24.3%
RSI(14)
38.9
趋势
中性
SMA 20 / 50 / 200
4,376.88 / 4,558.38 / 4,436.44
MACD / 信号
-90.379 / -92.125
MACD 金叉 (1 天前)

CL=F

WTI 原油期货

$75.39 -1.82%
5 日
-14.05%
距 52w 高
-36.9%
RSI(14)
30.3
趋势
中性
SMA 20 / 50 / 200
88.42 / 94.29 / 73.66
MACD / 信号
-5.031 / -3.354

USDCNY=X

美元 / 人民币

¥6.77 +0.16%
5 日
-0.06%
距 52w 高
-6.1%
RSI(14)
33.4
趋势
空头
SMA 20 / 50 / 200
6.77 / 6.80 / 6.96
MACD / 信号
-0.013 / -0.013
接近 52 周低空头排列
风险提示

过去走势不代表未来表现,所有技术指标均基于历史数据计算,不构成投资建议。本报告仅供技术指标解读参考,请结合其他分析工具综合决策。

Australia news live: Jim Chalmers admits losing ‘political paint’ over tax reform; NSW quake near goldmine

Treasurer says carve-outs to capital gains tax discount came after ‘getting the consultation right’. Follow today’s news live • Get our breaking news email, free app or daily news podcast There were two small earthquakes in New South Wales this morning, according to Geoscience Australia. A 3.4 magni

中文摘要 澳大利亚财长Jim Chalmers承认在税改中失去“政治光彩”,资本利得税折扣的例外条款是在“正确咨询”后推出。此外,新南威尔士州今晨发生两次小地震,靠近金矿。

Critics of Russia Say This Critic Isn’t Critical Enough

Alexander Sokurov has questioned the Russian president about government repression, but he is still controversial among his country’s exiles.

中文摘要 俄罗斯导演Alexander Sokurov曾就政府镇压问题质疑普京,但其流亡同胞认为他批判力度不够,引发争议。

Man Forced Boy, 3, Into Crocodile Pen at English Zoo, Police Say

A man is being held on suspicion of attempted murder after a boy was critically injured when he ended up in a crocodile enclosure at a zoo outside Cambridge, the authorities said.

中文摘要 英国剑桥附近一家动物园发生鳄鱼袭击事件,一名3岁男童被推入鳄鱼圈后重伤,一名男子因涉嫌谋杀未遂被拘留。

Iran announces plans to bring in maritime fees for strait of Hormuz

Tehran says fees to cover cost of managing waterway will come into effect at end of 60-day negotiation period Iran has announced plans to introduce a system of maritime fees in the strait of Hormuz in two months, after the 60-day period of negotiation that has been triggered by the signing of the me

中文摘要 伊朗宣布计划在霍尔木兹海峡引入海上收费制度,称60天谈判期结束后将开始收费,以覆盖航道管理成本。

Secretary of Defense Hegseth scolds NATO allies — again

Defense Secretary Pete Hegseth blasts NATO allies as a "paper tiger" while launching a six-month review of U.S. troop deployments in Europe and pressing for Europe to take the lead on its own defense.

中文摘要 美国防长Pete Hegseth再次批评北约盟国是“纸老虎”,同时启动为期六个月的驻欧美军部署评估,敦促欧洲主导自身防务。

Thirty-five killed as gunmen attack Niger's biggest airport

Niger has been fighting a militant Islamist insurgency for a decade and in January suspected jihadists attacked the same airport.

中文摘要 尼日尔最大机场遭武装分子袭击,造成35人死亡。该国已与伊斯兰极端分子交战十年,今年1月该机场也曾遭疑似圣战分子袭击。

Iran War Live Updates: Vance Issues Blunt Warning to Israel as He Defends Trump’s Deal

The vice president delivered a rebuke to Israeli critics of the agreement, warning them not to alienate their most important ally, as the administration sought to rebut mounting criticism of the deal.

中文摘要 美国副总统Vance就伊朗协议向以色列发出严厉警告,批评以方反对者,提醒不要疏远最重要的盟友。政府正试图反驳对协议日益增长的批评。

Are prices really dropping in the US, as Trump claims?

Donald Trump boasts on social media that since signing of MoU for peace deal, oil is flowing and prices are dropping.

中文摘要 特朗普在社交媒体上声称,自签署和平协议备忘录以来,石油流通、物价下降。但文章质疑美国物价是否真的如其所言在下跌。

Fallout From the Iran Deal

Trump’s agreement reads like a list of concessions to Iran. Will it damage him politically?

中文摘要 特朗普与伊朗的协议被视为一系列对伊让步,文章探讨此举是否会对其政治造成损害。

Gunmen Attack a Major Airport in Niger

A militant group claimed responsibility for the early-morning assault on Niamey airport, the second such attack in the West African country this year.

中文摘要 尼日尔尼亚美机场遭凌晨袭击,一武装组织声称负责。这是该国今年第二次此类攻击。

Brazil executes warrant against Lula ally in widening financial scandal

Left-wing Senator Jaques Wagner was the latest prominent figure to be swept up in investigations into Banco Master.

中文摘要 巴西警方对左翼参议员Jaques Wagner执行逮捕令,他是卢拉盟友,因Banco Master金融丑闻调查扩大而被牵连。

Standard Nuclear Files for US IPO In Reactor Fuel Push

Standard Nuclear Inc., a maker of fuel for nuclear reactors, filed for an initial public offering as the power needs to fund the artificial intelligence boom continue to grow.

中文摘要 核反应堆燃料制造商Standard Nuclear Inc.提交美国IPO申请,旨在为人工智能热潮带来的电力需求筹集资金。

Asian Stocks to Rise as US-Iran Deal Lifts Mood: Markets Wrap

Stocks in Asia were poised to rise Friday after a peace deal between the US and Iran boosted optimism that the reopening of the Strait of Hormuz will ease inflation pressures by restoring oil flows.

中文摘要 受美伊和平协议提振,亚洲股市周五预计上涨。协议有望重开霍尔木兹海峡、恢复石油供应,缓解通胀压力。

Oil Set for Deep Weekly Loss as Hormuz Traffic Starts to Pick Up

Oil headed for a substantial weekly drop as the US-Iran interim peace deal saw shipping through the Strait of Hormuz start to return to normal, easing the global crude market’s biggest ever supply shock.

中文摘要 油价本周大幅下跌,因美伊临时和平协议使霍尔木兹海峡航运逐步恢复正常,缓解了全球原油市场历史上最大供应冲击。

MSCI Says Flow of Market Information Worsening in Indonesia

Information flow has deteriorated in Indonesian markets, MSCI Inc. said ahead of a potential downgrade that could further compound issues for the world’s worst-performing equities.

中文摘要 MSCI表示印尼市场信息流通恶化,面临潜在降级风险,可能进一步打击已表现全球最差的股市。

UK government to overhaul ‘outdated’ home-selling process

Vendors will need to produce ‘sales packs’ containing information on the property’s condition prior to listing

中文摘要 英国政府将改革“过时”的房屋销售流程,要求卖方在挂牌前提供包含房产状况信息的“销售包”。

Schwab Warns of Margin Calls as Concerns Over Tax Bets Mount

Charles Schwab Corp. is imposing new margin requirements for clients using long-short investment strategies, adding more guardrails around a fast-growing tax-management trade.

中文摘要 嘉信理财对使用多空策略的客户实施新保证金要求,为快速增长的税务管理交易增加更多护栏。

Gulf oil shock speeds up India’s EV drive

Carmakers report increasing sales and showroom enquiries as fuel costs rise

中文摘要 海湾石油危机加速印度电动汽车推广,汽车制造商报告称随着燃料成本上升,销量和展厅咨询量增加。

The Knicks Win! So Do New York City, Big Merch And The NBA

The New York Knicks weren’t the only big winners of this year’s NBA finals. On today’s Big Take podcast, Jason Kelly and Randall Williams tell Sarah Holder what the series means for the league’s bottom line, New York’s economy and Jalen Brunson’s future as a pitchman. Listen wherever you get your po

中文摘要 纽约尼克斯队赢得NBA总冠军,除球队外,纽约市、周边商品销售和联盟收入均从中受益。

【今朝の5本】仕事を始める前に読んでおきたい厳選ニュース

ホルムズ封鎖解除、インテルとアップル、スペースX社債発行検討、アマゾン半導体、JERA

中文摘要 今朝精选五条新闻:霍尔木兹海峡封锁解除、英特尔与苹果动向、SpaceX考虑发行债券、亚马逊半导体业务、JERA相关动态。

SpaceX’s Epic Fundraising Campaign for AI Has Only Just Begun

When Elon Musk went into sales pitch mode ahead of the SpaceX IPO, the comparison he marshaled to justify valuing the company in the trillions of dollars, and tapping investors for tens of billions, raised a few eyebrows.

中文摘要 SpaceX在IPO前展开大规模融资活动,马斯克以万亿估值和数十亿美元筹资目标说服投资者,引发关注。

Everpure CEO: Data Has to Become Primary for Businesses

Charlie Giancarlo, CEO of Everpure discusses what he sees as the evolution for companies using apps to a unified data environment. (Source: Bloomberg)

中文摘要 Everpure首席执行官Charlie Giancarlo表示,企业应优先构建统一数据环境,而非分散应用。

对公益站祛魅了

感觉公益站本来很有意思很吸引人 自己开一个公益站果然就祛魅了,不过如此 开站!hlool公益站启动!(已允许注册,深夜福利) 开发调优 本帖使用社区公益推广,符合推广要求。我申明并遵循社区要求的以下内容 本站不保证使用,这波到什么时候封了就没了,目前测试200个号左右暂时不死,也可能随时就死 站点很小是小鸡,后面看看如果撑得久就再优化下,撑不住就就此告别 登录送100刀,看情况再继续优化 不会发渠道方案方式,希望可以活的久一点 仅gpt站点 上号太慢了太低估佬友们的速率了,加大下注册机封车,稍等优化完毕 稍微等下,L… 公益站Token滞销,救救我们 这回真睡了,希望可以帮到熬夜的佬们,严肃

关于“某公益站域名续费”等话题统一集中讨论

于今日凌晨出现多位佬友发帖分享某公益站域名续费等帖子,且该情况持续 我们明白你们的喜悦之情,但是该行为已污染论坛信息流 此类话题已有对应的集中贴:https://linux.do/t/topic/1799316 如再出现对应情况我们将采取包括但不限于删除帖子,禁言等处理 感谢佬友们的理解 1 个帖子 - 1 位参与者 阅读完整话题

[SubBoost] 开源 Clash 订阅转换、增强和管理工具丨 UI 可视化配置链式代理、精确分流和防 DNS 泄露

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: (点击了解更多详细信息) 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 1. 背景介绍 SubBoost 是一个 Clash/Mihomo 订阅转换、增强和管理工具。可以将机场订阅或自建节点转换为优化、统一的订阅,并自动更新。通过 UI 可视化,一键实现链式代理 、精确分流和防 DNS 泄露等高级功能。 经过半年的更新,目前 2.4.0 版本已经趋近于稳定,可以开放给大家自部署。再也不用担心因为 IP 限制而解析订阅失败啦。 当然,我部署的 在线版 依然会继续运行。 2. 部署教程 推荐部署在大陆服务器上,

感觉生图站没什么人用啊

一堆走4k接口的生1k图,生图站可以生4k图,就是记得选分辨率: 还是生4k价格图太高了吗? 19 个帖子 - 19 位参与者 阅读完整话题

Vibe coding需要知道的设计术语——文字排版

经常看到朋友和同事在 Vibe Coding 时吐槽 ai 听不懂人话——“高级感”,“眼前一亮”,“再改改”,然后页面越来越乱。很多时候不是 AI 能力不行,更像我们是在用玄学抽卡。 LLM在pre-train和RL都用了海量的代码训练,自然英文术语的语义锚点会强于中文(而且中文语义信息熵太高了),打算长期整理一份术语速查表分享给佬友们,方便查阅 1. 字体分类与基本术语 (Font Classifications & Basics) Typography (文字排版) 文本的字体选择、间距调整、版面布局,以及最终在屏幕或纸张上的阅读呈现方式。好的排版会先给读者一个入口,再让他们知道哪里可以

[富可敌国] 0门槛拿【御三家 API】!GitHub Star>1 就能领!最高送1000美金!

[富可敌国] 0门槛拿【御三家 API】!GitHub Star>1 就能领!最高送1000美金! 核心福利(先到先得) 星火普惠:GitHub Star > 1 就能直接申请参与,0门槛直接白嫖生态基础资源! 千星专享:Repo 只要过 1000 Star,官方直接砸 价值**$100 美金** 的API 额度! 冲榜争霸:活动结束时最终冲到 Top 10 的神仙项目,还能再获得额外激励,其中Top 1 能独揽 价值**$1000 美金** 的API 额度! 不搞虚的,只为同行:拒绝短期喧嚣,这次活动就是官方掏出真金白银致敬开源,帮开发者把 API 成本直接打到 0! 活动时间:6月18日

分享一下我现在用的比较舒服的一套claudecode等agent vibe coding工作流~2026/6/18

算下来使用 claudecode 两个月了,直到最近才有了一套自己用下来比较舒服的一套工作流,来跟佬友们分享一下哈 首先使用 plan 模式先把需求想法给到 agent,可以畅所欲言想到啥说啥,但注意要跟它说一句先交流交流或者先听听你的想法,不然它有可能听完需求直接开干了哈 然后等它听完我一开始给它的需求并进行初步思考后,我就会用最近非常爱用的 grill-me 这个 skills 让它开始连环询问,帮我思考项目细节。因为跟它一开始表达的时候很难把所有细节都考虑到,也不容易把内心中那个需求的模样给说清楚。 然后它就开始不断在我的想法框架基础上问我问题,大部分情况下它给的建议都挺不错了,只需要做

祝各位佬友端午安康

回家包粽子了,我是拍照的,包粽子的是我奶奶 69 个帖子 - 64 位参与者 阅读完整话题