每日简报

2026-06-28

← 历史归档

simplex-chat/simplex-chat

Haskell · ★ 13,815 · 🍴 792 · 📈 1,469 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 是首个无需任何用户标识符(手机号、邮箱等)的通讯网络,100% 隐私设计。基于点对点架构和端到端加密,覆盖 iOS、Android 和桌面端。面向极注重隐私的用户,用于完全匿名的即时通讯。

xbtlin/ai-berkshire

Python · ★ 4,095 · 🍴 576 · 📈 685 stars today

AI 时代的伯克希尔:基于 Claude Code 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 on Claude Code. 4 masters' methodologies + multi-agent adversarial analysis.

中文介绍 基于 Claude Code 的价值投资研究框架,融合巴菲特、芒格、段永平、李录四大师方法论,用多 Agent 并行执行研究任务。面向价值投资者和股票分析师,自动化公司分析、生成投资报告,降低主观偏见。

commaai/openpilot

Python · ★ 62,065 · 🍴 11,064 · 📈 322 stars today

openpilot is an operating system for robotics. Currently, it upgrades the driver assistance system on 300+ supported cars.

中文介绍 Openpilot 是面向机器人的操作系统,现已升级 300 多款车型的驾驶辅助系统。采用计算机视觉和机器学习算法,为汽车增加自适应巡航、车道保持等功能。适合汽车爱好者和开发者,用于在支持的车辆上部署增强版辅助驾驶。

IceWhaleTech/CasaOS

Go · ★ 35,777 · 🍴 2,043 · 📈 502 stars today

CasaOS - A simple, easy-to-use, elegant open-source Personal Cloud system.

中文介绍 CasaOS 是一个简单易用的开源个人云系统,基于 Docker 容器化技术部署。提供 Web 界面管理文件、媒体、应用等。适合家庭和极客用户,用于在 NAS、树莓派等设备上快速搭建私有云盘、媒体中心。

ripienaar/free-for-dev

HTML · ★ 124,167 · 🍴 13,082 · 📈 459 stars today

A list of SaaS, PaaS and IaaS offerings that have free tiers of interest to devops and infradev

中文介绍 一份持续更新的 SaaS、PaaS 和 IaaS 免费层服务清单,涵盖数据库、监控、CI/CD 等。面向 DevOps 和基础架构开发者,用于在项目早期或预算有限时快速寻找高性价比的云服务。

google-labs-code/design.md

TypeScript · ★ 22,321 · 🍴 1,782 · 📈 1,541 stars today

A format specification for describing a visual identity to coding agents. DESIGN.md gives agents a persistent, structured understanding of a design system.

中文介绍 DESIGN.md 是一种规范格式,用于向 AI 编码代理描述视觉设计系统。通过 Markdown 文件持久化定义颜色、排版、间距等 token。适合前端团队结合 AI 编程助手,确保生成界面符合设计规范一致性。

microsoft/PowerToys

C · ★ 135,698 · 🍴 8,168 · 📈 57 stars today

Microsoft PowerToys is a collection of utilities that supercharge productivity and customization on Windows

中文介绍 Microsoft PowerToys 是 Windows 上的一系列生产力工具集合,包括窗口布局管理、批量重命名、快速颜色拾取等。面向 Windows 高级用户和开发者,用于提升日常操作效率与系统自定义能力。

hugohe3/ppt-master

Python · ★ 33,061 · 🍴 2,814 · 📈 589 stars today

AI generates a real, editable PowerPoint from any document — native shapes & animations, speaker notes voiced as audio narration, and the option to follow your own .pptx template, not slide images · by Hugo He

中文介绍 PPT-Master 用 AI 将任何文档(论文、报告等)自动生成可编辑的 PowerPoint 文件,保留原生图形、动画,并可将演讲者笔记转为音频旁白。支持自定义 .pptx 模板。适合需要快速制作专业演示文稿的学生和职场人士。

JCodesMore/ai-website-cloner-template

TypeScript · ★ 22,103 · 🍴 3,169 · 📈 750 stars today

Clone any website with one command using AI coding agents

中文介绍 一条命令即可克隆任意网站,底层调用 AI 编码代理(如 Claude Code)自动分析结构并重建。面向开发者,用于快速复制网站布局、学习前端架构或搭建原型。

garrytan/gstack

TypeScript · ★ 117,235 · 🍴 17,417 · 📈 674 stars today

Use Garry Tan's exact Claude Code setup: 23 opinionated tools that serve as CEO, Designer, Eng Manager, Release Manager, Doc Engineer, and QA

中文介绍 Garry Tan 的 Claude Code 配置集,包含 23 个严格定义的工具,扮演 CEO、设计师、工程经理等角色。面向使用 Claude Code 的团队,通过多角色 AI 代理改善代码审查、文档编写、产品设计等协作流程。

NanmiCoder/MediaCrawler

Python · ★ 53,765 · 🍴 11,009 · 📈 394 stars today

小红书笔记 | 评论爬虫、抖音视频 | 评论爬虫、快手视频 | 评论爬虫、B 站视频 | 评论爬虫、微博帖子 | 评论爬虫、百度贴吧帖子 | 百度贴吧评论回复爬虫 | 知乎问答文章|评论爬虫

中文介绍 支持爬取小红书、抖音、快手、B 站、微博、贴吧、知乎等平台的内容和评论。基于 Python 的多线程爬虫框架,用于社交媒体数据采集与分析。适用于市场研究、舆情监测、内容搬运等场景。

Anil-matcha/Open-Generative-AI

JavaScript · ★ 21,378 · 🍴 3,640 · 📈 255 stars today

Unrestricted Open-source alternative to AI video platforms — Free AI image & video generation studio with 200+ models (Flux, Midjourney, Kling, Sora, Veo). No content filters. Self-hosted, MIT licensed.

中文介绍 一个开源的 AI 图像与视频生成工作室,集成 200+ 模型(Flux、Midjourney、Kling、Sora 等),无内容过滤,支持自托管和 MIT 许可证。面向需要灵活创作的用户,用于生成不限主题的视觉内容。

topoteretes/cognee

Python · ★ 23,994 · 🍴 2,254 · 📈 780 stars today

Cognee is the open-source AI memory platform for agents. Give your AI agents persistent long-term memory across sessions with a self-hosted knowledge graph engine.

中文介绍 Cognee 是开源的 AI 记忆引擎,为 agent 提供跨会话持久长时记忆。基于自托管知识图谱存储用户历史,让 AI 对话系统记住上下文。适合构建个人助理、客服机器人等需要长期记忆的Agent 应用。

dbt-labs/dbt-core

Rust · ★ 13,215 · 🍴 2,439 · 📈 18 stars today

dbt enables data analysts and engineers to transform their data using the same practices that software engineers use to build applications.

中文介绍 dbt 核心产品,让数据分析师和工程师用软件工程实践(版本控制、测试、文档)进行数据转换。通过 SQL 和 Jinja 模板实现 ELT 流程。广泛用于数据仓库建模,提升数据团队协作效率和代码可维护性。

luongnv89/claude-howto

Python · ★ 38,616 · 🍴 4,649 · 📈 141 stars today

A visual, example-driven guide to Claude Code — from basic concepts to advanced agents, with copy-paste templates that bring immediate value.

中文介绍 一份视觉化、示例驱动的 Claude Code 指南,从基础用法到高级 Agent 实战,附带可直接复制的模板。面向初学者,用对比图和代码片段帮助快速掌握 Claude Code 的核心功能与编程模式。

anomalyco/opencode

TypeScript · ★ 179,740 · 🍴 22,099 · 📈 392 stars today

The open source coding agent.

中文介绍 一个开源的编码代理,可嵌入 IDE 或命令行,自动理解需求并生成代码。面向开发者,用于加速日常编程任务,如生成函数、修复 bug、编写测试等,提升开发效率。

Fission-AI/OpenSpec

TypeScript · ★ 57,113 · 🍴 3,984 · 📈 177 stars today

Spec-driven development (SDD) for AI coding assistants.

中文介绍 OpenSpec 推行规范驱动开发(SDD),让 AI 编码助手依据结构化规范生成代码。通过编写高层次的规范文档,确保 AI 输出符合预期。适合追求文档与代码一致的团队,用于复杂项目的 AI 辅助开发。

HKUDS/Vibe-Trading

Python · ★ 13,702 · 🍴 2,588 · 📈 92 stars today

"Vibe-Trading: Your Personal Trading Agent"

中文介绍 Vibe-Trading 是一个个人交易代理,利用 AI 分析市场情绪与技术指标,自动执行策略。面向个人投资者,用于自动化选股、择时与回测,降低人力监控成本。

keycloak/keycloak

Java · ★ 35,263 · 🍴 8,554 · 📈 20 stars today

Open Source Identity and Access Management For Modern Applications and Services

中文介绍 Keycloak 是一个开源的身份与访问管理(IAM)平台,支持 OAuth2、OpenID Connect、SAML 等标准协议。提供单点登录、用户联合、多因素认证等功能。适用于现代应用和微服务架构,统一认证与授权。

every-app/open-seo

TypeScript · ★ 3,377 · 🍴 372 · 📈 239 stars today

Open source alternative to Semrush and Ahrefs

中文介绍 Semrush 和 Ahrefs 的开源替代品,提供关键词研究、竞争对手分析、外链监控、排名追踪等 SEO 功能。适合网站所有者、SEO 人员自托管,避免付费订阅,灵活扩展。

Are We Ready For An Agent-Native Memory System?

👍 104

Memory for large language model (LLM) agents has rapidly evolved from simple retrieval-augmented mechanisms into a data management system that supports persistent information storage, retrieval, update, consolidation, and dynamic lifecycle governance throughout agent execution. Despite this evolutio

中文介绍 探讨LLM智能体的记忆系统从检索增强向原生数据管理系统的演变,支持持久存储、检索、更新、整合和动态生命周期管理,分析当前技术成熟度与挑战。

ShutterMuse: Capture-Time Photography Guidance with MLLMs

👍 44

Real-world photography requires capture-time guidance for both camera framing and subject pose. Yet existing aesthetic cropping benchmarks mainly evaluate post-hoc crop prediction and overlook subject-side recommendations, leaving the capture-time guidance capabilities of multimodal large language m

中文介绍 提出ShutterMuse,利用多模态大语言模型在拍摄时提供相机构图和主体姿势指导,填补现有裁剪基准仅关注后期预测的空白,评估其抓拍引导能力。

In-Context World Modeling for Robotic Control

👍 42

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 提出上下文世界模型进行机器人控制,解决VLA模型在相机视角或机器人形态变化时的泛化问题,通过将系统配置作为变量纳入模型。

OPID: On-Policy Skill Distillation for Agentic Reinforcement Learning

👍 40

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(在策略技能蒸馏),结合结果基RL的稳定性和自蒸馏的密集token级监督,为语言智能体提供更好的中间决策信号。

Qwen-Image-Agent: Bridging the Context Gap in Real-World Image Generation

👍 40

While text-to-image (T2I) models have achieved remarkable progress, they struggle with real-world requests that are often underspecified, implicit, or dependent on up-to-date knowledge. We identify this challenge as the Context Gap: the mismatch between the user context and the sufficient generation

中文介绍 Qwen-Image-Agent旨在弥合真实世界图像生成中的上下文差距,处理用户请求中未指定、隐含或依赖最新知识的问题。

The Verification Horizon: No Silver Bullet for Coding Agent Rewards

👍 38

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 挑战编程智能体中“验证比生成更容易”的直觉,发现当模型能力增强时,验证复杂候选解决方案变得困难,指出奖励设计无银弹。

ViQ: Text-Aligned Visual Quantized Representations at Any Resolution

👍 37

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,实现文本对齐的任意分辨率视觉量化表示,平衡离散信号的底层细节保留与高层语义对齐。

MVTrack4Gen: Multi-View Point Tracking as Geometric Supervision for 4D Video Generation

👍 34

Synthesizing a novel-view video from a monocular reference video along a target camera trajectory requires both geometric consistency and motion fidelity with respect to the reference video. Existing methods based on explicit 3D representations are limited by the accuracy of off-the-shelf reconstruc

中文介绍 MVTrack4Gen利用多视角点追踪作为几何监督,从单目参考视频生成新视角视频,提升几何一致性和运动保真度。

JetSpec: Breaking the Scaling Ceiling of Speculative Decoding with Parallel Tree Drafting

👍 29

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 JetSpec通过并行树草稿打破推测解码的缩放天花板,在保持高接受率的同时增加草稿预算,显著提升加速比。

GUI vs. CLI: Execution Bottlenecks in Screen-Only and Skill-Mediated Computer-Use Agents

👍 25

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 构建匹配的执行层基准,对比GUI与CLI两种交互模态的计算机使用智能体,揭示执行瓶颈。

V-Zero: Answer-Label-Free On-Policy Distillation with Contrastive Evidence Gating for Fine-Grained Visual Reasoning

👍 25

Fine-grained visual reasoning requires multimodal large language models (MLLMs) to identify task-relevant visual evidence and ground their reasoning in local image regions. Existing agentic methods typically rely on reinforcement learning with verifiable rewards or supervised fine-tuning on large-sc

中文介绍 V-Zero采用无答案标签的在策略蒸馏与对比证据门控,实现细粒度视觉推理,无需大规模标注数据。

UnityShots: Memory-Driven Multi-Shot Audio-Video Generation with Boundary-Aware Gating

👍 24

Generating a coherent multi-shot video requires structured cross-shot memory. Subject appearance, scene context, and speaker identity must persist across cuts. Existing approaches either train end-to-end over fixed-length sequences and cannot scale, generate shot-by-shot with memory banks that grow

中文介绍 UnityShots通过记忆驱动和边界感知门控生成连贯多镜头音频视频,解决跨镜头主体外观、场景和说话人身份的一致性。

Fast LeWorldModel

👍 22

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 Fast LeWorldModel加速联合嵌入预测架构,通过优化局部一步潜变量过渡模型,提升视觉规划中的候选动作序列评估效率。

Running the Gauntlet: Re-evaluating the Capabilities of Agents Beyond Familiar Environments

👍 15

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 重新评估智能体在熟悉环境之外的能力,指出现有基准任务简单且能力范围狭窄,提出更全面的评测。

Why Multi-Step Tool-Use Reinforcement Learning Collapses and How Supervisory Signals Fix It

👍 15

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 分析多步工具使用强化学习坍塌的原因,并提出监督信号修复,避免RL单独训练导致的灾难性崩溃。

LISA: Likelihood Score Alignment for Visual-condition Controllable Generation

👍 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 LISA通过似然分数对齐改进视觉条件可控生成中的双分支范式,分析侧网络中间层特征融合的作用。

Autodata: An agentic data scientist to create high quality synthetic data

👍 11

We introduce Autodata, a general method that enables AI agents to act as data scientists who build high quality training and evaluation data. We show how to train (meta-optimize) such a data scientist agent, so that it learns to create even stronger data. We describe the overall formulation, and a s

中文介绍 Autodata让AI智能体扮演数据科学家角色,通过元优化学会构建高质量合成训练/评估数据。

Information-Aware KV Cache Compression for Long Reasoning

👍 9

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缓存压缩方法,超越传统注意力权重估计,更有效压缩长推理场景中的KV缓存。

Confidence-Aware Tool Orchestration for Robust Video Understanding

👍 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,解决视频推理模型对扰动帧的盲目信任问题,在真实场景扰动下提升15-30%准确率。

PhysiFormer: Learning to Simulate Mechanics in World Space

👍 9

We present PhysiFormer, a diffusion transformer for physically-plausible 3D object motion. Unlike video world models that operate in view-dependent pixel space, PhysiFormer represents objects as 3D meshes expressed in world coordinates. Given the initial vertex positions and velocities, as well as o

中文介绍 PhysiFormer是一种扩散Transformer,在世界坐标中表示3D网格物体,学习物理合理的运动模拟,给定初始顶点位置、速度和物体属性。

27 Hidden Claude Features, Settings & Shortcuts That Most Users Don't Know

@sairahul1 · 121.4K 粉丝 · 2.9M 阅 · 506 赞 · 97 转

Most people use Claude Code like a fancy autocomplete. They prompt. They wait. They accept the first output. They are leaving 90% of the power on the table. There are 27 specific moves that separate

中文介绍 Claude Code 的 27 个隐藏功能、设置和快捷键,帮助用户摆脱简单自动补全的局限,释放 Claude 90% 的潜力。

How To Build a One-Person Company Using Claude Cowork

@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 构建一人公司,聚焦知识工作中占 60% 时间的非核心任务(邮件、报告、整理等),实现自动化。

ORACLE: Official AI Agents Trade on Polymarket

@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 成为预测市场(Polymarket)的顶级策略,超过 30% 活动来自算法与 AI 钱包。推出官方 Oracle 代理进行交易。

ORACLE: Official AI Agents Trade on Polymarket

@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,自主 AI Agent 成为预测市场(Polymarket)的顶级策略,超过 30% 活动来自算法与 AI 钱包。推出官方 Oracle 代理进行交易。

This Week on Base: New Base MCP skills + now live in Perplexity, Base App on desktop, and more!

@base · 1.3M 粉丝 · 93.3K 阅 · 501 赞 · 127 转

News ■ Base MCP's new skills let agents transact, trade, lend, mint, and buy onchain across 13 apps: @yield @AskVenice @KyberNetwork @opensea @o1_exchange @Balancer @printr @bitrefill @flaunchgg

中文介绍 Base 本周更新:Base MCP 新增技能,Agent 可在 13 个应用(Yield、Venice、OpenSea 等)执行链上操作;Base App 桌面版上线并集成 Perplexity。

thoughts on why mcp didn't work, what's next

@RhysSullivan · 57.2K 粉丝 · 86.1K 阅 · 503 赞 · 25 转

mcp came out when the best models were sonnet 3.5 and GPT 4o not a lot was known about how to properly work with these tools yet, we were still incredibly concerned on models having access to tools,

中文介绍 反思 MCP(模型上下文协议)为何未达预期:推出时模型能力有限(Sonnet 3.5、GPT-4o),工具使用经验不足。展望下一步发展方向。

Life After Switching to Kimi

@0xDevin_ · 6.6K 粉丝 · 38.3K 阅 · 539 赞 · 5 转

Most AI tools are chatbots with a nice interface. Kimi is different. It is a full system: a browser automation engine called Claw that navigates websites like a human, an Agent Swarm that runs

中文介绍 切换到 Kimi AI 后的体验:Kimi 不只是一个聊天界面,而是包含浏览器自动化引擎 Claw 和 Agent Swarm 的完整系统,能像人类一样操作网页和执行复杂任务。

Karpathy's Agentic Engineering Finally Has Proper Tooling

@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 Google 构建了 Karpathy 提出的智能体工程(Agentic Engineering)的配套工具,并给出逐步指南。该概念将生产级 Agent 与氛围编码(vibe coding)区分开来,是 2026 年关键趋势。

Human in the /loop

@ericzakariasson · 76.0K 粉丝 · 32.3K 阅 · 518 赞 · 29 转

What I like most about coding with agents right now is the room to leave a few runs going and still get on with other work. When something finishes or needs a call, I show up. This post is a short

中文介绍 博主分享当前与 Agent 协作编码的感受:可以同时运行多个任务,只在需要时介入。这种 human-in-the-loop 模式提升了效率,能并行处理多个 Agent 任务。

Previewing GPT-5.6 Sol: a next-generation model

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,在编程、科学和网络安全方面能力更强,并搭配最先进的安全堆栈。

not much happened today

**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的三个变体:旗舰版Sol、中端Terra、低成本Luna。受美国政府限制,仅向可信合作伙伴推出,Sol版本强化了网络安全和安全特性。

Repositioning retail for the AI era

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 人工智能正在重塑零售业,最大变化不在消费者端的虚拟试穿或聊天助手,而是在后台决策方式,如产品搜索结果排序等。

not much happened today

**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在编程和智能体基准测试中领先,Code Arena前端得分1595,推理准确率34.29%且零失败。Databricks将其速度提升至392 tok/s。新模型Ornith-1.0采用MIT许可。

[AINews] It's Meta-Harness Summer

Move over, Harness Engineering, it is time for the harness of harnesses!

中文介绍 Latent Space文章称“元交配”夏季到来,强调交配工程之外的交配之交配。

How agents are transforming work

A new OpenAI research paper shows how AI agents are transforming work, enabling longer, more complex tasks and expanding productivity across roles.

中文介绍 OpenAI新研究论文显示AI智能体正在改变工作方式,能够执行更长更复杂的任务,并提升各岗位生产力。

市场总览

当前市场整体技术面偏弱,资产间分化明显。美股方面,标普500 ETF(SPY)和纳斯达克100 ETF(QQQ)均处于短期均线下方,RSI在46.5和49.9中性区域,短期内未有明确方向信号,但QQQ保留了多头排列结构。科技巨头如AAPL、MSFT、NVDA等普遍下跌,其中MSFT RSI 28.7进入超卖区且呈空头排列,显示短期动能疲弱。加密市场受恐慌贪婪指数仅18(极度恐慌)压制,总市值2.15T,24小时微跌0.21%,BTC主导率55.8%、ETH占8.8%;BTC-USD、ETH-USD均处于所有均线下方且MACD死叉,RSI在32附近,延续下行趋势。中概股承压明显,阿里巴巴RSI跌至16.7极度超卖,拼多多、京东同步超卖,技术形态均呈空头排列。商品市场黄金期货RSI 36.2处于下方区域,原油期货RSI 27.1超卖,短期面临压力。外汇方面,美元指数DXY受RSI 72.8超买及多头排列支撑,接近52周高位,而美元/人民币则呈现空头排列。宏观经济指标中,VIX恐慌指数位于18.41,RSI 52.6中性,MACD金叉暗示波动性可能提升;10Y美债收益率低于20日均线,表现中性。整体而言,偏下行资产占主导,仅有美元指数等少数标的呈现偏上行信号。

今日关注

DX-Y.NYB 美元指数 DXY
偏上行

当前价格101.37高于所有均线(sma20 100.02,sma50 99.15,sma200 98.75),RSI 72.8处于超买区,MACD值为0.5985高于信号线0.4306,趋势信号为bullish且呈多头排列,同时接近52周高,技术面显示强势上行动能。

BABA 阿里巴巴 (BABA)
偏下行

价格94.81远低于各均线(sma20 115.10,sma50 127.06,sma200 148.48),RSI跌至16.7极度超卖区域,MACD值为-8.0993低于信号线-6.3731,趋势bearish且为空头排列,近5日跌幅11.76%,技术面显示强烈下行压力。

SPY S&P 500 ETF
中性

价格728.99位于sma20(744.88)下方但高于sma200(690.14),sma50(733.77)附近有支撑,RSI 46.5处于正常区间,MACD 0.8288低于信号线3.3559但绿柱未明显放大,趋势判定为neutral,且无明确信号触发,技术状态无明显方向倾向。

全部资产

^VIX

VIX 恐慌指数

$18.41 -1.18%
5 日
-0.16%
距 52w 高
-47.8%
RSI(14)
52.6
趋势
中性
SMA 20 / 50 / 200
17.79 / 17.78 / 18.63
MACD / 信号
0.143 / 0.009
MACD 金叉 (2 天前)

^TNX

10Y 美债收益率 (%)

$4.37 -2.04%
5 日
-2.04%
距 52w 高
-12.5%
RSI(14)
47.4
趋势
中性
SMA 20 / 50 / 200
4.49 / 4.43 / 4.21
MACD / 信号
0.004 / 0.017
接近 52 周低

DX-Y.NYB

美元指数 DXY

$101.37 -0.24%
5 日
+1.27%
距 52w 高
-0.4%
RSI(14)
72.8
趋势
多头
SMA 20 / 50 / 200
100.02 / 99.15 / 98.75
MACD / 信号
0.598 / 0.431
RSI 超买接近 52 周高多头排列

SPY

S&P 500 ETF

$728.99 -0.58%
5 日
-1.62%
距 52w 高
-4.1%
RSI(14)
46.5
趋势
中性
SMA 20 / 50 / 200
744.88 / 733.77 / 690.14
MACD / 信号
0.829 / 3.356

QQQ

Nasdaq 100 ETF

$706.52 -0.58%
5 日
-2.21%
距 52w 高
-5.6%
RSI(14)
49.9
趋势
多头
SMA 20 / 50 / 200
726.22 / 701.41 / 631.52
MACD / 信号
5.105 / 8.563
多头排列

AAPL

Apple

$283.78 -3.17%
5 日
-4.11%
距 52w 高
-10.6%
RSI(14)
32.2
趋势
中性
SMA 20 / 50 / 200
299.73 / 291.15 / 269.21
MACD / 信号
-1.588 / 1.199

MSFT

Microsoft

$372.97 +2.05%
5 日
-1.57%
距 52w 高
-32.9%
RSI(14)
28.7
趋势
空头
SMA 20 / 50 / 200
402.81 / 411.74 / 448.61
MACD / 信号
-14.234 / -8.797
RSI 超卖空头排列

NVDA

Nvidia

$192.53 -3.25%
5 日
-5.92%
距 52w 高
-18.6%
RSI(14)
39.6
趋势
中性
SMA 20 / 50 / 200
208.85 / 210.21 / 190.53
MACD / 信号
-2.940 / -1.317

GOOGL

Alphabet

$337.39 -2.29%
5 日
-7.26%
距 52w 高
-17.4%
RSI(14)
36.2
趋势
中性
SMA 20 / 50 / 200
363.45 / 369.27 / 313.36
MACD / 信号
-6.254 / -3.301

TSLA

Tesla

$379.71 +1.11%
5 日
-4.21%
距 52w 高
-23.9%
RSI(14)
38.9
趋势
空头
SMA 20 / 50 / 200
404.67 / 405.01 / 417.80
MACD / 信号
-7.383 / -3.473
空头排列

META

Meta

$550.25 -1.33%
5 日
-3.05%
距 52w 高
-30.9%
RSI(14)
34.1
趋势
空头
SMA 20 / 50 / 200
587.54 / 615.32 / 651.10
MACD / 信号
-15.724 / -12.405
空头排列
加密恐慌贪婪
18
极度恐慌
加密总市值
$2.15 T
-0.21% / 24h
BTC 主导率
55.8%
ETH 8.8%
24h 成交量
$44.2 B
活跃币 17,441

BTC-USD

Bitcoin

$60,041.12 +0.04%
5 日
-6.12%
距 52w 高
-52.4%
RSI(14)
32.0
趋势
空头
SMA 20 / 50 / 200
63,051.86 / 69,828.18 / 75,837.99
MACD / 信号
-2,315.018 / -2,294.372
MACD 死叉 (今天)空头排列

ETH-USD

Ethereum

$1,573.87 -0.17%
5 日
-8.84%
距 52w 高
-68.2%
RSI(14)
31.9
趋势
空头
SMA 20 / 50 / 200
1,680.75 / 1,905.91 / 2,318.88
MACD / 信号
-79.249 / -78.231
MACD 死叉 (今天)空头排列

SOL-USD

Solana

$70.54 -1.80%
5 日
-1.90%
距 52w 高
-72.1%
RSI(14)
46.9
趋势
空头
SMA 20 / 50 / 200
69.61 / 77.66 / 95.47
MACD / 信号
-1.511 / -2.059
空头排列

BABA

阿里巴巴 (BABA)

$94.81 -5.00%
5 日
-11.76%
距 52w 高
-50.8%
RSI(14)
16.7
趋势
空头
SMA 20 / 50 / 200
115.10 / 127.06 / 148.48
MACD / 信号
-8.099 / -6.373
RSI 超卖空头排列

PDD

拼多多 (PDD)

$76.55 +1.07%
5 日
-4.14%
距 52w 高
-45.1%
RSI(14)
24.0
趋势
空头
SMA 20 / 50 / 200
81.84 / 91.76 / 109.62
MACD / 信号
-4.636 / -4.210
RSI 超卖空头排列

JD

京东 (JD)

$25.39 -0.35%
5 日
-9.03%
距 52w 高
-31.1%
RSI(14)
24.4
趋势
空头
SMA 20 / 50 / 200
28.17 / 29.79 / 30.14
MACD / 信号
-1.103 / -0.770
RSI 超卖空头排列

0700.HK

腾讯控股 (0700.HK)

HK$411.80 -2.28%
5 日
-6.45%
距 52w 高
-39.7%
RSI(14)
35.8
趋势
空头
SMA 20 / 50 / 200
445.59 / 461.46 / 561.50
MACD / 信号
-10.926 / -7.749
MACD 死叉 (4 天前)接近 52 周低空头排列

GC=F

黄金期货

$4,096.30 +1.63%
5 日
-3.03%
距 52w 高
-26.7%
RSI(14)
36.2
趋势
中性
SMA 20 / 50 / 200
4,271.85 / 4,488.10 / 4,448.21
MACD / 信号
-121.756 / -106.658
MACD 死叉 (4 天前)

CL=F

WTI 原油期货

$69.23 -3.74%
5 日
-9.62%
距 52w 高
-42.1%
RSI(14)
27.1
趋势
中性
SMA 20 / 50 / 200
83.24 / 91.99 / 73.90
MACD / 信号
-6.534 / -5.201
RSI 超卖

USDCNY=X

美元 / 人民币

¥6.79 -0.00%
5 日
+0.31%
距 52w 高
-5.9%
RSI(14)
56.5
趋势
空头
SMA 20 / 50 / 200
6.77 / 6.79 / 6.95
MACD / 信号
-0.002 / -0.007
接近 52 周低空头排列
风险提示

过去走势不代表未来表现,本报告仅基于公开行情数据的技术指标解读,不构成投资建议。任何交易决策应结合自身风险承受能力独立判断。仅供技术指标解读参考。

Iran war live: Trump threatens Tehran as US bombs Sirik, Qeshm for 2nd day

US president threatens to 'militarily complete the job' as the US strikes Sirik and Qeshm Island over ship attacks.

中文摘要 美国总统特朗普威胁将“从军事上完成任务”,美军连续第二天空袭伊朗锡里克和格什姆岛,回应船只遇袭事件。

Mideast Live Updates: U.S. Strikes Iran Again, Testing Truce

U.S. forces carried out retaliatory airstrikes on Iran for the second straight night, saying the latest barrage was in retaliation for an attack on an oil tanker on Saturday.

中文摘要 美军连续第二晚对伊朗发动报复性空袭,称此次打击是回应周六一艘油轮遭袭事件,考验停火前景。

Here’s the latest.

中文摘要 该实时更新条目暂无详细报道。

Australia news live: shark-spotting drones to patrol NSW beaches year-round; Chalmers grilled on ‘widow’s tax’

Premier Chris Minns says technology will be largest expansion of aerial shark surveillance in the world. Follow the day’s updates live Get our breaking news email, free app or daily news podcast Big tech not cooperating ‘as much as we’d like them to’: Watt Federal cabinet minister Murray Watt, says

中文摘要 新南威尔士州州长克里斯·明斯宣布鲨鱼监测无人机将在该州海滩全年巡逻,称这是全球最大规模无人机鲨鱼监视扩展;财政部长查默斯就“寡妇税”接受质询。

Here’s the latest.

中文摘要 该实时更新条目暂无详细报道。

12 Hours With Venezuelan Doctors Searching for Earthquake Survivors

A medical team set out from the capital to rescue people in the hardest-hit part of the disaster zone, La Guaira. They found silence in the ruins instead.

中文摘要 一组医疗队从首都出发前往重灾区拉瓜伊拉搜寻幸存者,却在废墟中只找到寂静;报道记录了12小时救援过程。

A Rush of Volunteers in Venezuela Has Slowed Rescue Efforts

The outpouring of volunteer aid after Venezuela’s earthquakes clogged the only road into the disaster zone, delaying rescue crews.

中文摘要 委内瑞拉地震后大量志愿者涌入,堵塞了唯一通往灾区的道路,反而延误了救援队伍的到达。

US launches second night of strikes on Iran after ship hit by drone

The attacks come after an alleged Iranian drone struck another commercial vessel in the Strait of Hormuz on Saturday.

中文摘要 美军连续第二晚对伊朗发动打击,此前一架疑似伊朗无人机在霍尔木兹海峡袭击了另一艘商船。

Israel strikes southern Lebanon as Hezbollah condemns new deal

Strikes have killed at least one person, state media say a day after Lebanon and Israel signed a framework agreement.

中文摘要 以色列空袭黎巴嫩南部至少造成一人死亡;此前一天黎巴嫩和以色列签署框架协议,真主党予以谴责。

Venezuela races to find survivors

The search for survivors is entering a critical stage in Venezuela. NPR'S Eyder Peralta reports from La Guaira where rescue crews are still hoping to find people alive.

中文摘要 委内瑞拉搜寻幸存者进入关键阶段;NPR记者从拉瓜伊拉报道,救援队仍希望找到生还者。

UK’s Bridgepoint nears deal to buy real estate unit in bet on US property

London-based buyout group is looking beyond its traditional speciality in corporate buyouts

中文摘要 总部位于伦敦的私募股权集团Bridgepoint接近达成收购一处房地产部门的协议,押注美国房地产市场,以拓展其传统的企业收购业务。

Warsh Makes Debut on the Global Stage With the Class of 2008

Kevin Warsh and three other veterans of the 2008 global financial crisis will share a stage this week as the danger of renewed turmoil continues to haunt central bankers.

中文摘要 凯文·沃什与另外三位2008年全球金融危机资深人士本周共同亮相,在央行官员持续担忧新一轮动荡之际登上全球舞台。

Tech Equity Sales Renew AI Debt-Binge Worries

Tech companies are selling stock like it’s the dot-com boom, and some investors fear that’s a bad sign for bondholders.

中文摘要 科技公司正以类似互联网泡沫时期的速度发售股票,一些投资者担忧这对债券持有人构成危险信号,引发对AI债务狂潮的重新忧虑。

Bloomberg This Weekend 6/27/2026

The news doesn’t stop when markets close. Hosts David Gura, Christina Ruffini and Lisa Mateo bring clarity, context and a bit of humor to the weekend’s biggest headlines, LIVE from New York. Joined by Fmr. UN Ambassador and Fmr. USAID Administrator Samantha Power and Oklahoma Senator Alan Armstrong.

中文摘要 Bloomberg This Weekend节目于2026年6月27日播出,主持人大卫·古拉、克里斯蒂娜·鲁菲尼和丽莎·马特奥从纽约直播解读周末重大新闻,嘉宾包括前联合国大使萨曼莎·鲍尔等。

Pointed News Quiz | Reddit, Music, YA Novels

David Gura, Christina Ruffini, and Lisa Mateo of “Bloomberg This Weekend” play Pointed! Wager your points, leverage your bets and answer wisely. A new quiz is available to play each week on Bloomberg.com (Source: Bloomberg)

中文摘要 Bloomberg This Weekend主持人参与「Pointed!」新闻问答游戏,每周在Bloomberg.com更新,本集主题涉及Reddit、音乐和青少年小说。

ECB’s Schnabel Sees Upside Inflation Risks Despite Peace Deal

European Central Bank Executive Board member Isabel Schnabel warned that price pressures could turn out stronger than anticipated even as a US-Iran peace deal reopens the Strait of Hormuz.

中文摘要 欧洲央行执委伊莎贝尔·施纳贝尔警告称,尽管美伊和平协议重新开放霍尔木兹海峡,但价格压力可能强于预期,通胀面临上行风险。

Sen. Armstrong: Permitting Is Key to Lower Energy Costs

Senator Alan Armstrong (R-Oklahoma) joins David Gura and Christina Ruffini on Bloomberg This Weekend and says permitting reform—not additional federal spending—is the most effective way to lower housing and energy costs, arguing regulatory delays have created infrastructure bottlenecks that drive up

中文摘要 俄克拉荷马州共和党参议员艾伦·阿姆斯特朗表示,许可改革而非额外联邦支出是降低住房和能源成本的最有效途径,认为监管延误造成了基础设施瓶颈。

America’s Farmers Need USMCA More Than Ever

For many American farmers, Canada and Mexico have become indispensable export markets at a time when trade disputes, weak commodity prices, and rising costs are already straining the agricultural economy. Iowa farmer Stu Swanson says many producers are operating on dwindling hope as financial pressu

中文摘要 对于许多美国农民来说,加拿大和墨西哥已成为不可或缺的出口市场。爱荷华州农民斯图·斯旺森表示,在贸易争端、商品价格疲软和成本上升的压力下,许多生产者正带着日渐减少的希望经营。

US Pet Market Set to Surpass $250 Billion

Americans are increasing spending on their pets beyond traditional preventative care, with the US pet market projected to grow from approximately $150 billion today to over $250 billion in the next decade. Bloomberg Intelligence consumer staples analyst Diana Rosero-Pena joins Bloomberg This Weekend

中文摘要 美国宠物市场预计将从目前的约1500亿美元增长至未来十年超过2500亿美元。美国人在宠物上的支出已超出传统预防性护理范畴,Bloomberg Intelligence分析师就此发表看法。

Venezuela Struggles with Earthquake Rescue Efforts as Death Toll Nears 1,000

Bloomberg News Andean Bureau Reporter Andreina Itriago is on Bloomberg This Weekend explaining concerns that the official casualty figures in Venezuela after the earthquake are likely underreported, with many more deaths expected as rescue efforts continue and bodies are recovered. Itriago explains

中文摘要 委内瑞拉地震死亡人数逼近1000人,救援工作仍在进行。Bloomberg新闻记者表示官方伤亡数字可能被低估,随着救援进展预计还有更多死亡。

AI’s Energy Crunch Has Investors Searching for Next IPO Winners

The artificial intelligence boom has a power problem, and Wall Street is betting billions on companies that promise to solve it — even if some of the technology hasn’t been fully developed yet.

中文摘要 人工智能热潮带来能源问题,华尔街正押注数十亿美元于承诺解决该问题的公司,尽管其中一些技术尚未完全开发,投资者在IPO热潮中寻找下一个赢家。

Schrödinger's Strait of Hormuz: Open or Closed?

Despite a ship being struck in the Strait of Hormuz on Thursday and another incident over the weekend, some vessels continue to travel the strait on both the Iranian and Omani sides. However, overall traffic remains significantly reduced compared to pre-conflict levels, with daily ship movements dow

中文摘要 尽管霍尔木兹海峡周四发生船只被袭事件,周末又有另一起事故,但部分船只仍从伊朗和阿曼两侧通行。不过总体交通量较冲突前大幅减少,每日通行量下降。

2026年最新免费.edu教育邮箱 【帕克大学申请图文教程】

帕克大学在线申请教程 更新时间:2026 年 6 月 28 日 一、先看清楚这几件事 申请前,请先记住下面几点:* 请使用你自己能登录的个人邮箱。 帕克大学是一所获得美国正规高等教育认证的大学,自1941年起便持续获得高等教育委员会(HLC)认证。通过其在线证书项目完成申请后,申请人会获得真实的学生学号,并配套开通一个以 @park.edu 结尾的学校邮箱。 二、申请前要准备什么 在打开申请页面之前,建议先打开下面的美国地址生成器网站把资料先生成。这样填写表格时直接复制粘贴即可。 你需要准备: 一个真实可用的个人邮箱 一个自己能记住的账户密码 使用到的工具: 美国地址生成器 https://g

LDLive 3.0 正式上线!

从 LDLive - 人人都可以直播 继续讨论 [!tip] 因为之前比较忙没时间维护,最近稍微有段时间给这个老项目重启并优化了 [!info] 跳转域名: LDLive - 智能线路分配 [!info] 最新域名: LDLive - 人人都可以直播 [!tip] 可在底下反馈报错或建议 22 个帖子 - 13 位参与者 阅读完整话题

《万事不通·L站指南》

[!todo]+ 真诚、友善、团结、专业,共建你我引以为荣之社区。《社区准则》 [!note]+ 简介 《万事不通·L站指南》是一个汇总L站各种指南的汇总帖,涉及内容包含信任等级、LDC 和 CDK、公益站等指南;无论新佬还是旧佬,都会需要查询资料,可以凭借一帖查资料;佬有所需。 传道授业解惑也! 一、入门指南 作为萌新,佬友需要了解最基础的社区知识,包含信任等级、LDC、CDK 等。 [!example]+ 目录 (1)搜索与书签 (2)信任等级 (3)LDC (4)CDK (点击了解更多详细信息) 二、公益指南 汇总公益站以及一些 AI 模型工具。 [!example]+ 目录 (1)公益

【公益站】猫猫公益站-众人拾柴火焰高

【公益站】猫猫公益站--渠道是老黄的免费key 福利羊毛 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: 我的项目是免费使用的,无收费(变相收费、赞助)部分: 是 我的帖子已经打上 公益推广 标签: 是 我的项目属于个人项目,与公司或商业机构无关: 是 我的项目不存在QQ、TG等群组引流: 是 我的项目不存在非运营必要的网站引流: 是 / 否 我的项目不存在为他人推广、AFF: 是 我的项目无关联的商业项目: 是 我的站点存在登录… 站点:https://new-api.rugao.me 渠道共享 加入映射了,可能要重新配置 148 个帖子 - 87 位参与者 阅读完整话

我去, 阿里也下场 输入法了

阿里千问输入法 macOS 版今日上线官网,支持最快 300 字/分的 AI 语音输入,可自动润色、将口语转为工整文字,并支持 9 种方言,纯净无广告。官方预告 iOS、Android、Windows 版将于近日发布。此前千问团队已于今年 5 月推出千问语音输入法(千问 App 内的组件),具备去语气词、纠错、格式化整理及基于上下文的智能回复等能力,而本次上线的输入法则定位为独立 App,填补千问在移动端 AI 输入法赛道的空白。 88 个帖子 - 78 位参与者 阅读完整话题

【公益站】猫猫公益站--渠道是老黄的免费key

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: 我的项目是免费使用的,无收费(变相收费、赞助)部分: 是 我的帖子已经打上 公益推广 标签: 是 我的项目属于个人项目,与公司或商业机构无关: 是 我的项目不存在QQ、TG等群组引流: 是 我的项目不存在非运营必要的网站引流: 是 / 否 我的项目不存在为他人推广、AFF: 是 我的项目无关联的商业项目: 是 我的站点存在登录,并已接入 LINUX DO Connect: 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 最近脑袋大开,创建了几十个英伟达免费

Openai是要倒闭了吗?

Bug多到难以置信,codex疯狂烧磁盘 然后cpu爆满 电脑差点废了 这垃圾公司天天在更新什么 每次更新一个版本就出严重问题 51 个帖子 - 44 位参与者 阅读完整话题

...朋友用一种匪夷所思的方式把密码都丢了

朋友被內容農廠和人工智能坑騙,把密碼丟了 情況如下: Ta沒有計算機知識,也不知道密碼管理器什麼的 他打算歸檔某帳號的密碼上×度網盤存著 但他想加密一下密碼: 搜索 “網站怎麼加密密碼”,得到某内容农场之文章(MD5加密) 發現 “Hash可以加密密碼” 實際上Hash加密是錯誤的表達,哈希是不可逆的,像Base64加密這樣的說法也是錯誤的 帶著錯誤的預設 (Hash加密)問AI 於是,他 把密碼全“加密”成了Hash 幾個月後,他才發現密碼全勾吧完蛋了 我聽聞後大受震撼 人类还有救嗎 46 个帖子 - 34 位参与者 阅读完整话题

哔哩哔哩 17 周年庆,陈睿称 B站社区是优质内容最好的土壤

哔哩哔哩近日迎来成立 17 周年,董事长兼 CEO 陈睿在周年庆现场发表《让好内容发生》主题演讲。他表示,尽管当下内容供给过剩,用户对优质内容的需求依然旺盛,而 B站社区正是培育优质内容的最佳土壤。 陈睿认为,短期流量只能吸引注意力,唯有留下记忆、形成心智才能实现长期留存,因此优质内容是平台持续增长的最有效方式。官方数据显示,B站五分钟以上视频观看时长占总时长的 65%,持续更新超过 5 年的 UP 主达 167 万人;过去一年,UP 主充电收入增长 71.1%,付费充电用户达 1063 万人。陈睿称,B站将坚持服务 UP 主、做好社区,让优质内容持续产生。 finance.sina.com.