2026-07-09

RynnWorld-Teleop: An Action-Conditioned World Model for Digital Teleoperation

👍 67

07/07 08:00

Scaling robot learning requires massive, diverse trajectory data, yet collection is currently bottlenecked by physical teleoperation, where every demonstration binds operator time to specific hardware and workspaces. We introduce digital teleoperation, a paradigm that decouples data collection from

中文介绍 提出数字遥操作范式RynnWorld-Teleop，使用动作条件世界模型将机器人数据收集与物理硬件解耦，实现跨工作空间的大规模多样化轨迹生成，突破数据采集瓶颈。

Hierarchical Sparse Attention Done Right: Toward Infinite Context Modeling

👍 36

07/03 08:00

Scaling modern large language models (LLMs) to long contexts is limited by the quadratic computation cost, and poor length extrapolation of dense attention. Chunk-wise sparse attention offers a promising alternative, but all existing methods fall short of full attention because of their inaccurate c

中文介绍 提出层级稀疏注意力机制，解决长上下文LLM中二次计算代价和长度外推不足的问题，实现高效无限上下文建模。

Vision as Unified Multimodal Generation

👍 30

07/07 08:00

We formulate computer vision as unified multimodal generation, where heterogeneous visual tasks are expressed in the native text and image generation spaces of a unified multimodal model, without task-specific architectures. Under this formulation, SenseNova-Vision uses natural-language instructions

中文介绍 将计算机视觉表述为统一多模态生成任务，SenseNova-Vision在统一模型的文本和图像生成空间处理各类视觉任务，无需任务特定架构。

Bridging Interleaved Multi-Modal Reasoning as a Unified Decision Process

👍 29

07/04 08:00

Unified multi-modal models (UMMs) have shown promising interleaved text-image reasoning capabilities, yet effectively optimizing such multi-turn generation via reinforcement learning (RL) remains an open challenge. Existing approaches apply RL exclusively to text steps, relegating image generation t

中文介绍 将交错多模态推理建模为统一决策过程，突破现有RL仅优化文本步骤的局限，联合优化文本与图像生成，提升多轮多模态生成效果。

Gemma 4 Technical Report

👍 26

07/02 08:00

We introduce Gemma 4, a new generation of open-weight, natively multimodal language models in the Gemma model family. Designed to advance compute efficiency and reasoning, the Gemma 4 model suite features dense and Mixture-of-Experts architectures, ranging from 2.3B to 31B parameters. Alongside impr

中文介绍 介绍Gemma 4系列开放权重原生多模态语言模型，包含Dense和MoE架构（2.3B–31B参数），在保持推理效率的同时显著提升计算效率与多模态能力。

Light-Omni: Reflex over Reasoning in Agentic Video Understanding with Long-Term Memory

👍 19

07/06 08:00

Agentic video understanding equips models with long-term memory to autonomously process and respond to continuous, long-horizon multimodal streams. However, advanced video agents often rely on ``detective-style'' iterative reasoning for action control (e.g., search) and evidence aggregation, incurri

中文介绍 提出Light-Omni，用反射机制替代侦探式迭代推理，配合长时记忆实现高效代理视频理解，大幅降低推理开销与响应延迟。

DSpark: Confidence-Scheduled Speculative Decoding with Semi-Autoregressive Generation

👍 17

07/06 08:00

Speculative decoding accelerates Large Language Model (LLM) inference by decoupling draft generation from target verification. While recent parallel drafters efficiently propose long token sequences in a single forward pass, they suffer from rapid acceptance decay due to a lack of inter-token depend

中文介绍 提出DSpark，通过置信度调度的半自回归生成，解决并行草稿推测解码中接受率快速衰减问题，加速LLM推理。

SkillOpt-Lite: Better and Faster Agent Self-evolution via One Line of Vibe

👍 17

07/03 08:00

While skill optimization for autonomous agents has gained traction, existing methods rely on complex pipelines. This leaves a fundamental question unaddressed: What constitutes a minimal viable pipeline for skill optimization, where every component is justified by theory or empirical necessity? We f

中文介绍 提出SkillOpt-Lite，以最简可行流水线（一行「vibe」）实现智能体技能自我优化，比现有复杂管道更快更好，理论必要组件均有依据。

Parallelized Autoregressive Decoding for Omni-Modal Dense Video Captioning

👍 16

07/03 08:00

Dense video captioning aims to generate temporally grounded descriptions of video events, benefiting both event-level video understanding and generation. In this domain, autoregressive video large language models have emerged as a prevalent paradigm due to their strong generative and cross-modal mod

中文介绍 提出并行化自回归解码方法，用于全模态密集视频描述，加速事件级别视频理解与生成，提升多模态大模型在时序定位描述上的效率。

Look Before You Leap: Distilling Tree Search into Action Evaluation for Frozen VLA Models

👍 15

07/04 08:00

Vision-Language-Action (VLA) models acquire broad embodied capabilities through large-scale pretraining, yet their generalization remains far more fragile than that of LLMs and VLMs. The prevailing remedy, post-training via supervised fine-tuning or reinforcement learning, improves task-specific per

中文介绍 提出将树搜索蒸馏为行动评估的方法，适用于冻结VLA模型，无需微调即可提升泛化能力，以最小代价改善机器人动作生成。

From Foundation to Application: Improving VLA Models in Practice

👍 11

07/07 08:00

Despite recent progress of VLA foundation models, the disparity between laboratory conditions and real-world applications continues to impede their practical implementation. To bridge this gap, we present LingBot-VLA 2.0, which advances LingBot-VLA through improvements in three functional domains. (

中文介绍 提出LingBot-VLA 2.0，在三个功能域（数据、泛化、部署）改进，缩小VLA基础模型从实验室到实际应用的差距，提升实用性能。

TurnOPD: Making On-Policy Distillation Turn-Aware for Efficient Long-Horizon Agent Training

👍 10

07/07 08:00

On-policy distillation (OPD) trains a student policy by matching a stronger teacher on the student's own trajectories, offering a promising framework for language agent training. However, its application to long-horizon agentic tasks remains insufficiently explored. We identify two key inefficiencie

中文介绍 提出TurnOPD，使在线策略蒸馏感知回合边界，解决长周期智能体训练中采样效率低下和早期回合无效匹配两个关键低效问题。

MentalThink: Shaping Thoughts in Mental SVG World

👍 10

07/03 08:00

We introduce MentalThink, a visual-symbolic reasoning paradigm that equips Multimodal LLMs (MLLMs) with an executable mechanism for "mental" visualization. The core of MentalThink is a think-with-SVG pipeline, where the model learns to generate, render, and interpret scalable vector graphics (SVG) c

中文介绍 提出MentalThink视觉符号推理范式，通过「think-with-SVG」管道（生成、渲染、解释SVG），赋予多模态LLM可执行的思维可视化能力。

Unified Audio Intelligence Without Regressing on Text Intelligence

👍 10

07/06 08:00

Audio intelligence involves understanding, reasoning about, and generating both audio and speech. In this work, we introduce Nemotron-Labs-Audex-30B-A3B (Audex), a unified audio-text LLM built on Nemotron-Cascade-2-30B-A3B, a strong text-only MoE LLM. Audex adopts a simple unified design with a sing

中文介绍 提出Audex，基于MoE文本LLM构建的统一音频-文本模型，在不退化文本智能的前提下实现音频理解、推理与生成。

CanvasAgent: Enabling Complex Image Creation and Editing via Visual Tool Orchestration

👍 8

07/06 08:00

Complex image creation and editing often require more than a single generation or editing model. A user request may involve synthesizing images, localizing objects, segmenting regions, editing selected content, compositing intermediate assets, reading text, and enhancing the final result. Such tasks

中文介绍 提出CanvasAgent，通过视觉工具编排（合成、定位、分割、编辑、组合等）实现复杂图像创建与编辑，应对多步骤用户需求。

PointDiT: Pixel-Space Diffusion for Monocular Geometry Estimation

👍 7

07/02 08:00

State-of-the-art single-image 3D reconstruction methods often rely on complex hybrid architectures and loss functions, or compress geometry into latent spaces in order to leverage pre-trained latent diffusion models. In this work, we show that such architectural overhead and intricate loss formulati

中文介绍 提出PointDiT，直接在像素空间进行扩散，无需复杂混合架构和潜在空间压缩，简化单目几何估计流程并取得SOTA性能。

CGGS: Consistency-Augmented Geometric Gaussian Splatting for Ego-centric 3D Scene Generation

👍 7

07/04 08:00

Challenges remain in ego-centric 3D scene generation due to limited view overlap and the dominant influence of individual perspectives on scene interpretation. These factors hinder the creation of viewpoint-consistent and semantically aligned visual content, as well as the construction of accurate g

中文介绍 提出CGGS，通过一致性增强的几何高斯散射，解决自中心3D场景生成中视角重叠有限和个体视角偏差问题，提升视图一致性与语义对齐。

SeKV: Resolution-Adaptive KV Cache with Hierarchical Semantic Memory for Long-Context LLM Inference

👍 7

06/30 08:00

Large language models increasingly operate over long contexts, where the KV cache becomes a dominant memory bottleneck: its size grows linearly with sequence length and must be retained throughout decoding, making full GPU caching prohibitively expensive without compression. Existing KV cache compre

中文介绍 提出SeKV，分辨率自适应KV缓存与层级语义记忆，根据上下文动态调整缓存精度，显著压缩长序列LLM推理中的显存占用。

TREK: Distill to Explore, Reinforce to Refine

👍 6

07/06 08:00

Group Relative Policy Optimization (GRPO) is effective when the current policy already samples useful reasoning trajectories, but it stalls on hard prompts whose correct solution modes lie outside the student's on-policy support. We propose TREK (Teacher-Routed Exploration via Forward KL), a simple

中文介绍 提出TREK，通过教师路由的前向KL探索困难问题，再以强化学习精炼，解决GRPO在硬提示上因策略外解空间缺失而停滞的问题。

Nemotron-Labs-Diffusion: A Tri-Mode Language Model Unifying Autoregressive, Diffusion, and Self-Speculation Decoding

👍 6

07/07 08:00

We introduce Nemotron-Labs-Diffusion, a tri-mode language model (LM) that unifies AR, diffusion, and self-speculation decoding within a single architecture. Trained with a joint AR-diffusion objective, Nemotron-Labs-Diffusion can switch modes to sustain high throughput across deployment settings and

中文介绍 提出Nemotron-Labs-Diffusion，三模态语言模型统一自回归、扩散和自推测解码，通过联合训练灵活切换模式，在不同部署场景下保持高吞吐。

The Task Economy - Data will be the next $1 Trillion Category

@EverettRandle · 19.4K 粉丝 · 484.8K 阅 · 507 赞 · 56 转

07/08 00:15

The Token Economy When we talk about AI today, tokens are king. Specifically, inference tokens have emerged as the primary proxy for tracking the growth of the AI ecosystem. Public companies report

中文介绍 分析AI经济中推理token成为核心度量，提出「任务经济」概念，预言数据将是下一个万亿美元类别，超越token本身。

Continual Learning for Agents

@pirroh · 15.0K 粉丝 · 314.8K 阅 · 540 赞 · 59 转

07/06 21:10

Everyone talks about Continual Learning as if it means one thing only: updating model weights. But there's an inconvenient truth about the agent ecosystem — the vast majority of agents in production

中文介绍 指出持续学习在智能体领域被窄化为权重更新，但生产环境中的大多数智能体实际并非如此，揭示这一尴尬现实。

Everyone is wrong about open source AI in the enterprise

@thejessezhang · 59.4K 粉丝 · 274.2K 阅 · 512 赞 · 54 转

07/06 23:31

The prevailing story right now is that open source is eating the enterprise. The capability gap between the best closed and open models has shrunk to low single digits. A third of the Fortune 500 has

中文介绍 反驳「开源吞噬企业」的主流观点，称闭源与开源模型能力差距已缩至个位，但财富500强实际采用情况并非如此。

Getting started with loops

@ClaudeDevs · 564.6K 粉丝 · 252.4K 阅 · 2.9K 赞 · 239 转

07/07 03:08

There’s a lot of talk right now about "designing loops" instead of prompting your coding agent. If you spend some time on X trying to pin down what a loop actually is, you'll come across multiple

中文介绍 介绍编码智能体中「设计循环」替代直接提示的理念，解析循环的实际含义与实现方法，以提升智能体性能。

This prompt will change your life

@EXM7777 · 123.5K 粉丝 · 233.6K 阅 · 554 赞 · 38 转

07/07 22:31

for most people, fable has a few hours left to live... this is the smartest model that ever existed and everyone's spending its last night building one more app i made it do something 100x more

中文介绍 在Fable模型即将退出免费计划之际，分享一个改变人生的提示词：利用最后几小时提取模型智能，做比建应用更有价值的事。

How to become an applied AI engineer

@eyad_khrais · 20.2K 粉丝 · 225.8K 阅 · 508 赞 · 38 转

07/07 23:42

I wrote this as the guide I would have wanted before making the transition into Applied AI Engineering. The role broadly overlaps with traditional software engineering, but it adds a few important

中文介绍 提供从传统软件工程转型为应用AI工程师的完整指南，强调角色差异与关键技能路径，适合想要入行的人。

The Heretic’s Guide to AI’s Stars Part III: AI Demand, Offshore Financing, & Compression Too

@michaeljburry · 2.0M 粉丝 · 209.2K 阅 · 508 赞 · 51 转

07/08 08:23

“They are just flying empty airplanes around.” - Son NVIDIA is the North Star, Orion, the whole Milky Way. NVIDIA is benefitting from strong demand, but is selling into a concentrated set of buyers

中文介绍 以NVIDIA为例分析AI需求、离岸融资与压缩，指出买家高度集中，可能存在「空飞飞机」的泡沫风险。

Expanding Managed Agents in Gemini API: background tasks, remote MCP and more

@GoogleAIStudio · 184.1K 粉丝 · 165.4K 阅 · 504 赞 · 62 转

07/08 00:38

Today we’re announcing new capabilities in Managed Agents in Gemini API, including background execution, remote MCP server integration, custom function calling and refreshing credentials across

中文介绍 Google AI Studio宣布Gemini API中Managed Agents新功能：后台执行、远程MCP集成、自定义函数调用和跨会话凭证刷新。

You have 1 day to clone Fable 5 into Opus 4.8.

@alex_prompter · 277.4K 粉丝 · 125.9K 阅 · 595 赞 · 77 转

07/07 02:25

Tomorrow is the last day Fable 5 sits inside your plan for free. After July 7 it moves to pay-per-use credits, and most people are about to spend the week arguing over whether it's worth keeping. That

中文介绍 提醒Fable 5即将转为按量付费，建议在免费最后一天将其克隆到Opus 4.8，并提供迁移策略避免争论。

Model and effort in Claude Code: knowing more vs. trying harder

@ClaudeDevs · 564.6K 粉丝 · 95.9K 阅 · 1.5K 赞 · 100 转

07/09 00:55

Claude Code gives you two settings that both seem to "make the answer better": the model, and the effort level. But what do these actually do to the output? And how do you know whether to reach for a

中文介绍 解析Claude Code中模型选择与努力水平的差异：前者增加知识，后者增加尝试次数，指导用户何时该换模型或提高努力度。

Do this on your last day with Fable

@EXM7777 · 123.5K 粉丝 · 86.9K 阅 · 510 赞 · 37 转

07/06 23:48

You can extract Fable 5's intelligence out of the model before it's gone... and i'm going to teach you the 5 workflows that do it, with every prompt ready to paste because tomorrow Fable 5 leaves your

中文介绍 在Fable 5付费前最后一天，分享5个工作流及可直接粘贴的提示词，用于提取模型智能，避免能力消失。

You can go from zero to hireable AI engineer in 4 months. Here's the exact path.

@free_ai_guides · 11.1K 粉丝 · 59.1K 阅 · 503 赞 · 114 转

07/08 00:32

AI engineering is one of the best-paid, fastest-growing jobs in tech right now. And the door to it just got wider than it's ever been. Most guides get this wrong. They hand you a wall of theory, tell

中文介绍 宣称4个月从零到可雇佣AI工程师，提供精确路径，批评其他指南过于理论化，贴合当前高薪趋势。

How to Distill a $10M+ AI Researcher

@JIACHENLIU8 · 2.9K 粉丝 · 48.9K 阅 · 508 赞 · 28 转

07/08 02:16

tldr: Research Process > Research Outcomes Everyone says the best researchers have great intuition. They seem to know which questions are worth asking before the field recognizes them as important.

中文介绍 强调研究过程重于研究结果，顶尖研究者的直觉来自问题选择，提供从高产AI研究者身上提炼方法的技巧。

Unicity Labs and Quant Partner to Bring Verifiable Execution to Agentic Trading

@unicity_labs · 127.3K 粉丝 · 4.5K 阅 · 520 赞 · 827 转

07/08 22:22

Today we're announcing a partnership between Unicity Labs and @tryquantio to build the secure foundation for agentic finance. Quant - the AI intelligence and execution layer for the agentic future

中文介绍 Unicity Labs与Quant合作，为智能体交易提供可验证的执行层，构建智能体金融的安全基础。

A Google Knowledge Panel Isn't a Vanity Badge. Here's What It Actually Is - and How I Earned Mine.

@krudhrah · 1.2K 粉丝 · 1.3K 阅 · 517 赞

07/08 00:11

Search my name — "Rudhrah Keshav" — and Google doesn't just show you links. It shows you a panel: my photo, my profession, my date of birth, my education, my podcast, my profiles. All assembled by

中文介绍 解释Google知识面板是结构化信息展示而非虚荣徽章，分享作者获得个人面板的经验与自动聚合条件。

Why AI Infrastructure must evolve for Agent Experience — Akshat Bubna, Modal CTO

Latent Space · 07/09 06:55

2 years after our first coverage, we return with Modal's other cofounder to explore why Agent Experience is working now, and everything they have learned building the new agent cloud.

中文介绍 Modal联合创始人兼CTO Akshat Bubna探讨AI基础设施为何必须为智能体体验演进，分享构建新智能体云的经验。

Data for Agents

Hugging Face Blog · 07/09 01:16

中文介绍 NVIDIA在Hugging Face博客发文，探讨为AI智能体提供开放数据的关键问题。

Our approach to government and national security partnerships

OpenAI News · 07/08 21:30

Learn how OpenAI approaches government and national security partnerships, with principles for responsible AI use, democratic accountability, and public safety.

中文介绍 OpenAI阐述其政府及国家安全合作方式，强调负责任AI使用、民主问责和公共安全原则。

Separating signal from noise in coding evaluations

OpenAI News · 07/08 21:00

A new analysis from OpenAI reveals issues in SWE-Bench Pro, a popular coding benchmark, raising concerns about reliability and accuracy in evaluating AI models.

中文介绍 OpenAI分析指出热门编码基准SWE-Bench Pro存在问题，质疑其在评估AI模型时的可靠性和准确性。

Helping K–12 educators build practical AI skills

OpenAI News · 07/08 18:00

OpenAI Academy and the Walton Family Foundation are bringing hands-on AI Skills Jams to help K–12 educators build practical AI skills for the classroom.

中文介绍 OpenAI Academy与沃尔顿家族基金会合作，通过实操AI技能工作坊帮助K-12教师培养课堂实践AI技能。

[AINews] Lilian Weng summarizes 35 papers on Harness Engineering for RSI

Latent Space · 07/08 10:20

a quiet day lets us read some condensed insight

中文介绍 Lilian Weng总结35篇关于Harness Engineering for RSI的论文，提供精炼见解。

Introducing GPT-Live

OpenAI News · 07/08 08:00

A new generation of voice models for natural human-AI interaction, now powering ChatGPT Voice.

中文介绍 OpenAI推出GPT-Live，新一代语音模型，支持自然人机交互，现为ChatGPT语音功能提供动力。

Native-speed vLLM transformers modeling backend

Hugging Face Blog · 07/08 08:00

中文介绍 Hugging Face博客介绍vLLM的原生速度transformers建模后端。

GPT-5.6 Thursday ⭐️, Claude Cowork mobile 📱, Gemini API agents 🤖

TLDR AI · 07/08 08:00

中文介绍 TLDR AI报道GPT-5.6发布、Claude Cowork移动应用及Gemini API智能体。

From Hugging Face to Amazon SageMaker Studio in one click

Hugging Face Blog · 07/08 05:15

中文介绍 Hugging Face推出从平台一键部署到Amazon SageMaker Studio的功能。

Hugging Face Models on Foundry Managed Compute

Hugging Face Blog · 07/07 23:20

中文介绍 微软Foundry托管计算现支持Hugging Face模型。

The foundational elements of AI architecture that IT leaders need to scale

MIT Tech Review AI · 07/07 19:10

With the rapid progress of AI capabilities and the move to agentic systems, organizations are expanding their use cases as the technology continues to grow. That constant evolution also introduces risk, leaving IT leaders to wonder which investments will prove valuable even six months into the futur

中文介绍 MIT Tech Review文章指出，随着AI能力快速进步和智能体系统兴起，IT领导者需关注AI架构基础要素，谨慎选择投资方向。

[AINews] The Field Guide to Fable

Latent Space · 07/07 12:44

a quiet day lets us digest the world's most significant model launch... to date.

中文介绍 Latent Space发布关于Fable模型发布的现场指南，称其为迄今最重要模型发布之一。

Australian Payments Plus moves faster with ChatGPT and Codex

OpenAI News · 07/07 08:00

See how Australian Payments Plus uses ChatGPT Enterprise and Codex to move faster through payments complexity. AP+ saves time, improves quality, and keeps human judgment central.

中文介绍 澳大利亚支付平台Australian Payments Plus借助ChatGPT Enterprise和Codex加速处理支付复杂性，节省时间、提升质量，并保持人为判断。

MUFG aims to become AI-native with OpenAI

OpenAI News · 07/07 08:00

MUFG uses ChatGPT Enterprise to build an AI-native organization, improve workflows, and deliver new AI-powered financial services at scale.

中文介绍 三菱日联金融集团（MUFG）与OpenAI合作，使用ChatGPT Enterprise构建AI原生组织，改善工作流程并规模化提供AI驱动金融服务。