2026-07-14

Accurate, Interdisciplinary and Transparent Structure-property Understanding with Deep Native Structural Reasoning

👍 85

07/08 08:00

Structure-property relationships are foundational to biology, chemistry and materials science, where function, reactivity and physical response emerge from spatial, chemical and periodic organization. Mechanistically explaining these relationships requires interpreting structural evidence through sc

中文介绍 提出深度原生结构推理（Deep Native Structural Reasoning）方法，从空间、化学、周期性等结构化证据中解释功能与反应性，提升跨学科结构-性质理解的准确性与透明度。

Video-Oasis: Rethinking Evaluation of Video Understanding

👍 58

07/02 08:00

The inherent complexity of video understanding makes it difficult to determine whether Video-LLM benchmark performance stems from visual perception, linguistic reasoning, or knowledge priors. While many benchmarks have emerged to assess high-level reasoning, shared criteria for evaluating video unde

中文介绍 提出Video-Oasis基准，重新审视视频理解评估，区分视觉感知、语言推理与知识先验的贡献，建立共享评价标准以统一视频理解评估。

Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading

👍 47

07/09 08:00

AI agents have become capable of autonomously completing short, well-specified tasks. However, existing terminal benchmarks largely focus on simple problems that finish within minutes and are evaluated only by their final outcome. This setup overlooks intermediate progress and partial solutions, yie

中文介绍 提出Long-Horizon-Terminal-Bench，针对长时域终端任务采用密集奖励评分衡量中间进展，弥补现有基准仅评估最终结果的不足。

Scalable Visual Pretraining for Language Intelligence

👍 41

07/10 08:00

The rapid progress of large foundation models has been driven predominantly by pretraining on large-scale text corpora. However, many forms of knowledge are conveyed through visual representations, where figures, typeset equations, and page layouts carry rich information that cannot be faithfully or

中文介绍 针对文本预训练难以捕获图表、公式等视觉知识的问题，提出可扩展视觉预训练方法，利用视觉表示增强语言智能，提升知识获取的完整性。

Video Generation Models are General-Purpose Vision Learners

👍 40

07/10 08:00

Driven by next-token prediction, NLP shifted from task-specific models into powerful generalist foundation models. What, then, is the equivalent catalyst needed to achieve a general-purpose model in computer vision? In this paper, we contend that large-scale text-to-video generation serves as a stro

中文介绍 论证大规模文本到视频生成可作为计算机视觉通用模型的催化剂，使视频生成模型成为通用视觉学习者，迁移至多种下游任务。

Ideas Have Genomes: Benchmarking Scientific Lineage Reasoning and Lineage-Grounded Idea Generation

👍 32

07/09 08:00

Scientific ideas rarely start from a blank page. They inherit mechanisms, repair known limitations, and recombine pieces of earlier work, much like biological genomes. Current benchmarks still say little about whether AI systems can follow this inheritance structure. We present IdeaGene-Bench (IG-Be

中文介绍 提出IdeaGene-Bench，评估AI对科学思想继承与重组结构的推理能力（谱系推理），以及基于谱系生成新思想的能力。

UniClawBench: A Universal Benchmark for Proactive Agents on Real-World Tasks

👍 31

07/09 08:00

The rapid development of large language models and multimodal large language models has accelerated the emergence of proactive agents capable of operating everyday tools and assisting users in real-world environments. However, existing benchmarks struggle to evaluate such agents effectively, as they

中文介绍 提出UniClawBench通用基准，评估主动agent在真实世界任务中操作工具和协助用户的能力，弥补现有评估在动态环境中的不足。

LongE2V: Long-Horizon Event-based Video Reconstruction, Prediction, and Frame Interpolation with Video Diffusion Models

👍 29

07/09 08:00

Recovering high-quality video from sparse event streams is a challenging task. Regression methods often blur textures, while existing generative models struggle with long-term stability. We propose LongE2V, a novel approach that leverages pre-trained video diffusion priors to jointly handle event-ba

中文介绍 提出LongE2V，利用预训练视频扩散先验，从稀疏事件流进行长时域视频重建、预测和帧插值，解决纹理模糊与长期不稳定问题。

Enhancing In-context Panoramic Generation via Geometric-aware Pretraining

👍 20

07/09 08:00

In this work, we present Canvas360, a two-stage framework for in-context panoramic generation that combines geometry-aware pretraining with downstream task-specific fine-tuning. To address the lack of large-scale, high-quality training data tailored to in-context panoramic tasks, we propose Canvas36

中文介绍 提出Canvas360两阶段框架，通过几何感知预训练与任务微调实现上下文全景生成，并构建大规模高质量数据集Canvas360-1.4M。

Trust Region Policy Distillation

👍 17

07/06 08:00

Big goals are hard to achieve all at once; breaking them into small steps is wiser. We present Trust Region Policy Distillation (TOP-D), which transforms the notoriously unstable, high-variance On-Policy Distillation (OPD) into a stable training paradigm by dynamically constructing a proximal teache

中文介绍 提出Trust Region Policy Distillation（TOP-D），通过动态构建近端教师将高方差的在线策略蒸馏转化为稳定训练范式，实现逐步知识迁移。

OpenCoF: Learning to Reason Through Video Generation

👍 17

07/09 08:00

Reasoning has become a core capability for large models, especially when reliable decisions require understanding logical consequences. Recent video generation models offer a reasoning path distinct from previous Chain-of-Thought (CoT): reasoning can unfold through temporally connected frames, known

中文介绍 提出OpenCoF，利用视频生成模型通过时间连接帧进行推理，区别于链式思维，开辟基于视觉时序的新推理路径。

KronQ: LLM Quantization via Kronecker-Factored Hessian

👍 16

07/08 08:00

Post-training quantization (PTQ) is a widely adopted technique for compressing large language models (LLMs) without retraining. Existing second-order PTQ methods, including GPTQ, construct quantization objectives exclusively from input activation statistics, effectively assuming that all output chan

中文介绍 提出KronQ，利用Kronecker分解Hessian矩阵进行LLM后训练量化，考虑输出通道重要性，弥补现有方法仅依赖输入统计的不足。

Jet-Long: Efficient Long-Context Extension with Dynamic Bifocal RoPE

👍 16

07/08 08:00

Modern LLMs are increasingly deployed in long-context applications such as retrieval-augmented generation, repository-level coding, and agentic workflows whose accumulated reasoning and tool traces routinely push the input an order of magnitude past the pretraining window, making zero-shot context e

中文介绍 提出Jet-Long，采用动态双焦RoPE实现高效长上下文扩展，支持RAG、agent工作流等超出预训练窗口的零样本应用。

Single-Rollout Asynchronous Optimization for Agentic Reinforcement Learning

👍 16

07/08 08:00

Reinforcement learning (RL) is becoming increasingly important for post-training large language models (LLMs). Previous RL pipelines for LLMs were mostly synchronous and batch-interleaved, which is inefficient for long-horizon agentic tasks. Recently, asynchronous RL has emerged as a more efficient

中文介绍 提出单轨迹异步优化（Single-Rollout Asynchronous Optimization），解决同步RL在长时域agent任务中的低效问题，提升LLM后训练效率。

CineMobile: On-Device Image-to-Video Diffusion for Cinematic Camera Motion Generation

👍 14

07/04 08:00

The growing demand for image-to-video creation on mobile devices has increasingly focused on cinematic motion effects like bullet time, dolly zoom, slow motion, etc. While Diffusion Transformers (DiTs) exhibit strong performance in video generation, their large parameter sizes and multi-step iterati

中文介绍 提出CineMobile，面向移动设备的图像到视频扩散模型，高效生成子弹时间、镜头推拉等电影级相机运动，解决DiT参数大、迭代多的问题。

Automating the Design of Embodied Agent Architectures

👍 12

07/03 08:00

Embodied agents are typically built as hand-designed compositions of perception, memory, planning, and action modules. This modularity exposes a large architectural design space, but current systems still rely on researcher intuition to choose where information is stored, how observations are proces

中文介绍 提出自动化设计具身agent架构的方法，自动搜索感知、记忆、规划、行动模块的最佳组合，减少手工设计依赖。

Remember When It Matters: Proactive Memory Agent for Long-Horizon Agents

👍 10

07/09 08:00

In long-horizon tasks, decision-relevant state is often scattered across an expanding trajectory, while the action agent must surface it and act. As trajectories grow, task requirements, environment facts, prior attempts, diagnoses, and open subgoals can be buried in the context window or pushed bey

中文介绍 提出Proactive Memory Agent，在长时域任务中主动记忆和检索关键状态（任务需求、环境事实、先前尝试等），防止信息被上下文窗口淹没。

Sparse Delta Memory: Scaling the State of Linear RNNs through Sparsity

👍 10

07/08 08:00

Linear attention models allow a fixed state size and a fixed amount of compute per token. However, due to their limited state size, linear attention models fall behind in long-context recall compared to softmax-attention-based transformer architectures. Increasing the state size of linear attention

中文介绍 提出Sparse Delta Memory，利用稀疏性扩展线性RNN的状态大小，提升长上下文召回能力，弥补与softmax Transformer的差距。

AgentLens: Production-Assessed Trajectory Reviews for Coding Agent Evaluation

👍 8

07/07 08:00

We present AgentLens, a production-assessed benchmark for interactive code agents. Most code-agent benchmarks reduce a run to a single bit -- did the task pass? -- but the people who actually use these agents experience the entire trajectory: how the agent follows instructions, uses its tools, verif

中文介绍 提出AgentLens，基于生产环境评估的代码agent轨迹审查基准，不仅判断任务通过与否，还评估指令遵循、工具使用、验证等完整轨迹质量。

Towards Mechanistically Understanding Why Memorized Knowledge Fails to Generalize in Large Language Model Finetuning

👍 7

07/09 20:17

Fine-tuning LLMs to inject new knowledge faces a critical challenge: LLMs can quickly memorize new facts, yet fail to use them for downstream reasoning tasks. We formalize this failure as the \textbf{Knowing--Using Gap}, characterized by an accuracy gap and a temporal lag between memorization and ge

中文介绍 定义微调LLM中「Knowing-Using Gap」，即新知识被记忆但无法泛化推理，通过机制理解揭示准确率差距与时间延迟，为弥合该鸿沟提供方向。

The Great Wave Has Arrived (from GLM CEO Jie Tang)

@bingxu_ · 6.5K 粉丝 · 318.6K 阅 · 505 赞 · 88 转

07/11 23:10

-- Bing Xu's Note --- I came across an internal GLM letter on the Chinese app RedNote, purportedly written by @jietang, and translated the Chinese text in the images into English. --- Letter Start

中文介绍 Bing Xu 翻译了 GLM 创始人唐杰在红书发布的内部公开信，信中宣告 AI 大浪潮已至，呼吁团队做好准备。原文为中文，经翻译后在英文社区传播，引发对 GLM 战略和国产大模型发展的关注。

The Reverse Information Paradox

@satyanadella · 7.0M 粉丝 · 146.5K 阅 · 1.1K 赞 · 162 转

07/12 23:09

In the age of intelligence, how should firms protect their core IP? Nobel Prize winning economist Kenneth Arrow famously described a paradox in the market for information. “Its value for the

中文介绍 微软 CEO 萨提亚·纳德拉引用阿罗信息悖论，讨论智能时代企业如何保护核心知识产权。他提出「反向信息悖论」——信息越有价值，越难通过市场交易定价和获利，建议企业重新思考 IP 策略。

gpt-5.6-sol without hitting limits

@theo · 356.6K 粉丝 · 77.0K 阅 · 774 赞 · 27 转

07/12 06:59

I've burned over $200,000 of tokens with gpt-5.6-sol. It's a great model. That said, it's a bit too easy to hit your limits on the $200 Codex Pro sub. OpenAI has been generous about resets, but that

中文介绍 用户 theo 在 gpt-5.6-sol 上花费超 20 万美元 token，认为模型性能出色，但 200 美元/月的 Codex Pro 订阅极易撞上限，虽 OpenAI 常给重置，仍建议优化配额机制以避免体验断档。

How I'd make $10 million with AI agents

@gregisenberg · 682.7K 粉丝 · 62.2K 阅 · 578 赞 · 36 转

07/14 02:22

2 markets worth trillions are opening at the same time. I've been through 3 platform shifts living in Silicon Valley building/selling VC-backed companies and I've NEVER seen this. Mobile apps came

中文介绍 连续创业者 Greg Isenberg 认为两个万亿级市场（AI Agent + 新平台）同时开启，他结合三次平台转型经验，提出用 AI Agent 赚取 1000 万美元的具体路径，强调抓住移动应用之后的最大机遇。

Motion-Graphics Skill

@HeyGen · 106.3K 粉丝 · 58.5K 阅 · 516 赞 · 56 转

07/13 03:00

Day 7 of 30: Motion Graphics by Example /motion-graphics makes short, design-led motion graphics: usually under 10 seconds, no narration, motion is the message. Kinetic type, stat count-ups, charts,

中文介绍 HeyGen 推出「Motion Graphics」新技能，可快速生成 10 秒以内、无旁白的设计导向动态图形，如动态文字、数据计数、图表动画等。以运动为主体传达信息，适用于短视频和品牌内容创作。

GPT‑5.6 by our community

@OpenAIDevs · 373.8K 粉丝 · 58.5K 阅 · 532 赞 · 34 转

07/12 03:14

GPT‑5.6 has only been in builders’ hands for a couple of days. Here’s what the community is already building with it: A voxel Manhattan, built autonomously with GPT‑5.6 Sol: GPT‑5.6 Sol got to

中文介绍 OpenAI Devs 展示社区用 GPT-5.6 的早期成果，包括自主构建体素曼哈顿城市（Voxel Manhattan）。模型在代码生成、空间理解和自主编程方面表现突出，开发者正探索更多复杂创意应用。

How To Actually Design With AI

@LexnLin · 17.6K 粉丝 · 32.3K 阅 · 503 赞 · 27 转

07/13 05:44

Design has always been one of the areas where AI struggles to match human ability. But most people still misunderstand what designing with AI actually means. Designing with AI does not mean asking it

中文介绍 Lexn Lin 指出多数人对「用 AI 设计」存有误解——并非让 AI 直接出图。真正的 AI 设计是将其作为协作工具，人类把控方向与美感，AI 负责执行与迭代。他分享了提升人机协作效率的方法。

What Anthropic’s latest AI discovery does—and doesn’t—show

MIT Tech Review AI · 07/14 02:00

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. Anthropic—currently the world’s most valuable AI company, with a nearly $1 trillion valuation—has a reputation for publishing strange and heady research. It’s loo

中文介绍 Anthropic（估值近1万亿美元的全球最有价值AI公司）的最新AI发现展示的内容及其局限性。

Claude Code browser 🌍, Cursor general agent 🤖, Claude Fable extension ⏳

TLDR AI · 07/13 08:00

中文介绍 当日AI动态包括Claude Code浏览器发布、Cursor通用代理及Claude Fable扩展。

not much happened today

Smol AI News · 07/11 13:44

**Prime Intellect** released **verifiers v1**, a redesigned environment stack for **agentic reinforcement learning** and evaluations, improving efficiency by storing rollout traces as **message DAGs** to reduce complexity from **O(n²)** to **O(n)**. This enables practical long-horizon multimodal rol

中文介绍 Prime Intellect发布verifiers v1，重构智能体强化学习环境栈，通过消息DAG存储rollout轨迹，将复杂度从O(n²)降至O(n)，支持实际长程推理。

[AINews] not much happened today

Latent Space · 07/11 10:53

a quiet day after a week of nonstop model releases

中文介绍 当天AI领域较为平静，此前一周有连续不断的模型发布。

How Deutsche Telekom is rewiring telecommunications with AI

OpenAI News · 07/10 15:00

How Deutsche Telekom is becoming an AI-native telco with OpenAI-transforming customer service, employee workflows, network operations, and the future of voice.

中文介绍 德国电信与OpenAI合作，利用AI改造客服、员工工作流、网络运营及语音服务，向AI原生电信公司转型。

[AINews] OpenAI launches GPT 5.6 Sol/Terra/Luna, Codex becomes ChatGPT superapp

Latent Space · 07/10 14:19

A big day for OpenAI.

中文介绍 OpenAI推出GPT-5.6版本（Sol/Terra/Luna分层），同时Codex成为ChatGPT超级应用。

not much happened today

Smol AI News · 07/10 13:44

**OpenAI** rolled out **GPT-5.6** featuring a new model stratification with tiers **Luna / Terra / Sol** and effort levels including **Max** and **Ultra**, introducing complex configuration options. The launch faced UX challenges with the **ChatGPT Work / Codex** split, prompting rapid corrective ac

中文介绍 OpenAI推出GPT-5.6，模型分层为Luna、Terra、Sol，并引入Max和Ultra努力级别。但发布面临ChatGPT Work与Codex分拆带来的用户体验挑战。

Getting started with ChatGPT

OpenAI News · 07/10 08:00

Learn how to use ChatGPT, start your first conversation, and discover simple ways to write, brainstorm, and solve problems with AI.

中文介绍 OpenAI学院发布ChatGPT入门指南，教授如何开始对话及使用AI进行写作、头脑风暴和问题解决。

Profiling in PyTorch (Part 3): Attention is all you profile

Hugging Face Blog · 07/10 08:00

中文介绍 Hugging Face博客发布PyTorch性能分析系列第三篇，主题为注意力机制剖析。

GPT-5.6 🚀, Muse Spark 1.1 ✨, ChatGPT Work 💼

TLDR AI · 07/10 08:00

中文介绍 当日AI动态包括GPT-5.6发布、Muse Spark 1.1更新、ChatGPT Work推出。

Anthropic found a hidden space where Claude puzzles over concepts

MIT Tech Review AI · 07/10 04:22

The AI firm Anthropic has developed a technique that has given it the clearest glimpse yet at what’s really going on inside large language models as they answer questions or carry out tasks. What they found ranges from the mundane to the unnerving. Researchers at the company built a tool called the

中文介绍 Anthropic研究发现大语言模型内部存在隐藏空间，Claude在其中对概念进行推理，该技术提供了模型内部运作的最清晰视图，结果有平凡也有令人不安的。

GPT-5.6 is now the preferred model in Microsoft 365 Copilot

OpenAI News · 07/09 21:00

Learn how GPT-5.6 powers Microsoft 365 Copilot with stronger AI capabilities across Word, Excel, PowerPoint, Chat, and Cowork for faster, higher-quality work.

中文介绍 GPT-5.6成为Microsoft 365 Copilot首选模型，增强Word、Excel、PowerPoint等应用的AI能力，提升工作效率和质量。

GPT-5.5 Bio Bug Bounty

OpenAI News · 07/09 18:00

Details about the OpenAI Bio Bounty program

中文介绍 OpenAI推出GPT-5.5生物漏洞赏金计划，旨在发现与生物安全相关的潜在风险。

GPT-5.6: Frontier intelligence that scales with your ambition

OpenAI News · 07/09 18:00

More intelligence from every token, stronger performance per dollar, and more capability on demand for your hardest work.

中文介绍 OpenAI发布GPT-5.6，宣称每个token带来更智能输出，单位成本性能更强，可按需扩展能力以应对复杂任务。

ChatGPT is now a partner for your most ambitious work

OpenAI News · 07/09 18:00

ChatGPT Work is an agent that can take action across your apps and files, stay with a project for hours if needed, and turn a goal into finished work.

中文介绍 OpenAI推出ChatGPT Work智能体，可跨应用和文件执行操作，长期持续处理项目，将目标转化为完成的工作。