🔥 AI热点
实时追踪 AI 行业热点动态 · 共 100 条资讯
论文：Coding Agents with an Obstacle-Aware Harness for Safe Robot Manipulation
Coding agents have emerged as a promising paradigm for robot manipulation: a language model writes the robot controller as a program, and agents built in this way now operate robots without robot-specific training.Whether this paradigm is also safe, however, has not been asked. We evaluate coding agent under a safety constraint, where each task pairs a manipulation goal with an obstacle the robot
0 阅读原文论文：Region-Level Policy Optimization for Fine-grained MLLM Perception
Region-Level Policy Optimization for Fine-grained MLLM Perception
0 阅读原文论文：Embedding Models Measure in Peculiar Ways
Embedding spaces define notions of semantic similarity and distance. We study whether those embeddings reflect physical measurements of mass, distance, time and volume, which admit a unique, objective notion of semantic equivalence and distance. We find that physical measurement is only weakly modeled in the embedding space, and that instead quite peculiar measurement patterns can be observed. Fur
0 阅读原文论文：SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent Harness
SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent Harness
0 阅读原文论文：Workspace Models: Lightweight Robotic Memory via Saliency-Driven Supervision
Complex robotic manipulation tasks frequently require a long-term memory of past events and actions. As conditioning on full histories renders policies prone to spurious correlations and degrades performance, many approaches to policy memory involve compressing historical information through expensive VLM queries in-the-loop to process only task-salient information. In this paper, we propose an al
0 阅读原文论文：An Empirical Study of Harness Design for Coding Agents
An Empirical Study of Harness Design for Coding Agents
0 阅读原文科大讯飞推出全国产语音基座大模型 Spark-Audio-1.0-Preview
科大讯飞推出全国产语音基座大模型 Spark-Audio-1.0-Preview ，基于纯国产算力训练，采用0.65B音频编码器搭配30B-A3B MoE语言模型，使用1300万小时音频数据训练。模型能直接理解语音的语义、情绪与场景，支持99种语言、202种方言识别，在高噪声、小音量等复杂场景表现领先，Fleurs中文集取得SOTA。 来源：科大讯飞
0 阅读原文论文：FAMOS: Feed-Forward 3D Articulation Modeling from Sparse Observations
Modeling articulated objects from sparse monocular views is challenging because each observation reveals only partial geometry and motion evidence. Most feed-forward methods infer articulation from a single observation and therefore rely heavily on learned category-level shape priors. We present FAMOS, a feed-forward model that predicts movable-part segmentation and joint parameters from a sparse,
0 阅读原文Trail of Bits 用 Agent 为 Miden zkVM 审计自建 LSP、反编译器和 Lean 形式化证明
Trail of Bits 在审计 Miden zkVM 前，让 Agent 用六个月从零构建了 MASM 的 LSP 服务器、反编译器、静态分析引擎和 Lean VM 执行器模型。这些工具发现了可让恶意 prover 伪造 Falcon 签名盗取资金的高危漏洞，静态分析定位了 400 多处类型验证缺陷，Lean 工作产出 95 个机器验证的正确性证明，还发现两个单元测试未捕获的细微 bug。 🔗 阅读原文 via AIHOT · https://aihot.news/items/cmu6w30lt0dnhrowkh7qwiped
0 阅读原文Product Hunt：AEXGrid
Coordinate AI agents. From anywhere. Discussion | Link
0 阅读原文GitHub 热门：TencentCloud/Octop
A smarter, self-hosted AI assistant — multi-user, multi-agent.
0 阅读原文智谱推出高速推理模型 GLM-5.3-FlashX
智谱正式推出 GLM-5.3-FlashX 模型，最高推理速度达200 tokens/s。模型此前以”Ox Alpha”代号与全球开发者见面，调用量持续攀升。新模型基于10万张国产芯片的推理算力，进一步加大对Infra侧投入与推理优化。 GLM-5.3-Flash 长期保持同尺寸最强智能水平，具备极高性价比，本次提速形成智能、价格、速度的全面竞争力。 来源：智谱
0 阅读原文3人团队用前沿模型以不到3000美元token成本入侵OpenAI员工账户
一个3人团队在7月25日利用两个漏洞接管了OpenAI员工的 ChatGPT/Codex 账户，并可访问 Outlook、Slack、GitHub 等关联服务，他们用向 OpenAI 内部代码库提交 PR 的方式证明了漏洞，全程不到72小时。 🔗 阅读原文 via AIHOT · https://aihot.news/items/cmu6n3z980c4qro0fjx53zz4h
0 阅读原文Product Hunt：WhaleRead
Locally translate TXT, Markdown, and EPUB Discussion | Link
0 阅读原文GitHub 热门：Fission-AI/OpenSpec
Spec-driven development (SDD) for AI coding assistants.
0 阅读原文阿里千问推出新一代原生全模态模型 Qwen3.8-Omni-Flash
阿里千问推出新一代原生全模态模型 Qwen3.8-Omni-Flash ，支持文本、图像、音频、视频输入及 1M 长上下文，30 项评测较上代平均提升超 26%，音频能力整体超越 Gemini 3.8 Flash，API 音频输入价格降幅超 98%。模型可端到端完成视频剪辑、短剧翻译、电影解说、会议纪要执行等长程工作流。 来源：千问大模型
0 阅读原文WSJ：三名研究人员用 Claude Opus 5 借 Discourse 漏洞访问 OpenAI 私有代码
据 WSJ 报道，三名研究人员使用 Claude Opus 5 将 Discourse 的一个漏洞串联成对 OpenAI 私有代码的访问。 researchers 在 Discourse 服务器上获得了包括 OpenAI 员工在内的认证 token，部分 forum token 可用于 ChatGPT 并触达 OpenAI 的 GitHub 服务。 🔗 阅读原文 via AIHOT · https://aihot.news/items/cmu6i7irl000qro0fik1cusru
0 阅读原文Product Hunt：AI Class by Kanary
Knight or Ninja? Your Codex & Claude logs decide Discussion | Link
0 阅读原文GitHub 热门：supermemoryai/supermemory
Memory and context engine + app that is extremely fast, scalable, and can be run fully locally. The Memory API for the AI era.
0 阅读原文Hacktron 复盘利用 libheif 漏洞与 OpenAI SSO 缺陷入侵 OpenAI 论坛并接管员工 ChatGPT 账号
Hacktron 团队披露 2026 年 7 月 25 日 chained libheif 堆缓冲区溢出与 OpenAI SSO 身份缺陷，攻破 community.openai.com 并接管多名员工 ChatGPT/Codex 账号，用员工 Codex 在 OpenAI 内部 monorepo 开出无害 PR 作为证明，全程不到 72 小时。 🔗 阅读原文 via AIHOT · https://aihot.news/items/cmu6idvti04i9ro0fktczvi9g
0 阅读原文