AI 情报文章归档

This is a submission for the Sanity Challenge, Path One: Ship an Agent That Queries Real Content ...

OpenAI CFO Sarah Friar 在 CNBC 上确认，Q3 总收入运行率增长超过 70%，业务收入较季度初的 7 月翻倍。增长由更好的模型、新的 100 美元小企业计划、编码、网络安全和生命科学业务驱动。

This is a submission for the Sanity Challenge, Path One: Ship an Agent That Queries Real Content ...

Ditch brittle OCR regex pipelines and expensive SaaS fees. Build a deterministic multi-modal invoice extractor in n8n with Gemini Flash.

中央网信办印发通知，在全国部署开展为期 2 个月的“清朗 · 整治短视频虚假人设乱象”专项行动，重点整治假扮弱势群体、虚构职业资质冒充医生律师、利用 AI 仿冒公众人物、包装富二代名媛鼓吹奢靡迷信、编造冲突剧情等 5 类问题。治理措施包括从严查处违规账号和 MCN 机构批量炮制行为、强化虚假人设类短视频的内容标注要求、禁止在榜单推荐位置推送相关内容，并对诱骗老年人等行为严格处置。

You ask your AI assistant how to do something. It gives you clean, confident code, with an install...

今晚在 Kona 的收获

3 条赛道。3 位获胜者。去创作吧。🥇

BlueQubit 团队使用 IBM 120 量子比特的 Nighthawk r2 处理器，在 36 个周期、918 个双量子比特门配置下仅用 19 秒完成 100 万次随机量子线路采样，研究人员估算 Frontier 超级计算机复现同等任务约需 110 年，约 1.2 × 10²⁷ 次计算操作。

地平线正式发布 HSD 全场景辅助驾驶系统 V2.1，新增全场景倒车 Beta，为国内量产首发端到端全场景倒车，首批搭载车型 iCAR V27 将于 10 月 8 日起分批推送升级。本次升级还新增闪灯鸣笛警示、自定义车位遥控泊入泊出、实体钥匙遥控泊车，并升级主动安全视觉神经网络，提升横穿、转弯低速跟车和目标 Cut in 等场景避撞能力。

Perplexity 发布自研 Rust 检索与排序引擎 Photon，已承接全部生产流量，p99 检索排序延迟从约 800 ms 降至约 65 ms，服务机器减少约 20%。

三星 SDI 在 DCWA 亚洲数据中心展 2026 上首次公开其圆柱形磷酸铁锂（LFP）电池，基于独有结构技术在提升输出功率和容量的同时强化安全性。同场展品还包括采用底部泄压孔设计的高功率高容量 21700 三元锂电池、面向 AI 数据中心应急电源优化的圆柱形电池，以及用于 ESS 储能的方形磷酸铁锂电池。

宜鼎（Innodisk）发布工业级 DDR5-8000 RDIMM 内存模组，较此前 7200MT/s 型号提速 1/9，提供 16~64GB 容量选择。该模组集成 TVS 二极管与 eFuse 电子保险丝，并采用 30μ" 金手指与抗硫化设计。

市场监管总局宣布，JJF 2385—2026《智能眼镜声学性能测试规范》国家计量技术规范将自 10 月 2 日起正式实施，为智能眼镜声学性能检测划定统一技术标尺。

I've been building CORTEX, an AI agent that runs entirely on my own laptop. It plans a task, runs...

I’ve been experimenting with an AI-Driven Development workflow at Relipa, and one thing became clear:...

智元机器人与网易云音乐达成生态合作，旗下远征A3完成网易云音乐系统级原生集成，成为全球首个原生集成音乐平台的全尺寸人形机器人。自10月15日起，用户可通过远征A3语音点播直接调用网易云音乐正版曲库，无需外接设备或手动开启应用。智元已在横琴长隆飞船乐园部署超300台全系机器人，部分点位开启远征A3音乐交互试点。

智元与长隆在横琴长隆飞船乐园落地全球首个大规模具身智能主题乐园，300 多台人形机器人与四足机器狗在 100 多个点位提供文娱商演、科普研学、导览等七大核心场景服务。

Time series anomaly detection (TSAD) is increasingly deployed in streaming settings, where data arrive sequentially and may exhibit non-stationarity. As a result, several works from the recent literature propose streaming anomaly detection…

前 Google CEO Eric Schmidt 建议大学生用 AI 放大自己的产出：非技术背景者可用这些工具把个人梦想与影响力规模化，技术背景者则可用其发明改变世界的东西。他认为当前入行成本前所未有地低，唯一限制是好奇心和冒险意愿。

OpenClaw 官方当地时间 29 日官宣 OpenClaw Enterprise，这是一个开源中立的敏感环境持久性智能体管理平台，即将发布 1.0 正式版。该平台为智能体引入企业级控制平面，支持多租户、严格的安全边界和标准化的智能体原语，并在智能体生命周期内增强治理和审计能力，同时确保核心原语可替换为第三方解决方案及内部实现。

The scariest security failure isn't the guardrail that's down, or the one that's weak. It's the one that's running, passing every health check, and quietly configured to catch nothing — green by construction. I found one in a benchmark of …

We study additive regression under a potentially non-product random design on $[0,1]^d$, allowing the dimension $d$ to grow with the sample size $n$. We introduce coupled smoothness classes that separately control the regularity of the mar…

豆包个人助理产品定名“小豆”，并计划推出独立 App 版本，名称已于今年暑期确定，目前仍处内部测试阶段，最终对外版本可能调整。该产品可视为豆包将手机端积累的 Personal Agent 能力向独立应用形态延伸，此前豆包手机助手消费者版本已于 2026 年 9 月落地努比亚 NaviX Ultra。

Prologue: Like Jev — A Harness for Cheap Brains Source version of CodeSmith: v0.5.0...

SpatialClaw 重新思考智能体空间推理的动作接口 论文：https://huggingface.co/papers/2606.13673

Large Audio Language Models (LALMs) rely on effective audio encoders for multi-task performance. We introduce UniAE-MoE, a unified audio encoder designed to model cross-domain audio representations and achieve outstanding downstream unders…

这个图补充得很不错

VRAM for LLMs is a bandwidth problem: every token streams the whole model from memory. Bandwidth per tier, the 20x offload cliff, and what fits in 16, 24 or 48 GB.

Wireless split learning (SL) reduces on-device computation by offloading upper layers to a server, yet transmitting high-dimensional intermediate features at each iteration remains a major communication bottleneck. Existing methods select …

A deep classifier is defined not only by the decision it produces, but also by the sequence of transformations through which that decision is formed. Treating this evolution as a dynamical system across layers provides a natural framework …

Trustworthy Legal AI requires systems that can answer legal questions while grounding their responses in authoritative sources. However, existing Vietnamese legal benchmarks provide limited coverage of real-world legal consultations. We in…

Developmental psychology holds that certain priors are given to infants prior to experience rather than induced from data, and that the influence of such priors is suppressed under strong, well-constrained conditions but reasserts itself u…

Mamba-style and hybrid language models compress their past into a fixed-size recurrent state that is rewritten at every generated token. Storing this state in low precision saves memory bandwidth, but every rounding error is fed back into …

Fiber orientation and compartmental microstructure are central to the characterization of white matter tissue in diffusion MRI, yet existing methods either resolve fiber orientations without quantifying microstructure, or quantify microstr…

World Models are appearing as the next major frontier in computer vision. However, their robustness is currently largely unexplored. We identify the phenomenon of hallucination in latent World Models: given a state and an action, the predi…

LG 电子在首尔微软行业峰会上展示与微软合作研发的智能家居 AI 语音助手，基于语音到语音智能体技术并融合微软 Voice Live，支持用户插话提问时立即中断原回答并切换对话。该助手将应用于智能家居中枢 ThinQ ON，LG 计划年内推出搭载 Voice Live 的版本并为现有用户更新，场景还可扩展至酒店、办公室、零售门店等 B2B 空间。

Deep neural networks tend to rely on simple features that may be spurious and thus fail to generalize. We study this problem in the setting of linear probes, where a (generalized) linear model is fitted on the representations of a (pretrai…

郭明錤供应链调查显示，英伟达已开始测试以碳氢树脂为主的无玻纤HC覆铜板（CCL），替代此前测试的无玻纤PTFE基CCL，搭配无玻纤HC PP，用于Rubin Ultra NVL576交换机托盘PCB，该八机架互联方案预计2H27量产。

郭明錤最新产业调查显示，Nvidia 开始测试以碳氢树脂（HC）为主的無布 CCL 替代此前 PTFE 無布 CCL，搭配 HC 無布 PP，评估能否满足预计 2H27 量产的 Rubin Ultra NVL576（8 机柜互连方案）switch tray PCB 需求。