←搜索结果

按匹配度 · 共 3412 条结果 清除搜索

星期四

  1. Robotics & Automation News

    AutoStore 发布 2026 秋季新品:FlexBinsPort、AutoCase 与 AI 软件推进无人化履约

    AutoStore 公布 2026 秋季硬件与软件新品,包括可在单一工作站处理所有标准箱高的 FlexBinsPort(预计提升存储利用率 10-15%,2026 年第四季度发布),以及 2027 年第一季度起支持最大 600 x 400 mm 箱型的 AutoCase。软件侧推出基于生成式 AI 的 CubeExplore、CubeStudio 中的 AI 设计助手和 CubeFleet 车队维护系统,后者据估算可将非计划停机减少 35-45%。

    Francis AutoStore has unveiled a series of new and upgraded hardware and software products designed to reduce manual intervention in warehouse operations and move its automated storage systems closer to “lights-out” fulfillment. The company’s Fall 2026 product announcement includes new technology for handling different bin and case sizes, robotic picking for smaller businesses, AI-assisted system design and investment planning, and predictive maintenance capabilities. AutoStore says the addition

    匹配字段正文

  2. The Guardian Technology

    AI 会如何改变生物武器开发的风险与门槛

    英国军情五处负责人 Ken McCallum 警告 AI 可能“改变”现有乃至新型生物武器的风险格局。文章梳理了两类风险:无技术基础的恐怖组织可借 AI 聊天机器人补齐知识与实验排错,资源充足的国家行为体则可用 AI 放大已有专业能力,例如识别让病原体更致命的基因改动。专家同时指出,隐性知识、生物材料与设备获取、实验验证、规模化与投放仍是实质障碍,目前各国尚无针对此类滥用的正式政府政策。

    dles that remain. How could AI help people make bioweapons? It depends on who wants to make them. One risk is the rogue, unskilled bad actor or terrorist group that works from a garage laboratory. AI chatbots could help them skill up, by compiling relevant information, planning experiments and troubleshooting problems. It’s likely the work would focus on using existing pathogens and toxins. The other risk comes from well-resourced organisations and nation states that have plenty of skilled scien

    匹配字段正文

  3. Fortune AI

    Taser 制造商 Axon 强制推行 AI 使用,要求研发团队提效 25%-30%

    Taser 制造商 Axon 正强制要求约 2200 人的研发团队使用 AI,目标到 2026 年底将产品管理、设计和软件工程的工作量压缩 25% 至 30%。每个团队须实现 10% 的强制生产力提升,若加速超过 15% 可获得股票奖励。其 AI 工具包括用生成式 AI 起草警方报告的 Draft One、支持 50 多种语言的 Axon Assistant,以及自动打码的录像脱敏工具。

    When Rick Smith founded Taser International in an Arizona garage in 1993, his original vision involved taking a device that was originally created by a NASA scientist and redesigning it in a way that would make it palatable to law enforcement agencies. More than three decades later, Smith’s company—now called Axon Enterprise—has sold more than one million Taser devices and diversified its portfolio to include body cameras, drones, and license plate reader cameras to law enforcement, emergency re

    匹配字段正文

  4. Fortune AI

    FBI 称 OpenAI 工程师用妻子 ChatGPT 账号策划秘密录音设备

    美国司法部称,OpenAI 工程师 Abdulrahman Alfozan 于 9 月 24 日至 30 日通过与其妻子关联的 ChatGPT 账号生成搭建持续录音设备的提示词,并于 9 月 28 日将设备藏在旧金山办公楼一间高管会议室的电视后。法庭文件还显示,约六周前他曾用 OpenAI 内部生产力智能体查询可容纳六人以上、常用于高机密会议的会议室。该设备为 Seeed XIAO ESP32-S3 Sense,9 月 30 日被员工发现,存储卡文件无法恢复或解密;他于 10 月 9 日被捕,面临一项联邦窃听指控,最高可判五年。

    a TV in the room during a routine sweep, sources familiar with the incident told Fortune. It was a Seeed XIAO ESP32-S3 Sense, a thumb-sized board with a microphone and microSD storage, made by a Shenzhen-based company. This prompted cooperation between OpenAI and the FBI, and by Oct. 5, a camera had been placed in the room. In the photo in the complaint (and above), Alfozan was caught in the video at 1:20:47 p.m. that day, checking behind the TV. That same day, court documents say he left a bag

    匹配字段正文

  5. CNET

    Siri AI 将登陆部分旧款 Apple TV,不必急着换新硬件

    Apple 在 Apple TV 4K 新闻稿脚注中确认,第二代和第三代 Apple TV 4K 也将在今年晚些时候获得 Siri AI,该功能已出现在 TVOS 27.2 beta 中,首批支持美国、澳大利亚、加拿大、印度、新西兰、新加坡、南非和英国的英语用户。但旧款 HomePod Mini 无法获得 Siri AI,只有搭载 S10 处理器的 HomePod Mini 2 才能处理相关请求;单独购买的新款 Siri Remote 售价 59 美元,11 月 13 日发货,可兼容 Apple TV 4K 与 Apple TV HD。

    features of the third-generation model that was released in 2022. You’re paying for a faster A19 processor, internal microphones to pick up Siri commands and queries, 128GB of storage and an Ethernet port (the only configuration of this model). More from CNET It also comes with a revamped Siri Remote, which looks and acts like the previous one but has precision finding and a chime that can play when you’re using FindMy to figure out which piece of furniture is hiding it. The second- and third-g

    匹配字段正文

  6. CNET

    FBI 逮捕一名涉嫌在 OpenAI 会议室安装窃听设备的员工

    FBI 逮捕并起诉 OpenAI 产品团队员工 Abdulrahman Alfozan,他被指在仅限高管使用的会议室电视端口藏匿带摄像头、麦克风和 microSD 存储的微型开发板,用于窃录公司 AI 相关会议。OpenAI 于 9 月 30 日发现首个设备并通知 FBI,10 月 9 日将其解雇并逮捕,该指控最高可判五年监禁和 25 万美元罚款。OpenAI 称未发现产品安全或客户数据受影响,并称其前沿 AI 被用于分析该设备的取证图像。

    nt, filed in the Northern District of California, says Alfozan allegedly placed a “thumb-sized development board with a detachable camera sensor, digital microphone and microSD storage.” It was wrapped in toilet paper and stuck into a TV port. A second recording device was found in a bag Alfozan allegedly left in a printer room along with his wallet and ID. The conference room where he allegedly planted the bug could be reserved only by executives and their assistants — Alfozan did not have acce

    匹配字段正文

  7. Artificial Analysis

    Artificial Analysis 评测:Grok Voice 3.0 登顶语音到语音指数 86.9

    Artificial Analysis 评测显示,SpaceXAI 的 Grok Voice 3.0 以 86.9 分登顶其 Speech to Speech Index,高于 GPT-Live-1(Astra, medium)的 83.2 和 Gemini 3.8 Live Extended Thinking(High)的 82.6,较 Grok Voice 2.0 的 81.3 提升 5.6 分。该模型通过 SpaceXAI Realtime API 提供,每分钟 0.08 美元,输入音频每小时 4.80 美元;在 Speech Agent Arena 任务成功率上以 94.8% 居首,Big Bench Audio 音频推理得 98.5%,Big Bench Audio 平均首音频延迟 0.81 秒,慢于 Grok Voice 2.0 的 0.70 秒。

    2.6 and GPT-Live-1 (Sol, low) at 80.1. It is up on Grok Voice 2.0 at 81.3 (+5.6 points) and Grok Voice 1.0 at 73.7 (+13.2 points), released just under six months ago. The Index averages Speech Reasoning (Big Bench Audio), Agentic Performance (Tau Voice), Arena Preference, and Task Success Rate ➤ Speech Agent Arena: Grok Voice 3.0 ranks 3 in preference at Elo 1070, behind Gemini 3.1 Flash Live Minimal (1100) and Gemini 3.8 Live (1088) and ahead of GPT-Live-1 (Sol, low) at 1057, and 1 on Task Succ

    匹配字段正文

  8. AWS Machine Learning

    用 Amazon Bedrock AgentCore 构建可自我改进的文档处理系统

    AWS 介绍如何用 Amazon Bedrock AgentCore 构建智能体式 IDP 系统,由 8 个专用智能体在两条图工作流中完成文档分类、提取、校验与学习,通常可在 20 分钟内完成部署。新文档类型无需手工配置模板,智能体会自动生成提取提示词并保存复用,后续同类文档直接进入提取流程;系统结合 Amazon Textract、S3 Vectors、Aurora DSQL 与 AgentCore Runtime、Gateway、Identity、Observability 等能力。

    xtract and identifies the document type and sender. The Matcher agent searches for similar documents using vector similarity in Amazon S3 Vectors, a capability of Amazon Simple Storage Service (Amazon S3); Training workflow (new document types): When no match is found, the Instructions agent generates extraction prompts, the Extractor agent extracts structured data, and the Validator agent checks results against schemas and business rules in Amazon Aurora DSQL. If validation fails, the Instructi

    匹配字段正文

  9. Ai2 Blog

    Ai2 发布 HiRO-ACEv2,以约 1/300 成本生成全球 3 km 气候数据

    Ai2 发布 HiRO-ACE 系统的新版本 HiRO-ACEv2,其降尺度模型 HiROv2 可同时预测多个高分辨率气象场,生成全球 3 km 气候数据的成本约为高分辨率全球大气模型 X-SHiELD 的 1/300。在给定粗化的 X-SHiELD 数据时,HiROv2 对目标模拟高分辨率统计量的还原误差在 3% 以内,即使事件在百万个数据点中仅出现一次;它还能估算热带气旋总能量,包括训练中未出现过的风暴。目前 HiRO 只适用于与 X-SHiELD 匹配的 ACE 版本,尚不能用于更暖未来或 ERA5 等历史气候,团队计划训练包含升温 4°C 世界的新 X-SHiELD 数据集,并增加近地面温度、雪冰等输出。

    HiROv2 realistically generates strong winds flowing down mountain slopes. We also trained an updated version of ACE for this system. Over a 10-year simulation, its hurricanes, average climate patterns, and overall distributions of rain, wind, and pressure all line up well with X-SHiELD. That tells us ACE's coarse simulations are a solid foundation for producing long-term high-resolution climate data. Putting HiROv2 and ACE together, we generated a full year of global 3 km data. Its statistics c

    匹配字段正文

  10. The Decoder

    OpenAI 今年或发布一到两款新模型,旗舰仍需对齐工作

    据 The Information 援引知情人士,OpenAI 员工称年底前可能发布一到两款新模型,但公司正在开发的重大升级旗舰是否在其中尚不清楚。该旗舰模型仍需完成对齐工作并加强测试基础设施,且按特朗普关于先进 AI 系统评估的行政令,OpenAI 同意在模型交付外部伙伴前最多 30 天向美国政府开放访问。该旗舰与 Astra 一样采用循环 Transformer 技术,OpenAI 内部称其为 Recurrent Depth,可提升数学与编码表现并降低成本,但会把部分思考移入人类审查者无法查看的内部数值表示,削弱思维链监控,OpenAI 因此刻意限制了 Astra 的循环次数。

    old Microsoft in private. Microsoft took the section down after realizing it wasn't meant to be public. Looped transformers remain a sensitive part of OpenAI's model development, both during training and at inference, The Information reports. OpenAI chief scientist Jakub Pachocki has already acknowledged that chain-of-thought monitoring is growing "fragile" and "unfortunately trending in a negative direction." He also warned against an industry-wide race toward models without readable reasoning.

    匹配字段正文

  11. NVIDIA Developer · Generative AI

    NVIDIA NeMo Gym 优化西门子芯片设计智能体

    西门子 EDA 用 NVIDIA NeMo Gym 评估并改进 Siemens Fuse EDA AI Agent 的 Liberty 文件生成能力,基准综合得分从约 55% 提升至约 95%,智能体轮次从 155 降至 53,输出 token 从约 81K 降至 29K。NeMo Gym 是开放评估环境,通过领域专用验证器检查最终 Liberty 产物的覆盖率、有效性、相对验证、准确性和完整性,任一环节失败即给零奖励。

    t tokens fell from approximately 81K to 29K after refining skills and Model Context Protocol tool descriptions based on verifier diagnostics; A domain-specific verifier checks coverage, validity, relative validation, accuracy, and integrity of final Liberty artifacts, assigning zero reward if any gate fails; The same evaluation method applies to engineering agents in EDA, CAE, and other domains by grading final artifacts and optimizing skills through repeated rollouts Next Steps Run the NeMo Gym

    匹配字段正文

  12. Google Research

    Google 发布地理空间推理智能体技术报告,并开放 Google Earth 与 A2A 服务早期访问

    Google Research 发布地理空间推理智能体的技术报告,该智能体把 Gemini 与 Earth AI 基础模型、Google Maps Platform、Earth Engine 等数据连接,通过模块化技能、前置规划和独立空间工作区完成多步空间分析。报告在四个基准上对比四种配置,结果显示专用 Earth AI 技能对定量空间任务提升明显,而开放式任务还需要前置规划;团队同时发布 GeoKTools 和 GeoExpertQueries 两个新基准的开发集。Google 还扩大早期访问,在 Ask Google Earth 中预览深度分析能力,并向开发者提供实验性的 Agent-to-Agent API,WHO 非洲区域办事处、People in Need、GiveDirectly 等伙伴已参与测试。

    Putting Earth AI to work: Geospatial agents for our planet's biggest challenges

    alls only when a user question needs them. Before running complex operations, the agent builds a step-by-step analytical plan, selects appropriate fallback datasets when local coverage is sparse, and processes large map layers in a dedicated workspace. This approach avoids overwhelming the agent context with massive amounts of spatial data, while preserving accuracy and tracking data attribution. Every response shows the agent’s plan and the data sources used so experts can check assumptions or

    匹配字段正文

  13. Apple Machine Learning Research

    Apple 研究:用语音基础模型的语言选择性维度检测语码转换

    Apple 提出一种直接从语音中检测语码转换(CS)的轻量方法,利用预训练语音编码器帧级的语言信息判断一句话中是否同时出现两种语言。在 CS-FLEURS 的 14 个语言对上,Whisper-small 编码器取得平均 F1 0.798,高于 CNN 的 0.776 和 HuBERT 线性基线的 0.751。该方法只需少量 CS 数据调整决策规则,可用于筛选多语言语料。

    Bibtex Code-switching (CS), the alternation between languages within an utterance, is common in multilingual speech but underrepresented in pretraining corpora. Increasing CS coverage requires finding relevant utterances in large speech collections so that they can be included in training data. One way to find such utterances is to transcribe the speech and detect language alternation in the text, but this depends on obtaining accurate multilingual transcriptions. We propose a lightweight metho

    匹配字段正文

星期二

  1. Databricks Blog

    OLTP vs. OLAP:如何在事务型与分析型数据库之间选择

    Databricks 博客对比了 OLTP 与 OLAP 两类数据库:OLTP 面向支付、订单、库存等单条记录的高并发低延迟读写,依赖 ACID 保证数据完整性,常用 PostgreSQL、MySQL、SQL Server、Oracle Database 等关系型数据库;OLAP 则面向跨时间、产品、地区等维度的大规模历史数据扫描与聚合分析。两者通常并存,需通过 ETL 或 CDC 同步,随着数据新鲜度要求提高会带来延迟与管道开销,架构选择应跟随工作负载。Databricks 的 Lakebase 是面向事务型应用的全托管 Postgres 数据库。

    ai

    ness, data consistency, query breadth, and data movement part of the architecture decision. This article compares OLTP and OLAP systems across query patterns, data models, data storage, and latency, then shows how data pipelines, hybrid processing, and lakehouse architectures connect them so data teams can choose a better fit for the workload. TL;DR OLTP (online transaction processing) fits real-time transaction processing on individual records, such as online banking and e-commerce orders, whil

    匹配字段正文

星期三

  1. Databricks Blog

    Databricks 谈 AI 智能体 ROI 困境:采用快于回报证明

    Databricks 的 Amber Roberts 指出,企业 AI 智能体存在“ROI 缺口”:采购 AI 的人和日常使用它的人往往不是同一批人,多数组织仍处于试错阶段,在摸索哪些用例真正有效。她认为智能体的价值来自业务上下文——企业自己的数据、历史和约束,否则只能给出泛泛建议;同时“更快”不是衡量智能体生产力的正确指标,生产力提升必须伴随质量不下降才有意义。

    ai

    ments. It's no longer impressive to see how much someone produced alone in four hours, because there's a visible difference between that and what someone produces when they're leveraging AI well. That's the new baseline. But it doesn't replace deep knowledge of the subject matter. It just means the same instinct, the same taste, the same expertise, is now applied at a different scale. Where I see the biggest jump is when people take everything that used to eat their day, the recurring, schedulab

    匹配字段正文

  2. MIT News · Artificial Intelligence

    MIT 团队用机器学习改进 2 至 6 周天气预报

    MIT 等机构研究人员在 Nature 发表论文,提出一种机器学习后处理层,学习并修正现有天气预报模型的误差,在 2 至 6 周次季节区间把预报准确率提高一倍,并首次在该区间超过气候平均值。该方法作为可加装的 wrapper,与 ECMWF 模型结合后在国际预报竞赛中获第一名,对极端天气的预测也有统计显著提升,模型开源且运行成本低。肯尼亚气象部门已联系希望将其用于季风季预报。

    rnational forecasting competition. In a new paper in Nature, the researchers describe their approach and show how it enables forecasts that for the first time beat climatologic averages between two and six weeks. “Not only does this do better on time averaged forecasts for this range, which is already very useful for farmers and members of the energy industry trying to match supply with demand, but it also improves predictions of extreme weather events, whether it’s flooding or cold spells or so

    匹配字段正文

  3. TechCrunch AI

    Deezer 将推出过滤 AI 生成音乐的开关

    Deezer 正在开发一个开关,让用户从搜索结果中过滤掉 AI 生成的曲目、专辑和相关艺人,该功能预计今年年底前上线。Deezer 去年已加入 AI 检测工具,今年 7 月称其平台每日上传曲目中超过 50% 为 AI 生成,平均每天 9 万首。

    at time, AI-generated tracks represented 10% of daily uploads . In July, the company said that more than 50% of daily uploads to its platform were AI-generated tracks , with an average of 90,000 tracks per day. Other music streaming services have also created their own policies in relation to AI music. Bandcamp has completely banned AI tracks, while Tidal cut off monetization for people uploading AI music. Spotify has taken a leaner stance, only tagging AI tracks and artists more prominently on

    匹配字段正文

  4. Tomasz Tunguz 博客

    前沿模型智能接近,单位任务成本相差 3.8 倍

    五款前沿模型的智能水平相差不到 5 分,但单位任务成本相差 3.8 倍。Google 的 Argon 以每任务 1.99 美元处于最低位,比最贵的 Fable 5.1 低 74%,比 Opus 5.5 低 67%、比 Muse Wonder 低 56%、比 Astra 低 39%。作者认为这与 Google 自研 TPU 十余年的积累有关,并援引 Microsoft Maia 300 与 OpenAI 推理芯片 Jalapeño 的公开数据,判断随着每瓦智能每六个月翻倍,token 价格会继续下行,前沿模型利润率仍可维持在 40% 至 70%。

    months, pricing should continue to deflate in the world of tokens, with margins remaining in the 40-70% range, particularly for frontier models9. Cost per task is the weighted average cost to run one Intelligence Index task, aggregating input, cache hit, cache write, reasoning & answer token prices across the ten evaluations in Artificial Analysis Intelligence Index v4.3.2, weighted by each evaluation’s index weight. Source : Artificial Analysis, retrieved October 2026. All models measured at m

    匹配字段正文

  5. Rohan Paul

    Confido Health 融资 2250 万美元建医疗 AI 智能体

    Confido Health 累计完成 2250 万美元 A 轮融资,其中新增 1250 万美元由 z47_vc 领投,用于构建连接患者、医疗服务方、保险公司与药房的 AI 智能体,自动化预约、保险、转诊和处方等流程。美国每年医疗人力支出超 8000 亿美元、医疗 IT 支出超 2000 亿美元,到 2030 年医护人员缺口可能达 300 万。

    0B+ spent on healthcare IT - Potential shortage of 3M (!!!) healthcare workers by 2030 Yet appointments, insurance, referrals, and prescriptions still require coordination across fragmented systems. Confido Health is building AI agents that connect patients, providers, insurers, and pharmacies to execute and automate these workflows. If AI can reduce the administrative burden for existing healthcare teams, the impact extends beyond productivity... to the capacity of the healthcare system itself.

    匹配字段正文

  6. Tencent WorkBuddy

    WorkBuddy Library 让 AI 接续编辑文件

    WorkBuddy 推出 Library 功能,把文档、表格、幻灯片、HTML 和 Markdown 集中存放,让 AI 能接着上次的进度继续编辑已有内容、协作并保持工作流连贯。

    Your files shouldn’t just sit in storage. WorkBuddy Library keeps your docs, spreadsheets, slides, HTML, and Markdown in one place, so AI can pick up where you left off, edit existing content, collaborate, and keep the workflow moving. Create. Save. Edit. Reuse. https://t.co/AxSK5RYAcd

    匹配字段正文