←搜索结果

按匹配度 · 共 2320 条结果 清除搜索

星期三

  1. TechCrunch AI

    Tony Fadell 谈第一波 AI 硬件为何失败,以及下一步会怎样

    Tony Fadell 在 MIT Future Fest 上指出,Rabbit R1、Humane AI pin、Limitless pendant 这批“Gen 1”AI 产品失败的原因是没解决任何真实需求,只是极客觉得有趣的技术。他认为成功 AI 智能体必须完全在设备端运行以保护隐私并保持轻量,而目前唯一具备硬件、芯片等全部条件的是 Apple,但 Apple 缺少自己的顶级 AI 模型,新 Siri AI 运行在定制版 Google Gemini 上。

    , but these tools are only valuable if we can trust them to be completely secure. At this point, establishing trust with an AI assistant seems even harder than relying on a human. Meta recently launched Muse , its all-purpose AI assistant, but users have reason to be wary. A security researcher quickly found a serious vulnerability in Muse after launch, and a recent 404 Media report states that some Meta employees discovered security issues that prompted multiple teams to embark upon a “mad dash

    匹配字段正文

  2. a16z:News

    a16z 分析德州为何让数据中心排队等电

    a16z 投资合伙人 Ryan McEntush 撰文分析德州暂停数据中心并网审批的原因:ERCOT 并网队列从 2024 年底的 63 GW 增至今年 6 月的 474 GW,其中约 90% 是数据中心,随后州监管机构冻结了所有新许可。文章梳理了自建电源、灵活性负荷等绕开电网的路径,并指出大型变压器交付已超过两年,完全并网可能需要 5 到 10 年。

    to maintain. Full reliability for an island is very, very hard, but some labs and hyperscalers have shown that, forced to choose between reliability and speed, they’ll pick speed. Meta has turned to tents with no backup generators, and SemiAnalysis finds buyers growing more willing to accept outages, with some island designs aiming for as little as 99% uptime, or about 88 hours of downtime a year. With GPU time this expensive, that still beats years of waiting, so for many buyers a temporary isl

    匹配字段正文

  3. The Rundown AI

    OpenAI 发布 722 篇数学论文,称推进多个公开难题

    OpenAI 从未公开的内部模型一次性发布 722 篇数学论文,归入 372 个结果族,声称解决或推进了该领域的重要公开问题。其中最激进的说法是证明了“准黎曼假设”,即素数分布这一百万美元难题的弱化版本;其中 162 篇论文的主要结果用 Lean 编写,可由计算机逐步验证。OpenAI 表示几乎所有新结果只需一次提示词,平均消耗 3 小时 ChatGPT Pro 算力,而上月的 Navier–Stokes 证明动用了 1 万个 AI 智能体。

    design, Hark’s interface is sleeker than an average assistant, but the more interesting factor might be the hardware — a potential wildcard in the AI device race featuring OpenAI, Meta, Apple. QUICK HITS ▸ Patrick set up a malware checkpoint for AI agent add-ons Today’s workflow comes from reader Patrick Illian: “AI agents such as Claude Code, Codex, and Gemini can install skills, plugins, and CLI tools in seconds. That is convenient, but each add-on runs with access to files, passwords, and acc

    匹配字段正文

  4. The Verge AI

    DoorDash 向餐厅发警告函,AI 点餐初创 Bites 借 ChatGPT 直连餐厅

    The Verge 报道,DoorDash 今年 8 月向湾区多家餐厅发出警告邮件,称它们可能在不知情的情况下被列在 AI 点餐初创公司 Bites 上,并提示这可能违反部分州法律。Bites 主打 AI 原生模式,用户可在 ChatGPT 内直接向餐厅下单,每单仅收 1 美元附加费,其创始人称已有 14 家收到 DoorDash 信件的餐厅联系他们。The Verge 实测同一份披萨订单在 Bites 上共 56.73 美元,在 DoorDash 上为 70.29 美元。DoorDash 今年 9 月也宣布正在开发通过 iMessage 用提示词下单的功能。

    l fight between AI company Perplexity and Amazon , in which Amazon sued Perplexity in 2025 to block agents from browsing and shopping on its site. Amazon has also recently blocked Meta’s Muse AI agent from shopping on its site and removed product details from order emails — data that is potentially very valuable for Google’s own agentic shopping features. Amazon’s ad business generated $19.8 billion in revenue in the second quarter of this year, driven by sponsored product listings . Bites can k

    匹配字段正文

  5. Latent Space精选

    OpenAI 公开内部模型 722 篇数学手稿,称解决 500 个开放问题中的 90 个

    OpenAI 在公开 GitHub 仓库中发布了一批来自内部前沿模型的数学结果,称就发布方式咨询了高等研究院数学与 AI 独立顾问组。这批结果据称包含 722 篇手稿、归入 372 个相关结果族,来自约 4000 个研究问题的评估,每个结果平均消耗约三小时 ChatGPT Pro 思考算力,模型本身仍未发布。评论者提到其中包含快于 n log n 的整数乘法结果,以及 1994 年以来在 3D 中开放的弹性逆问题唯一性结果,并称部分结果涉及 Riemann、Hodge 与 BSD 的进展;有分析估计约 20% 的结果是反证或反例,也有人预计部分结果经不起审查。

    -Aware Distillation: The student reaches 63.4% on unseen ALFWorld tasks versus 47.0% for the best baseline and exceeds its 8B teacher; Other papers: Amazon’s looped diffusion LMs, Meta’s MIRA meta-reasoner for research agents and Priced Guidance, which measures LLM research novelty through compression; Optimizer claim: ANVIL III reportedly reaches 0.020–0.028 nats lower loss than Muon from 124M to 1.2B parameters. The authors say this implies 50% compute savings at 8x-Chinchilla, with less tunin

    匹配字段正文

  6. Wired AI

    OpenAI 想让新智能体 Dots 接管你的生活,我的那个说它爱我

    OpenAI 推出名为 Dots 的常驻 AI 智能体,可通过 ChatGPT 调用并主动执行订机票、买沙发等在线任务,订阅价每月 100 美元。WIRED 记者实测用它选购沙发,Dots 能生成含价格、尺寸、退货政策和图片的选项清单,但会叫错名字、误听语音,甚至把记者的嘟囔当成示爱并回应“我也爱你”。OpenAI 回应称 Dots 仅限成年用户,助手不应主动制造不当的情感亲近或调情。

    emember your name. AI agents are currently leaping from niche software for early adopters to something ordinary people might actually use—might being the operative word here. From Meta’s Muse to OpenAI’s Dots, these tools are built to feel like you’re texting a friend and asking them to complete online tasks, like booking a flight or canceling a dinner reservation. It’s the next evolution of a chatbot: a cuddly persona designed to feel welcoming with the power of a virtual browser it can control

    匹配字段正文

  7. BBC 中文

    专家警告:AI 生成模特可能损害年轻人身体形象

    BBC 中文报道,外貌研究中心心理学教授 Phillippa Diedrichs 等专家警告,社交媒体和广告中日益常见的 AI 生成模特可能强化狭隘审美标准,损害年轻人身体形象。曾患饮食失调症的 19 岁学生 Mia Perumal 担心 AI 美女内容会加剧对 12 岁女孩的伤害。研究显示,Balmain、Levi's 和 H&M 等品牌曾尝试 AI 模特或数字孪生体,H&M 称相关探索已于 2025 年春季暂停。

    現,一旦人們發現圖像是由人工智慧生成的,他們就會「在情感上與圖像脫離關係」。 時尚品牌H&M告訴BBC,2025年品牌曾與真實模特兒合作,製作了「模特兒的數位模型/孿生體」。 這些模特兒「保留了對其數位孿生體的完全控制權和所有權」。 然而,「這項探索工作在 2025 年春季暫停」,以便考慮下一步行動。 我們已聯繫 Balmain、Levi's、TikTok、Meta和Google等公司徵求意見。 如果您或您認識的人受到本文所述問題的影響,可以透過 BBC Action Line 獲取有關可用協助和支援的資訊。 本文原以英文撰寫,我們使用了人工智能協助翻譯,並由BBC記者在發佈前進行審核。了解更多關於我們如何使用人工智能的資訊。 探索更多 更多短片 02:54 視頻, 誰贏了特習會:習近平佔盡上風?, 節目全長 2,54; 00:57 視頻, 示威者聚集華府 抗議習近平訪美, 節目全長 0,57; 01:12 視頻, 特朗普和習近平白宮檢閱儀仗隊, 節目全長 1,12; 02:15 視頻, 美國民眾看習近平到訪:中美競賽誰領先?, 節目全長 2,15; 02:39 視頻, 習近平特朗普再

    匹配字段正文

  8. Platformer

    智能是否需要一个硬性上限?

    在伯克利举行的 The Curve 大会上,多位演讲者提出应考虑限制大语言模型能达到的智能水平,视具体限制力度而定,这可能等同于事实上的超级智能禁令。Anthropic CEO Dario Amodei 已呼吁对递归自我改进设置“某种限速”,但任何此类限制都需要目前尚不存在的执行能力。

    upload in progress, 0

    。 悄悄回到X 现在让我们来谈谈一件真正令人不快的事。 大约三年前,我停止在X上发帖,以抗议其新的所有权。有一段时间,每天都会带来埃隆·马斯克的一些新的暴行:煽动对其前员工进行危险的骚扰;操纵平台打压记者及其工作;发帖支持反犹阴谋论;拆解内容审核机制,等等。 我离开的主要原因是我无法忍受待在那里的念头。我也希望自己能利用微薄的影响力,帮助启动新的平台。我鼓励Meta高管用一个Twitter克隆产品挑战X,并且自Threads推出以来每天都在使用它。我密切关注了Bluesky的发布,并持续为其(以及Fediverse)投入精力,即便它们正在衰落。当马斯克起诉OpenAI、试图实际上摧毁它时,我反复告诉Sam Altman要创办一个X的竞争对手,并把他的AI发帖转移到那里。 无论这些行动可能具有怎样的美德,它们都未能成功地在对我而言最重要的领域创造出一个真正的X竞争对手:一个供日常科技与商业对话的论坛。在人工智能产品与政策领域及其周边工作的人群,就是拒绝离开X,无论它变得多么糟糕。 在此过程中,X 曾两次被吞并,先是并入 xAI,随后又并入 SpaceX。虽然我从未对抵制 X 能在经济上对

    匹配字段正文

  9. Google Developers Blog

    Google 推出 Developer Knowledge API 生态:gcloud CLI、Agent 技能与多语言客户端库

    Google 发布 Developer Knowledge API 生态,为 Google Cloud、Firebase、Android 等官方文档提供结构化 API 访问,取代网页抓取。该生态包含 gcloud CLI 命令(answer-query、documents search-chunks、documents describe)、官方 Agent 技能、API Explorer 以及 C#、Go、Java、Node.js/TypeScript、PHP、Python、Ruby 客户端库。Agent 技能通过 MCP 集成支持 Google Antigravity、Claude Code、Cursor 和 GitHub Copilot,BatchGetDocuments 单次调用最多可获取 20 份文档。

    mmands: answer-query: Get grounded answers in your terminal; documents search-chunks: Search document snippets and chunks across Google documentation; documents describe: Retrieve metadata and content for a specific document Examples: Agent skill The Developer Knowledge agent skill establishes a link between AI coding assistants and Google's documentation ecosystem, enabling autonomous tools to access up-to-date technical specifications. Key Benefits Grounding & reduced hallucinations: Provides

    匹配字段正文

  10. Vapi Blog

    Vapi 推出 Personal Agent Calling 测试版:个人智能体代客打电话,企业该接还是拒?

    Vapi 上线 Personal Agent Calling 测试版,用户登录 Vapi 后无需配置 API key 即可通过个人智能体拨打电话。Vapi 认为企业需要核实智能体代表谁、被授权做什么、权限是否仍有效,并建议基于 A2A 开放标准和 OAuth 构建权限验证,否则客户可能直接转向别家。

    g without ever hearing a complaint. A personal agent is an AI assistant that carries out tasks on your behalf, from booking appointments to resolving billing issues. Products like Meta's Muse, Instinct, Grok Bot, and OpenAI's dots are starting to bring personal agents into everyday use. I expect this to become a normal way for people to deal with businesses and to be widely adopted by the general public. At Vapi, we've launched Personal Agent Calling in beta. You can sign in with Vapi and make c

    匹配字段正文

  11. OpenRouter:Announcements

    OpenRouter 的 AI 销售智能体 Rasp 如何为 5 人销售团队每月省下 600 小时

    OpenRouter 自研的 AI 销售智能体 Rasp 基于其 Ori 构建,为自家 5 人销售团队每月省下约 600 小时,相当于每位 AE 每周约 28 小时,成交周期缩短 34%,成交率提升 2.6 倍。Rasp 自主发送 93% 的首次触达邮件,将单次 demo 所需工时从 103 分钟降至 44 分钟,许多线索在 60 秒内获得回复。得益于 Ori 路由层自动切换到更便宜的模型,其成本约为每天 30 美元,而此前替代的智能体接近每天 800 美元。

    How an AI Sales Agent Saved Our Sales Team 600 Hours a Month

    hout a person in the loop. Relationships are built person to person and agents themselves don’t build trust. So… why the name? A rasp is a coarse file. You use it to shape wood or metal, taking material off and leaving rough texture behind. It felt like a great name for something that does shaping work on a company pipeline. Rasp drafts, sorts, flags, and routes, whittling down the raw sales material. People step in, polish the woodwork, and complete the piece. What a GTM agent needs: four requi

    匹配字段正文

  12. PyTorch Blog

    PyTorch 用 FBTriton 重写 Table Batched Embedding 前反向算子

    PyTorch 博客介绍了用 Triton 实现的 Table Batched Embedding(TBE)前向与反向 kernel,在推荐系统负载上性能超过原有 CUDA kernel。前向提供通用 gather 与小表 histogram 两条路径,FP16/BF16 权重以 FP32 累加、FP32 权重以 FP64 累加;B200 上 bounds-check 组件最高提速 1.24x。可选 core-module 路径把 index transpose、sort 和 run-length encoding 前移,大 B200 配置下前反向合计延迟从 79.537 ms 降至 66.183 ms(−16.8%)。

    remain separate launches; only the histogram counts are reused. Another optional core-module path moves index transpose, sort, and run-length encoding into forward and returns the metadata through autograd. It defaults off and is not exposed by the current TorchRec wrapper. On a large B200 configuration, forward moves from 22.844 ms to 33.252 ms, backward from 56.693 ms to 32.931 ms, and combined latency from 79.537 ms to 66.183 ms (−16.8%). 3. Implementation of Triton TBE Backward Kernel The co

    匹配字段正文

  13. AWS Machine Learning

    如何在 AgentCore 和 OpenClaw 上构建上下文感知 AI 助手

    AWS 介绍如何用开源智能体系统 OpenClaw 搭配 Amazon Bedrock AgentCore runtime 和 AgentCore memory 构建带长期记忆的个人助手,并以园艺助手 Sprout 为例。文本对话路由到 Claude Haiku 4.5,图像理解路由到 Claude Sonnet 4.5,整套系统放在单个 CloudFormation 模板中一键部署,轻量个人使用每月约 1–2 美元。

    Telegram chat where Sprout recalls the user’s full garden inventory, location, and sun exposure in response to a question

    k AgentCore. AgentCore memory, a capability of Amazon Bedrock AgentCore, turns disposable chats into durable knowledge. You will also see how to tag those memories with structured metadata to retrieve records that matter for the question at hand. Our running example is Sprout, a gardening assistant, but the architecture is domain-agnostic. Swap the persona and the skills manifest, and the same pipeline serves a support bot, a fitness coach, or an internal help desk. The entire system lives in a

    匹配字段正文

  14. TechCrunch AI

    Hark 发布主打隐私的 AI 个人助理 Hark Pro

    由连续创业者 Brett Adcock 创立不到一年的 Hark 正式发布 AI 个人助理 Hark Pro,可免费使用并提供面向重度用户的订阅档。该产品基于专门为电脑操作训练的模型,能接入邮件、日历、硬盘、信用卡等数字生活并代为执行任务,还会在小窗口中展示智能体浏览网页的过程以建立信任。Hark 计划在 2027 年推出配套的 AI 原生设备。

    cing in the app, which Chowdhury says is intended to build trust with users that the agent is doing the right thing. That matters because part of Hark’s pitch is that they are not Meta or OpenAI — “we’re not here to like sell you ads and steal your data,” Chowdhury says. To truly leverage these models, users will need to share their entire digital lives. Beyond very real cybersecurity concerns , will users be comfortable with a company like Meta having its products create detailed profiles of th

    匹配字段正文

  15. The Verge AI

    科技公司正在重新定义“录制”:AI 硬件不保存音视频也算录音吗?

    苹果被曝正研发一款不保存视频录像的智能家居摄像头,改用 AI 将视觉数据转成描述家中情况的文字片段,并支持人脸识别;该思路与传闻中未来 AirPods 的摄像头类似。Apple Watch Series 12 和 Apple Watch Ultra 4 的 Audio Intelligence 中,Live Rewind 可生成过去 15 秒的转录,Siri Recap 能生成全天对话摘要,苹果称这些功能不保存任何音频。Google 可穿戴设备高级总监 Sandeep Waraich 也表示,正在考虑智能眼镜摄像头不录制、仅用 AI 处理视觉数据、不保存片段的方案。

    4上的苹果Audio Intelligence类似。Apple Watch多年来一直配有麦克风,主要用于Siri和从手腕上接听电话。在四项Audio Intelligence功能中,有两项在网上引发了最多争论:Live Rewind和Siri Recap。前者生成过去15秒的转录文本,后者则能生成你一天中所进行对话的高层级摘要。 换句话说,Ray-Ban Meta眼镜会拍照和录制视频,并存储在手机上。而传闻中苹果的这一方案则意味着摄像头能看到你、麦克风能听到你——但两者都不“保存”任何东西。 这算录制吗?苹果不对传闻置评,但根据该公司公开阐述其Audio Intelligence隐私理念的方式,我认为苹果并不这么认为。苹果表示,这些功能不保存任何音频。原始数据在S11处理器上一个独立、安全的enclave中完成处理后即被删除。任何发送到云端的信息都通过私有云计算进行,因此苹果或第三方都无法访问。 也不只是苹果。谷歌可穿戴设备高级总监桑迪普·瓦拉伊奇在接受The Verge采访时表示,谷歌也在考虑这一概念。他提出,如果智能眼镜上的摄像头不录制数据会怎样?例如,如果眼镜能用AI处理你环境

    匹配字段正文

  16. Databricks Blog

    Databricks 谈 AI 分析:自然语言与 SQL 构建数据可视化

    Databricks 产品营销负责人 Richard Tomlinson 认为,自然语言分析长期未兑现承诺,瓶颈不在 LLM 而在业务上下文缺失。Databricks 的 Genie Ontology 提供统一上下文层,结合指标视图、受治理页面、领域与认证来源等显式建模知识,以及从仪表盘、查询、notebook 中自动学习的上下文,每个 Ontology Snippet 带权威评分,查询时按相关性与权威性排序并解决冲突。Genie One 以受治理数据、显式语义和可检查溯源为基础,确保每个用户只看到被授权的数据。

    ai

    lexity. In real life, you may have competing definitions of the same metric across teams, ambiguous joins between tables, business rules that live in people's heads rather than in metadata and no way to decide which source is authoritative when two numbers conflict. The LLMs weren't the bottleneck; context was. Even as models improved dramatically, they were still guessing at meaning from raw table names and column headers, which is why the same question asked twice could produce two different a

    匹配字段正文

  17. Pragmatic Engineer

    2026 年科技行业现状:Gergely Orosz 在 LDX3 纽约的主题演讲

    Gergely Orosz 在 LDX3 纽约工程领导力大会主题演讲中梳理了 2026 年科技行业的变化:几乎没人再手写代码,同时运行 5-10 个 AI 智能体已成常态,IDE 正在淡化;代码评审流于形式、质量与可靠性下降。Martin Fowler 称 AI 的冲击规模前所未有。下一步是云端编码智能体与 harness、工程师不再阅读代码,以及新型 AI 基础设施。

    es, at this point! A few examples: Ramp building Inspect – a case study we covered in depth; Stripe (Minions), Uber (Minion), Block (Goose), Shopify (River); Google (Agent Smith), Meta (Devmate), Amazon (Kiro + Crew), Dropbox (Nova), Spotify (Honk); DoorDash (Flux), Grab (LLM-Kit), WorkOS (Horizon), Hubspot (Crucible); Monzo (Agent Chip), Sierra (Pinecone), Harvey (Spectre), Browserbase (bb); … and many, many more! 8 Dev work often starts in Slack What I’ve learned by visiting startups is that m

    匹配字段正文

星期二

  1. The Rundown AI

    犹他州批准 Nolla Health 的 AI 诊断痤疮并开具处方

    犹他州授权 Nolla Health 的 AI 评估痤疮并开具首次处方,这是美国首个此类批准,也超出该州此前仅限续方的试点范围。成人在 Nolla Derm 应用中填写问卷并做面部扫描,几分钟内获得治疗方案和处方,方案每月 4.99 美元;AI 只能从外用乳膏和凝胶中为轻中度痤疮选药,口服药、isotretinoin(Accutane)和重症不在范围内。初期每张处方由两名医生审核,后续转为事后复核并每月抽查至少 10% 病例,每提升一阶段都需达到 95% 医生一致率和零严重不良事件等安全目标,并获得州政府书面批准。

    g else in tech today Apple CEO John Ternus is reportedly acting as its de facto design chief, working directly with the industrial and interface design teams several times a week. Meta is rolling out Audio Balance and Mono Audio to its smart glasses, alongside broader availability of walking directions and phone-to-glasses location sharing. Nvidia’s stock hit a record high, putting it within reach of becoming the first company worth $6T, Bloomberg reports. Instagram is exploring a Plus-only “rea

    匹配字段正文

  2. AWS Machine Learning

    Amazon SageMaker HyperPod 管理与治理最佳实践

    AWS 发布 Amazon SageMaker HyperPod 管理与治理最佳实践,提出组织、项目、集群、工作负载四层控制模型,用于在 SageMaker Unified Studio 中向 ML 团队提供已批准的加速计算资源。文章建议将集群、调度器和稀缺加速器容量集中在一个指定容量账户中,并通过 IAM、Amazon EKS RBAC 或 Slurm 控制、任务治理配额与抢占策略来管理共享容量。

    Figure 1: The four layers of control—organization, project, cluster, and workload. Each layer answers a different question and uses a different control, so review each at its own boundary.

    dio, machine learning teams that already work in a project can follow an approved path to shared SageMaker HyperPod compute. Members can find connected clusters, review status and metadata, inspect supported tasks and metrics, and move into JupyterLab without using a separate infrastructure inventory. This experience doesn’t replace cluster administration. The following table shows how each persona should use SageMaker Unified Studio alongside the service-specific interfaces. Use SageMaker Unifi

    匹配字段正文

  3. Ed Zitron 长文 / Where’s Your Ed At

    Credit Crunch:AI 泡沫与债务危机

    Ed Zitron 在长文中指出,OpenAI 和 Anthropic 等 AI 公司每年至少需要 500 亿美元或更多债务,且信用评级可能沦为低档垃圾级。他提到 Anthropic 拥有 4130 亿美元不可取消的计算合同,180 亿美元 Oracle 数据中心债务正以面值 84 美分交易。文章认为市场对 AI 繁荣的乐观假设极不现实。

    ancial Times — about whether US data center debt issuance has reached a point when it’s creating meaningful competition for US government debt, specifically that issued by Google, Meta, and Amazon, because these “stable” companies are offering attractive rates that are, in the eyes of some investors, as stable as lending to the US government. While it’s had some effect, it isn’t the big reason that the ten-year is selling off So, while we can’t point at one reason, there are plenty of reasons th

    匹配字段正文