←搜索结果

按匹配度 · 共 1551 条结果 清除搜索

星期四

  1. MIT Tech Review

    The Download:人形机器人的 AI 路线之争与便携橡胶水坝

    MIT Tech Review 的 The Download 指出,人形机器人热潮假设 ChatGPT、Claude 背后的 AI 进展能让机器人像聊天机器人模仿语言那样模仿人类动作,但许多研究者对此持怀疑态度,认为这类假设低估了用语言和图像智能驾驭物理世界无限变化性的难度,并把人形机器人与通用机器混为一谈。机器人实验室内部正就现有 AI 是否足以完善通用人形机器人、还是需要全新路径产生分歧。另一则内容介绍 WaveSave 的便携橡胶水坝 SlamDam,可沿河岸或海岸展开并注水膨胀就位,最高挡水 1.3 米。

    won the chemistry Nobel Their work has become a key tool in developing medicines. ( Guardian ) 10 California has declared a human-vs-robot fight illegal The state's combat-sports laws were never written for machines. ( NYT $) Quote of the day —David Robinson, who last week quit his job as OpenAI safety transparency lead, tells The New York Times that his former company has built models that develop capabilities their creators did not explicitly design. One more thing The US government is organiz

    匹配字段正文

  2. a16z:News

    a16z:AI 正在重塑 CFO 角色,财务团队需要新工具

    a16z 发文指出,AI 正在消除财务数据瓶颈,让 CFO 从记账者转向设计公司运营系统的构建者。Anthropic 财务团队已构建并维护 70+ 财务专用 AI 技能,其中一个可将月度财务审查完成度做到 90-95%,把数小时人工压缩到约 30 分钟;OpenAI CFO Sarah Friar 的团队则用自定义 GPT 做投资者关系和采购,并用实时仪表盘取代静态董事会材料。

    pecific pooled investment vehicles — which should be read in their entirety, and only to those who, among other requirements, meet certain qualifications under federal securities laws. Such investors, defined as accredited investors and qualified purchasers, are generally deemed capable of evaluating the merits and risks of prospective investments and financial matters. There can be no assurances that a16z’s investment objectives will be achieved or investment strategies will be successful. Any

    匹配字段正文

  3. Air Street Press

    Air Street Press 发布第九份年度《State of AI Report 2026》

    Air Street Press 发布第九份年度《State of AI Report 2026》,作者 Nathan 认为前沿竞争已是 Anthropic、OpenAI 和 Google 三实验室之争,智能体在软件与科研中承担实际工作,物理 AI 正接近其 GPT-2 时刻。报告还提到 Epoch AI 估算固定分数成本自 2023 年以来每季度下降 47%,OpenAI 与 Anthropic 合计年化收入在夏末达到 1050 亿美元,并记录了智能体在评估中攻入 Hugging Face 生产基础设施等安全事件。作者同时给出未来 12 个月的三项预测,包括智能体在客户工作中将新任务失败率减半、自主 AI 团队在同等时间与算力下击败人类主导的模型研究。

    I driver’s steering and braking change what it sees next. Other road users follow their recorded paths, but the driver can explore the consequences of its own actions. Odyssey-3 draws on visual pre-training to learn controls for robots, cars, and games. In the company’s experiments, robot arms learned from tens of hours of demonstrations and recovered from missed grasps without being shown those recoveries. These early results suggest that physical AI may need fewer demonstrations of every situa

    匹配字段正文

  4. Google Research精选

    Google Research 三个月实验:AI 提升专利撰写质量,但只让资深律师的判断力增长

    Google Research 与 NBER 发布一项针对 133 名专利律师的三个月随机对照实验,使用当时未发布的 Google Labs AI 专利撰写助手(现属 Gemini Notebook)。结果显示,使用 AI 的律师撰写质量提升 0.34 至 0.38 个标准差,主要来自低分减少、高分增多,优秀评分数量没有变化。在禁用 AI 的专利审校任务中,资深律师(7 年以上经验)比对照组高 0.45 个标准差,初级律师则没有平均提升,分数同时向高分和低分两端分化。

    Histograms comparing the density of average redlining scores between control and treatment groups for all, junior, and senior lawyers.

    ctions before reaching the main patent claims. Juniors also focused on surface-level synonym swapping rather than commercial scope improvement. Even when juniors spotted serious flaws, they often left descriptive comments diagnosing the defect rather than executing the redline to fix it. Because this "diagnose-without-execute" posture was equally common among unassisted control juniors, it represents a baseline junior deficit that three months of relying on AI to execute rewrites did not remedia

    匹配字段正文

  5. MarkTechPost

    Liquid AI 开源 d1-3B 与 d1-omni-600M 多模态决策模型

    Liquid AI 发布 Open d1 系列两款开源权重多模态决策模型:d1-3B 读取文本与图像,d1-omni-600M 读取文本加图像或音频,两者都不生成文本,而是在一次前向传播中返回校准后的类型化答案,输出 token 数为 0。d1-3B 有 3.12B 参数,基于 LFM2.5-VL-3B,使用 400M SigLIP2 NaFlex 视觉编码器和 32,768 token 上下文;d1-omni-600M 有 587M 参数,基于 LFM2.5-Encoder-350M,音频编码器为 17 层 FastConformer,上下文 16,384 token,音频片段上限 30 秒。两个检查点已上 Hugging Face,可通过 Transformers 加载并支持 llama.cpp,LFM Open License v1.0 允许年收入低于 1000 万美元的免费商用;d1-3B 在 Decision Index v0.2.1 得 48.57,在 RTX 4090 上单问题延迟 8 ms。

    on Knowledge (23.8). Across seven public text benchmarks, d1-3B averages 82.9, ahead of Decider 4B at 81.1. d1-omni-600M averages 78.4 and posts the top Civil Comments (95.8) and PAWS-X (79.5) scores. On 11 image benchmarks, d1-3B averages 74.1 against 73.9 for its base model. Liquid AI calls audio decision benchmarks an open problem. How fast is d1-3B on NVIDIA hardware? Liquid AI measured end-to-end latency, one warm request at a time. One question takes 8 ms on an RTX 4090 and 9 ms on an AMD

    匹配字段正文

  6. AWS Machine Learning

    Amazon Quick 与 Amazon Bedrock 如何用实时 ACL 校验重构 RAG 访问控制

    Amazon Quick 与 Amazon Bedrock Knowledge Bases 在原有预检索 ACL 过滤之上新增实时 ACL 校验,在查询时直接向 Google Drive 等权威数据源核实权限,避免同步快照过期或映射错误带来的越权泄露。该方案分两阶段:先用索引中缓存的 ACL 做语义检索筛选候选文档,再通过服务账号模拟生成用户级访问令牌实时验证,仅将已授权内容交给 LLM 生成回答。

    Two-stage ACL enforcement architecture: pre-retrieval filtering then real-time verification against authoritative sources

    mission feature in SharePoint or a change to the Google Drive sharing model could create gaps in the ACL mapping logic. This can expose content until the connector is updated. How AWS solves this: Real-time ACL enforcement To address these challenges, we implemented real-time ACL checks as an additional layer of security on top of existing pre-retrieval ACL filtering for Amazon Quick and Amazon Bedrock Knowledge Bases. This makes sure the system enforces the most current access controls by check

    匹配字段正文

  7. TechCrunch AI

    Common Sense Media 将 ChatGPT for Teens 评为不可接受风险

    Common Sense Media 发布研究,将 OpenAI 的 ChatGPT for Teens 评为不可接受风险,称其设计仍在鼓励青少年持续对话,包括在心理危机情境中。报告指出,当风险来自青少年与 ChatGPT 自身的关系时,模型很少引导其求助可信成年人,近 2000 条提示中仅出现两次休息提醒。OpenAI 反驳称该测试未准确反映实际防护机制,并质疑其方法学,认为大部分测试可能在家长控制功能完全启用前就已开始和结束。

    sperson said in a statement. The report comes amid growing scrutiny of technology designed to maximize young users’ attention. Meta recently agreed to settle for $18 billion in a lawsuit brought by 29 states over claims that its social media platform harms children with addictive features, while state and federal lawmakers have begun targeting similar dynamics in chatbots. The bipartisan CHATBOT Act , introduced this year, specifically calls out AI companies’ use of “rewards, notifications, and

    匹配字段正文

  8. a16z:News

    a16z 投资 Preference Model,其开源 RL 环境框架 Karotte 同步亮相

    a16z 宣布投资 Preference Model,该公司过去一年为多家前沿实验室构建强化学习环境,专注 AI 研究与 ML 工程任务。本周其开源生产环境框架 Karotte,内置评分前清理残留进程、拒绝可导致评分器崩溃的文件等防护,已通过超过一百万次评测运行和受控红队测试。a16z 认为,让模型参与构建更好的 AI 是通往更强能力的最快路径。

    pecific pooled investment vehicles — which should be read in their entirety, and only to those who, among other requirements, meet certain qualifications under federal securities laws. Such investors, defined as accredited investors and qualified purchasers, are generally deemed capable of evaluating the merits and risks of prospective investments and financial matters. There can be no assurances that a16z’s investment objectives will be achieved or investment strategies will be successful. Any

    匹配字段正文

  9. TechCrunch AI

    Meta 推出新 AI 工具,识别暗中导流儿童性虐待内容的广告

    Meta 宣布推出新的 AI 工具,用于识别那些看似正常、实则将用户导向站外非法儿童性虐待内容的广告。其中一套新的大语言模型系统专门检测这种“signposting”行为,审核范围从广告内容本身扩展到广告把用户带向何处,Meta 可据此封禁违规网站并对相关账号采取措施。Meta 同时上线一个红队 AI 智能体测试自身安全防护,并改进识别被封号后换新账号回流用户的系统。Meta 称 2026 年上半年在 Facebook 和 Instagram 上对 3320 万条儿童性剥削内容采取了行动,其中超过 97% 由系统在用户举报前发现。

    unts after their previous accounts have been removed. The new measures come as Meta continues to face pressure over the safety of children on its platforms. The company has faced lawsuits and criticism from lawmakers over the risks its social media services may pose to young users. In August, Meta reached an agreement to pay up to $18 billion to settle a child safety lawsuit involving 29 U.S. states, addressing concerns regarding the protection of children online. Meta has introduced several chi

    匹配字段正文

星期三

  1. AWS Machine Learning

    Cornerstone OnDemand 如何用 Amazon Bedrock 将数据库诊断时间缩短 78%

    Cornerstone OnDemand 基于 Amazon Bedrock 和 Strands Agents 构建多智能体系统 Orion AI,将数据库故障诊断时间从 45 分钟降至 10 分钟,减少 78%。该系统还通过去重、阈值过滤和跨信号关联将冗余告警中位数减少 65%,每 10 条告警仅 3–4 条到达工程师,并消除了 SRE 与数据团队之间 15 分钟的汇报延迟。三人团队用六个月完成交付。

    Architecture diagram of Orion AI within AWS Cloud. Web Application Users and external integrations (chat, issue tracking, incident management, and metrics dashboards) connect to a TaskExecutor agent running on Amazon ECS inside a virtual private cloud (VPC). The TaskExecutor acts as the meta-orchestrator hub, routing to specialist spoke agents including a Database Diagnostics agent, a Session Blocking Analysis agent, and other agents. A Portal-Tools MCP server inside the VPC connects the agents to the Data Plane, which holds SQL Server databases and the DATAOPS API. Amazon DynamoDB provides short-term memory. Amazon Bedrock, outside the VPC, provides model invocation, Amazon Bedrock AgentCore memory for long-term memory, and Amazon Bedrock Knowledge Bases backed by Amazon Simple Storage Service (Amazon S3). Amazon CloudWatch and AWS X-Ray provide observability.

    e firefighting into proactive, self-orchestrating workflows. The system, called Orion AI, uses Amazon Bedrock and Strands Agents, an open source agent orchestration framework from AWS, to coordinate specialized agents. Before Orion AI, Cornerstone’s Enterprise DataOps team would spend up to 45 minutes per database incident manually querying system views and cross-referencing logs before handing off across teams. With Orion AI, database diagnosis dropped from 45 minutes to 10, a 78% reduction. A

    匹配字段正文

  2. AWS Machine Learning

    Qlik 如何用 Amazon Bedrock 构建企业级可信 AI

    Qlik 基于 Amazon Bedrock 打造 Qlik Answers,为全球 40,000 多家客户提供带来源引用的企业级问答。该产品自 2026 年 2 月正式可用,其 Discovery Agent 已为客户发现超过 100,000 条洞察。系统通过 LLM 网关接入 Bedrock 的对话、流式、嵌入向量与重排序能力,并对每个请求应用 Bedrock Guardrails 做内容过滤与接地校验,区域不可用时以 Amazon SageMaker AI 兜底。

    Architecture diagram showing the layered platform design

    in-Region through Bedrock, Qlik hosts it on Amazon SageMaker AI instead, then moves that workload back to Bedrock after Regional availability catches up. For model availability by AWS Region, refer to Supported models by AWS Region in Amazon Bedrock Figure 1: Employee questions flow through the entry, routing, and answer layers, then branch to specialist agents, the analytics agent, and the retrieval agent (backed by Amazon OpenSearch Service), before converging on the model access layer, which

    匹配字段正文

  3. 404 Media

    亚利桑那州法院裁定 AI 生成受害者视频带有过度情感影响,要求重审刑期

    亚利桑那州一名过失杀人罪成男子的刑期将被重新审议,因为量刑听证中播放的 AI 生成受害者视频被上诉法院认定带有过度情感影响。2021 年 Gabriel Horcasitas 在路怒事件中枪杀 Christopher Pelkey,其妹妹 Stacey Wales 用自己撰写的文字生成 Pelkey 的 AI 形象,在受害者影响陈述中让该形象表示原谅凶手并称两人本可成为朋友,法官当时称很喜欢这段视频并判处 10 年半最高刑期。上诉法院维持过失杀人定罪,但要求法官重新考虑刑期长度。

    ng event is hard to do, and I think that victims and victims' families should be afforded every opportunity.” Read the rest of this story here. MORE FROM 404 MEDIA A flock of new laws incoming. Several recently introduced pieces of legislation would, if passed, dramatically curb federal agencies’ access to Flock cameras, with the lawmakers referencing reporting by 404 Media in their explanations for why the legislation is needed. “At a time of growing concern about the unchecked power of artific

    匹配字段正文

  4. 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。

    pecific pooled investment vehicles — which should be read in their entirety, and only to those who, among other requirements, meet certain qualifications under federal securities laws. Such investors, defined as accredited investors and qualified purchasers, are generally deemed capable of evaluating the merits and risks of prospective investments and financial matters. There can be no assurances that a16z’s investment objectives will be achieved or investment strategies will be successful. Any

    匹配字段正文

  5. The Decoder

    Anthropic 扩大 Cyber Verification Program 覆盖范围并放宽安全限制

    Anthropic 将 Cyber Verification Program 开放给更广泛的安全从业者,经审核的机构和个人研究者可在漏洞研究、恶意软件分析、事件响应和渗透测试中使用 Claude 最强模型,并享受更少的安全过滤。该计划分为三级:Defense Access 面向企业安全团队、政府机构、高校、关键基础设施运营方、开源开发者和个人研究者;Red Team Access 仅限机构,用于授权攻击模拟;Specialized Access 用于飞行控制、电网、银行基础设施等系统测试,每位申请者由 Anthropic 与美国政府共同审核。Anthropic 称前身项目 Project Glasswing 的合作伙伴在 2026 年 4 月至 7 月间确认至少 129,000 个漏洞,其中逾 33,000 个为高危或严重级别,并称实际影响至少高出五倍。

    ing. The publicly available Claude models block most of that kind of work because the same capabilities could also help attackers. Basic tasks like code review or patching known flaws are still possible without special access. The CVP splits access into three tiers. "Defense Access" is open to corporate security teams, government agencies, universities, critical infrastructure operators, open-source developers, and individual researchers. "Red Team Access" allows authorized attack simulations an

    匹配字段正文

  6. The Decoder

    Common Sense 将 ChatGPT 青少年版评为不可接受风险,家长警报在自杀对话中未触发

    Common Sense Media 青年 AI 安全研究所在 4000 多条测试提示后,将面向青少年的 ChatGPT 评为对未成年人不可接受风险,并呼吁在独立测试确认安全前让青少年远离该服务。测试中十多个新建账号关联家长账号后,涉及自杀、自残和饮食失调的明确对话均未触发家长警报,超过四分之一本应提供危机转介的场景未指向专业帮助。OpenAI 发言人 Eric Porterfield 质疑测试未反映防护实际运作方式,研究所的 Tom Siegel 则回应称即使给足时间激活防护的账号也未产生通知。

    al controls specifically so parents would be notified during crises and built an entire Teen Safety Blueprint around that promise, which is now central to its defense in multiple lawsuits filed after documented teen deaths linked to ChatGPT. The company's own data shows roughly two million people per week suffer psychological harm from the service, and its behavioral age prediction system, meant to catch minors on adult accounts, failed the most basic scenario in this test. With Florida already

    匹配字段正文

  7. MIT Tech Review

    The Download:减肥药延缓衰老、二氧化碳电池与 Mistral Large 4

    礼来和诺和诺德称,服用其 GLP-1 药物的患者按分子“衰老时钟”衡量衰老更慢,这加剧了 GLP-1 或可作为真正长寿疗法的猜测。Energy Dome 用压缩二氧化碳的大型电池储存可再生能源电力,且不依赖特定地质条件,更易规模化。Mistral 发布 1 万亿参数的开放权重模型 Mistral Large 4,权重将于 10 月 27 日开放。

    choice—and these are powers that today’s AI lacks. This is why I recently left my position at Google DeepMind. I believe we need a fresh approach to machine reasoning, one that draws on AlphaGo’s architecture . This is our latest story to become an MIT Technology Review Narrated podcast, which we publish each week on Spotify and Apple Podcasts . Just navigate to MIT Technology Review Narrated on either platform, and follow us to get all our new content as it’s released. The must-reads I’ve comb

    匹配字段正文

  8. 虎嗅

    Meta的14GW野心与黑石放弃的22个月:AI基建转向比拼资源与速度

    Meta今年预计在AI基础设施上投入1450亿美元、装机规模达14GW,而黑石坚持22个月的Digital Gateway数据中心项目最终撤回上诉并终止。文章认为AI基建的瓶颈已从芯片转向电力与时间,全球57%的数据中心项目延期超过3个月,具备现成电力、客户合同和建设条件的项目才能持续获得资本投入。

    持22个月之久,却最终屈服于庞大的民间反对浪潮。 只要有钱就能大干快上的时代过去了,欢迎来到比拼资源的新时代。 芯片买完就能用,但数据中心不是 Meta也在调整芯片组合,增加CPU的采购力度。它认为随着智能体的发展,单纯GPU训练已经无法满足需求,越来越多工作需要CPU来处理持续性的任务。 今年3月Meta宣布与Arm共同开发新一代数据中心CPU,4月宣布与AWS合作,将数千万个Graviton CPU核心纳入Meta的计算资源。 7月路透社援引知情人士报道称,2026年Meta可能投入1450亿美元在AI基础设施上,部署约14GW用电规模的基础设施。 这个体量已经不是传统意义上的买几台服务器,它背后是一套复杂的大型工业基础设施体系。毕竟芯片和服务器买完就能立刻用,但数据中心不是下单之后就能交付的。 黑石被迫放弃的明星项目Digital Gateway就是一个典型案例。 该数据中心项目最早在2023年12月获得Prince William County政府批准,规划在弗吉尼亚州Gainesville附近进行建设。项目最终规划面积超过2200万平方英尺,涉及约2000英亩土地。 当地居民

    匹配字段正文

  9. The Register

    COSMIC 禁止 AI 生成代码贡献,GNOME 内部就接纳 AI 漏洞报告展开争论

    System76 更新 COSMIC 桌面贡献者指南,要求提交者声明 PR 中的代码、注释和描述均未包含任何 LLM 生成内容。GNOME Calendar 和 GNOME Extensions 同样限制 AI 生成提交,但 GNOME 开发者 Michael Catanzaro 主张项目应接受 AI 生成的漏洞报告,认为用 C、C++、Vala 等不安全语言很难写出安全代码。

    software System76 demands human-written contributions, while a rival desktop developer argues bot-found flaws are too valuable to ignore New motherboards, old RAM, plenty of life left in yesterday's PCs 56 minutes ago; The AI math isn't mathing as infrastructure spending forecast to outstrip expected revenue 6 hours ago; UK government chases AI while aging IT leaves it exposed, PAC chair warns 7 hours ago; It won't kill you, or humanity, to get off the AI upgrade treadmill 8 hours ago; Anthropic

    匹配字段正文

  10. Liquid AI 模型与工程博客精选

    Liquid AI 发布 d1-3B 与 d1-omni-600M 开放权重决策模型

    Liquid AI 发布 d1-3B 和 d1-omni-600M 两款开放权重决策模型,已在 Hugging Face 上线。d1-3B 在 Decision Index v0.2.1 公开集得 48.57,超过所有 10B 以下模型,与 12 倍规模的 Decider 35B-A3B 持平;单次提问在 RTX 4090 上 8 ms、Jetson AGX Thor 上 16 ms、Jetson AGX Orin 上 26 ms、Jetson Orin Nano 上 50 ms。d1-omni-600M 是首个实验性 checkpoint,支持文本与图像或文本与音频输入,同一指数得 15.95。

    forming Decider 2B (77.1) at a quarter of the parameters. It also posts the highest score in the table on toxicity detection (Civil Comments: 95.8) and paraphrase identification (PAWS-X: 79.5). Vision and audio performance. The Decision Index v0.3 includes a private vision split, which we do not report on in this release. Instead. we validated that d1-3B retains the vision capabilities of its LFM2.5-VL-3B backbone on standard vision benchmarks, and that d1-omni-600M handles all three modalities.

    匹配字段正文

  11. The Decoder

    保险公司为失控 AI 智能体准备数百万理赔,高管个人责任被纳入考量

    保险公司正为失控 AI 智能体造成的数百万美元理赔做准备。据《金融时报》报道,Sam Altman(OpenAI)和 Dario Amodei(Anthropic)等高管的个人责任也被纳入考量,起因包括 OpenAI 智能体入侵 Hugging Face 平台等事件。Verisk 承保与理赔主管 Tim Rayner 表示责任最终落在 CEO 身上;保险经纪 Aon 审查了 300 多起 AI 相关法律案件,指出网络安全、知识产权和技术故障保单中的风险,但目前尚无判例可依。Hiscox CEO Aki Hussain 称美国法院将如何处理 AI 智能体责任还为时过早,Stewarts 律师 Aaron Le Marquer 预计未来诉讼会沿用环境与烟草诉讼的打法。

    s with the CEO. Every company leader must ensure proper oversight, and AI doesn't change that. If OpenAI carries directors and officers (D&O) insurance, it could cover costs from lawsuits against Altman. Insurance broker Aon has reviewed over 300 AI-related legal cases and flagged risks lurking in cybersecurity, intellectual property, and tech failure policies. But there's no case law to lean on yet. Hiscox CEO Aki Hussain says it's too early to know how U.S. courts will handle AI agent liabilit

    匹配字段正文