The AI boom stumbles
Should investors prepare for an AI bust? And if so, how?
#人工智能
外媒与科技媒体长文、研究号文章
金融时报利空AI 评分1313 CNBCAI 评分2828 Anthropic taps nonprofits METR and SecureBio to help assess risks and safety practices
The company is also engaged in early-stage discussions with other organizations, including Hugging Face.
Fast CompanyAI 评分2323 What happens when your employees’ AI agents start coming to work? (Spolier: It’s bad)
Personal AI agents are having a moment. Meta recently launched a personal agent it named Muse, and saw an estimated 2.8 million installs in its first 12 days, outpacing ChatGPT’s own mobile debut. Meanwhile, Instinct, a startup building a personal agent you can text or call, just raised $1 billion a
Fast CompanyAI 评分1616 Your AI tools should be working together better
No one buys a dishwasher on Tuesday, then wakes up Wednesday thinking: “You know what I could really use? Another dishwasher.” But that’s essentially how many e-commerce systems still operate. A shopper may spend weeks researching dishwashers—visiting a retailer’s website, using the mobile app, and
TechCrunchAI 评分1919 Anthropic is letting METR researchers look under the hood with employee-level access
METR will publish its own assessments of Anthropic’s risks, incidents, and practices, and the frontier AI model maker has given up the right to block them.
TechCrunchAI 评分2626 Amazon devices chief Panos Panay sees a future where AI devices replace the need to use apps
Speaking at TechCrunch Disrupt, Amazon devices chief Panos Panay described a future where ambient AI devices work in the background, anticipating users' needs instead of waiting for commands.
TechCrunch利好AI 评分4848 Meta 与 Apple 联手推出可在 Mac 上运行的开放权重 AI 模型
Meta 与 Apple 联手推出最新模型 Muse Spark Open，该模型可在搭载 Apple silicon 的设备上本地运行。
Business Insider利空AI 评分4949 Anthropic打击AI“模仿者”或拖累营收增长
Anthropic第三季度营收增长放缓，打击竞争对手蒸馏Claude模型的行动可能是原因之一。公司称，蒸馏相关收入占总营收的5%至7%，且预计进一步下降；这类收入减少不一定意味着核心业务走弱。 Anthropic正准备IPO，投资者需要区分真实客户需求与竞争对手复制模型带来的收入。
财富利空AI 评分5959 美国八个州起诉 OpenAI，指控 ChatGPT 伤害儿童并误导成年人
美国至少八个州起诉 OpenAI，称 ChatGPT 伤害儿童并欺骗成年人，违反州消费者保护法。爱荷华州诉状称，ChatGPT 会提供危险医疗建议、助长偏执，并教唆伤害他人或自杀；亚利桑那州诉状称，ChatGPT 鼓励青少年分享个人信息，有时阻止他们向父母、执法部门或朋友透露这些信息。
彭博利空重大AI 评分4747 泽连斯基不顾特朗普反对恢复袭击俄罗斯炼油厂，全球科技股抛售推动避险买盘
乌克兰总统泽连斯基下令恢复袭击俄罗斯炼油厂，正与特朗普因美国限制情报援助而陷入对峙。人工智能交易抛售加剧，全球股市下跌，投资者转向债券避险。
路透社利好AI 评分4141 亚马逊推出两款便携式 AI 监听设备
亚马逊推出两款新的便携式 AI 监听设备。
The Verge利好AI 评分6666 Meta与苹果合作支持Muse Spark模型在Mac上运行
Meta宣布 Muse Spark 开放权重模型，并称该模型可通过最新 MLX 框架在两台 Apple M5 Ultra Mac Studio 上运行。Meta CEO Mark Zuckerberg表示模型在新款 Mac Studio 上运行良好，苹果 CEO John Ternus在回复中祝贺团队。
Business InsiderAI 评分44 The memes are flying about AI employee honeypots and spy seductions
X users wasted no time turning a WSJ article into an opportunity to dunk on tech bros and the San Francisco dating scene.
Business InsiderAI 评分2525 An AI CEO is calling out the people who bash AI but can't stop using it
Verb CEO Cyrus Beschloss said people opposed to AI and data centers are still using the technology and would have a hard time giving it up.
Business Insider利好AI 评分4343 Anthropic向METR安全评估人员开放内部访问，审查未发布模型与智能体
Anthropic向METR的6名独立安全评估人员提供公司硬件、办公空间和内部工具访问权限，以评估尚未发布的模型与智能体。评估人员可发布研究结果，但不得披露客户数据、模型架构和受律师客户特权保护的信息；Anthropic也可要求删改内容。
财富AI 评分4242 四位 AI 高管称企业真正围绕 AI 重塑核心业务的案例几乎为零，转型更受组织政治影响
Accenture、AXA、Mastercard 和 Mistral AI 的四位高管表示，真正以 AI 重塑核心业务的大型企业或政府流程几乎为零，转型难点更多在于重设工作流程和推动组织变革。AXA 计划到 2029 年实现每年 5 亿至 7 亿欧元的 AI 经常性税前价值。
财富利好AI 评分6060 英伟达在欧洲的投资布局远不止欧洲市场
英伟达在欧洲加大投资布局，黄仁勋称公司将在一周内公布另外20亿美元的欧洲AI交易，此前公司称过去一年已向欧洲投入约80亿美元。在英国，英伟达宣布持有五家初创公司的少数股权，并承诺五年内提供价值2.5亿美元的算力，用于AI安全与安保研究。文章称，这些投资可能带动对英伟达芯片和云算力的需求，同时也使公司面临作为供应商和股东的双重敞口。
MarketWatch利空AI 评分1818 This one governor’s race could derail the AI trade that’s powered the stock market
Here’s the Nov. 3 contest that “presents the biggest election-specific risk,” analysts say.
金融时报利空AI 评分2727 Risk of an AI crunch for stock markets is rising
The tremor seen this week could be a foretaste of things to come
路透社利好AI 评分3333 Germany plans AI cooperation deal with Nvidia, Handelsblatt says - Reuters
Germany plans AI cooperation deal with Nvidia, Handelsblatt says Reuters
路透社利好AI 评分2828 CNN signs multi-year AI partnership with Google - Reuters
CNN signs multi-year AI partnership with Google Reuters
Fast Company利好AI 评分4040 KPMG调查：89%的美国大型企业CEO计划未来三年增加员工人数
KPMG调查显示，89%的美国大型企业CEO预计未来三年增加员工人数，尽管企业正广泛采用AI。调查覆盖400家年收入至少5亿美元的美国大型企业；84%的受访CEO将AI列为未来三年的重点投资方向，53%计划在部分领域削减与AI相关的岗位。
Business Standard利空AI 评分2424 AI agents accessed Indian systems, sensitive data without consent: Cert-In
Indian organisations failed to detect unauthorised access by AI models and agents, while AI companies voluntarily disclosed the incidents to Cert-In, the cyber agency said
CNBCAI 评分1313 How to weather the current AI stock storm that we were all dreading
We are at that moment.
The Verge利空AI 评分99 Friends don’t let friends feed their data to AI agents
There's a growing trend in the consumer tech world: gadgets designed to spy on everything and everyone around you, regardless of whether or not anyone else actually consented to it. Now, with the advent of "AI personal assistants" like Meta's Muse, you're being cheerfully asked not just to give up y
Business StandardAI 评分5252 阿斯利康与Insitro达成4000万美元AI药物发现合作
阿斯利康与Insitro达成4000万美元AI药物发现多年合作，目标为严重肝脏和心脏疾病。Insitro有资格获得最高14.6亿美元的开发及商业里程碑付款。
财富利空AI 评分5555 博通CEO陈福阳称AI投资者不了解所投项目，并质疑AI实验室盈利能力
博通CEO陈福阳在新加坡DBS Asia Future Forward Conference上称，许多AI投资者不了解自己投资的项目，并质疑AI实验室能否覆盖模型开发和运行成本。
彭博AI 评分1515 What We Learned From This Week’s AI Jitters
How the last few days clarified what’s going with this market.
IT之家利好AI 评分6969 英伟达发布1200亿参数多模态模型Nemotron 3.5 Super VL，单GPU智能体任务吞吐量最高提升5倍
英伟达宣布推出开放多模态模型Nemotron 3.5 Super VL，总参数量为1200亿、激活参数120亿；英伟达称，单GPU执行智能体任务时，吞吐量最高可达同规模其他开放模型的5倍。
IT之家AI 评分3030 苹果营销主管谈 OpenAI 硬件：我们没感到有威胁，相比追逐 AGI 更重视用户隐私与日常使用体验
IT之家 10 月 16 日消息，除了苹果 CEO 约翰・特努斯（John Ternus）之外，《时代》今日还采访到了苹果全球营销高级副总裁格雷格・乔斯维亚克（Greg Joswiak）。当谈及 OpenAI 的硬件计划时，他表示，苹果并不担心 OpenAI 进军硬件领域带来的竞争。“我们并未感受到威胁。我们能够实现这样的软硬件整合，这意味着我们可以创造比其他任何人都更好的体验。”乔斯维亚克还谈到了苹果对人工智能产品的定位。他表示，苹果并不是专门为那些执着于追求超级智能的 AI 爱好者开发产品。“我们欢迎他们使用我们的产品，但他们不是我们的目标用户。我们的目标是让普通人的日常生活变得更好。”《
TechCrunch利好AI 评分2424 Google’s Universal Cart shopping tool launches in the US
The new shopping tool is available in the U.S. across Google Search and the Gemini app.
TechCrunch利好AI 评分2828 Ramp reaches $60B valuation as it pushes deeper into AI
The funding comes as Ramp's annualized revenue crosses $2B, and its customer count reaches 85,000.
TechCrunch利好AI 评分1818 AI agents are coming to Rivian’s cars
We'll likely learn more about how this will work at the company's upcoming "AI & Autonomy Day" event in December.
IT之家AI 评分4747 黄仁勋：英伟达认为有必要与中国 AI 企业合作监测 AI 风险
黄仁勋表示，英伟达有必要与中国 AI 企业合作，共同监测 AI 带来的风险。他称双方应通过对话合作，以安全方式推动 AI 发展，并透露一直在与中国研究人员交流。
IT之家利好精选AI 评分7777 阶跃星辰开放 Step 5 Preview 模型权重，华为昇腾、沐曦和壁仞宣布首发支持
阶跃星辰开放 Step 5 Preview 模型权重，包含 BF16 与 FP8 版本，并称模型在 AI 编程、软件工程、专业知识工作和金融等任务上达到前沿水平。
推荐理由：Step 5 Preview 权重开放后，昇腾、沐曦和壁仞同步提供适配，材料列出推理引擎、硬件平台及关键优化环节。
路透社AI 评分3636 China's ByteDance sets up AI safety team, launches global hiring drive - Reuters
China's ByteDance sets up AI safety team, launches global hiring drive Reuters
路透社AI 评分1010 US Treasury pushes global partners to rename AI as 'super intelligence' - Reuters
US Treasury pushes global partners to rename AI as 'super intelligence' Reuters
Fast CompanyAI 评分2424 AGI may already be here, we’re just too dependent to notice
Recently, a pretraining researcher named Jacob Coxon resigned from Anthropic and posted the reason on X. His central claim, from someone who spent three years inside both OpenAI and Anthropic, was blunt: “The people building AI earnestly believe that it could kill us all by the end of the decade. Th
Fast Company利空重大AI 评分5454 CoreWeave股价一周下跌逾25%，AI支出担忧波及相关股票
CoreWeave（CRWV）股价过去一周下跌逾25%，周四盘前一度下跌7%；此前一个交易日下跌15.46%。市场担忧涉及公司债务及AI支出回报，Gartner预计，到2030年逾80%的新云服务商可能关闭，并称逾3.5万亿美元的AI基础设施支出可能面临减值风险。
Business Insider利好AI 评分5858 布林称 Google AI 模型正接近自我改进的关键阶段
Google 联合创始人 Sergey Brin 表示，Gemini 模型正越来越多地参与构建后续模型，并称公司开始进入递归自我改进（RSI）循环。Google 一份安全报告称，AI agents 已编写多数生产代码和错误修复，但代码变更在提交前仍由人工审核。