OC

速度提高 100 倍：可以改变一切的 AI 原型 - Futura-Sciences

来自佛罗里达大学、加州大学洛杉矶分校和乔治华盛顿大学的研究人员开发了一款 AI 芯片，该芯片利用光来执行卷积运算，与传统电子电路相比，显着降低了能耗。该芯片集成了蚀刻在硅中的菲涅耳透镜，能够以几乎为零的额外功耗实现即时计算。这种方法将速度提高了 100 倍，精度与传统芯片相当，同时降低了能源消耗，这表明人工智能系统具有更高效、减少环境影响的潜力。

激进投资者表示，日本最大的厕所制造商被低估了人工智能的价值

《金融时报》提供各种订阅优惠，包括第一年数字访问 40% 的折扣、为期 4 周的试用期，然后按月付款以获取具有独家见解的完整数字访问，以及为个人和组织使用量身定制的其他计划。读者可以探索自己国家/地区提供的不同优惠。

欧洲隐私监管机构对埃隆·马斯克的 X 发起“大规模”调查 |美国有线电视新闻网商业频道

欧盟数据隐私监管机构已对其人工智能聊天机器人 Grok 生成的色情图像展开调查。上个月，法国和英国使用 Grok 创建了深度伪造图像（主要是妇女，也有儿童），导致全球强烈反对，随后法国和英国展开了调查。爱尔兰数据保护委员会 (DPC) 已对 X 是否遵守有关这些图像的欧盟数据隐私规则发起了“大规模”调查。此次调查正值监管压力和政府干预不断加大（包括禁令和罚款威胁）之际，迫使 X 限制 Grok 的图像生成能力。此外，英国信息专员办公室正在正式调查 X 因 Grok 生成的内容造成的潜在危害，而法国当局则突击搜查了 X 位于巴黎的办公室。这些行动凸显了人们对人工智能对年轻人潜在风险的担忧，并引发了对人工智能聊天机器人进行更严格监管的呼声。

Radnor 和 Council Rock 的学生用人工智能深度伪造了同学的照片。家长们表示学校未能保护他们的女儿。

12 月初，拉德诺高中的学生通过电话收到了关于同学拍摄的色情视频的令人不安的信息。这一事件以及涉及人工智能生成的深度伪造的类似案件引发了关于学校责任和家长担忧的争论。宾夕法尼亚州立法者正在推行针对深度造假的新限制，但学校认为解决这些问题的能力有限，尤其是校外事件。家长们认为学校应该采取更积极主动的措施，为受害者提供更好的支持。在拉德诺区和议会罗克区，家长们声称学校管理人员对这些事件轻描淡写，他们没有充分告知社区，也没有启动适当的调查，比如《第九条》性骚扰调查。建议修改法律，将人工智能生成的未成年人露骨图像归类为虐待儿童材料。学校面临着教育学生此类技术危险的压力，同时平衡与年轻人讨论敏感应用程序的潜在风险。

人工智能公司已买断西部数据 2026 年所有硬盘

西部数据的外置硬盘驱动器需求激增，尤其是来自主要人工智能相关客户的需求，导致个人消费者出现短缺。尽管有预测称传统硬盘将因 SSD 而衰落，但磁性物理学的进步和需求的增加重新燃起了人们对 HDD 的兴趣。西部数据首席执行官欧文·谭 (Irving Tan) 在财报电话会议上宣布，2026 年的产品至少在 2027 年之前就已售罄，优先考虑来自前七大客户的订单。这一转变凸显了数据中心和企业客户在消费市场中日益占据的主导地位，硬盘驱动器成为经济高效的云存储服务的关键。

印度人工智能峰会：代表们抱怨开幕日排长队且混乱

印度在德里举行的人工智能影响峰会在开幕当天面临后勤问题，包括会场排长队、过度拥挤和混乱。参与者报告了等待时间长、获得食物和水的机会有限以及摊位上的产品被盗等情况。尽管存在这些挑战，此次峰会仍旨在探讨人工智能的好处和坏处，并吸引了 100 多个国家的主要技术领导人参加。印度官员已对给您带来的不便表示歉意，并成立了“作战室”来解决报告的问题。

人工智能可以帮助解决气候问题的说法被斥为“绿色清洗”

根据 Beyond Fossil Fuels 和 Climate Action Against Disinformation 等非营利组织的一份报告，科技公司夸大了生成式人工智能在应对气候变化方面的作用。分析发现，没有证据表明谷歌的 Gemini 或微软的 Copilot 等流行工具可以显着减少排放。该报告的作者、能源分析师科坦·乔希 (Ketan Joshi) 表示，这些说法具有误导性，类似于“漂绿”。大多数断言来自科技巨头审查的国际能源署报告，缺乏有力的证据。Hugging Face 的 Sasha Luccioni 强调，虽然预测模型可能有利于气候努力，但生成式人工智能是资源密集型的。该研究显示，支持传统人工智能绿色主张的证据也很薄弱，只有 26% 引用了学术研究。批评者认为这种歪曲事实转移了人们对数据中心快速增长和能源消耗的注意力。

这位医生正在训练人工智能来完成她的工作。而且生意兴隆

Alice Chiao 博士曾是斯坦福大学的急诊医学讲师，现在通过强化学习来训练人工智能聊天机器人来模仿医疗专业知识。这个新兴行业被 Pitchbook 分析师 Dimitri Zabelin 估值为 170 亿美元，涉及法律、金融和娱乐等各个领域的专家，通过 Mercor 等平台教授人工智能模型。Chiao 强调确保人工智能提供准确和安全的医疗信息的重要性，并将其视为增强而不是取代人类医疗保健角色的工具。

A Once-in-a-Decade Investment Opportunity: 1 Magnificent Artificial Intelligence (AI) Software Stock to Buy Hand Over Fist Right Now | The Motley Fool Amazon's stock has been declining amid concerns over rising capital expenditures and an AI bubble, despite年初股价上涨并接近历史新高。公司计划今年投入200亿美元用于基础设施建设，远超预期，导致现金流增长放缓。然而，AWS业务表现出强劲的增长势头和高利润率，为亚马逊提供了财务灵活性以继续投资。通过与Anthropic等公司的合作，亚马逊正在构建一个高效的AI生态系统。尽管软件市场低迷，但基于市盈率趋势，亚马逊股票现处于人工智能革命以来的最低估值水平，被视作长期持有的良好机会。 Software stocks are cratering over fears of rising capital expenditures and artificial intelligence (AI) bubble concerns. It's been an odd year for Amazon ( AMZN +1.19% ) investors so far. In early January, the stock popped roughly 7% to $247 per share -- just inches away from notching a new all-time high . But over the last month, Amazon stock has been sliding -- now down about 14% on the year as software stocks continue to plummet. Let's dig into the driving forces behind Amazon's sell-off. From there, I'll detail why smart investors view this dip as a generational buying opportunity. Image source: Amazon. Why Amazon's $200 billion AI budget is worth it Prior to Amazon's fourth-quarter earnings call on Feb. 5, Wall Street was expecting management to guide for capital expenditures of about $150 billion for this year. Spoiler alert: Amazon plans to spend $200 billion, far more than investors were expecting. As Amazon continues to pour money into procuring GPUs from Nvidia , designing its own custom silicon, and building AI data centers, the company's free cash flow generation will decelerate. The combination of accelerating infrastructure spend and diminishing cash flow is what's causing investors to dump Amazon stock right now. Here's what investors are missing: During the fourth quarter, Amazon Web Services (AWS) generated $35.6 billion in revenue -- representing 24% growth year over year. This is the highest growth from AWS in 13 quarters , and data suggests the ride is just getting started. Backlog from AWS sits at a whopping $244 billion -- up 40% year over year and an impressive 22% from the third quarter. Importantly, AWS is a high-margin business for Amazon. While the company's profitability in the e-commerce division can experience pronounced swings due to seasonality and fluctuations in consumer buying patterns, AWS remains more resilient -- often generating operating margins in the mid-30% range. The consistency of AWS's profit margins is what fuels robust cash flow for Amazon -- providing the company with a high degree of financial flexibility to reinvest in the business, as it's doing now. Today's Change (1.19%) $2.37 Current Price $201.16 Beyond LLMs: How Anthropic is turning Amazon into an AI ecosystem One of the biggest contributors to the AWS renaissance is Amazon's relationship with Anthropic . Anthropic's Claude model is heavily integrated throughout the AWS ecosystem, particularly in a product called Amazon Bedrock. Moreover, Anthropic is also leveraging Amazon's custom Trainium and Inferentia chips. By partnering closely with a leading enterprise AI system and designing its own infrastructure, Amazon is quietly building a cost-efficient, vertically integrated AI stack. This strategy helps Amazon move beyond the world of basic chatbots and win over sophisticated, data-heavy enterprise workflows. Is Amazon's pullback a buying opportunity? I think most investors are overlooking the fact that Amazon is already bearing fruit from its AI-related investments. In my eyes, it makes complete sense for the company to double down and increase its infrastructure spend. While there may be a lag between the initial capital outlay and a meaningful return on investment for these buildouts, the evidence is already clear: AWS is becoming more than a capacity service, and the combination of Anthropic and custom silicon serves as strategic assets for Amazon's AI ecosystem. AMZN PE Ratio data by YCharts Nevertheless, the software bear market has brought Amazon stock to its cheapest levels during the entire AI revolution based on price-to-earnings (P/E) trends. I think Amazon is a no-brainer buy at this price point. The stock is heavily discounted relative to the company's upside potential, making it a compelling opportunity to buy and hold as the AI movement continues to benefit the hyperscaler.

人工智能就业启示录已经到来

人工智能在第四次工业革命中的崛起正在重新引发与第三次工业革命期间类似的关于工作岗位流失和经济混乱的担忧。与以往不同的是，这一波影响到了受过高等教育的专业人士以及蓝领工人，并可能导致不稳定的政治气候。随着人工智能开始加速取代白领工作，人们对失业的担忧加剧，有可能颠覆传统的政治联盟，并引发关于全民基本收入或股息的辩论。美国民主面临的挑战将是应对技术快速变革带来的社会变革，同时保持治理的稳定性和信任。