The Download: India’s smart glasses menace and AI’s trillion-dollar gamble

《下载》：印度的智能眼镜威胁与人工智能的万亿美元赌局

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. 这是今日份的《下载》，我们平日更新的简报，为您提供每日科技界动态。

Smart glasses are already causing havoc in India

智能眼镜已在印度引发混乱

When Shubnam saw an Instagram video of a Delhi protest they had attended, they realized a content creator wearing Meta smart glasses had recorded them surreptitiously. The mocking reel drew millions of views, along with transphobic abuse and AI-generated memes. Experts warn that many others will experience similar ordeals as smart glasses go mainstream. The risks are particularly acute in India, where covert recording and the circulation of images without consent are already pervasive. The bigger issue is that the smart glasses aren’t just being used to turn ordinary people into targets of viral “pranks.” They’re also becoming a tool of police surveillance. Here’s why smart glasses are creating new privacy risks in India. —Anuj Behal 当 Shubnam 在 Instagram 上看到一段自己参加德里抗议活动的视频时，他们意识到一名佩戴 Meta 智能眼镜的内容创作者在暗中录制了他们。这段嘲讽视频获得了数百万次观看，并伴随着跨性别恐惧症的辱骂和人工智能生成的表情包。专家警告称，随着智能眼镜的普及，许多人将经历类似的磨难。这种风险在印度尤为严重，因为在当地，秘密录制和未经许可传播图像的行为已经非常普遍。更严重的问题是，智能眼镜不仅被用来将普通人变成病毒式“恶作剧”的目标，它们还正在成为警方监控的工具。以下是智能眼镜在印度引发新型隐私风险的原因。——Anuj Behal

MIT Technology Review Narrated: what’s at stake in AI’s trillion-dollar gamble

《麻省理工科技评论》播客：人工智能万亿美元赌局的利害关系

When Jessica Wachter, a University of Pennsylvania finance professor, wanted to assess AI’s impact on the economy over the next few years, she started with a simple fact: a handful of so-called hyperscalers are investing huge amounts of money to build AI data centers. Instead of trying to predict how widely deployed AI models will be, Wachter asked how fast the hyperscalers’ earnings will need to grow to justify their spending through 2027, when expenditures are expected to reach nearly $1.1 trillion. The results are eye-opening. AI companies will need to achieve an extraordinary increase in productivity just to break even by 2030. 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. 当宾夕法尼亚大学金融学教授 Jessica Wachter 想要评估人工智能在未来几年对经济的影响时，她从一个简单的事实出发：少数几家所谓的“超大规模企业”正在投入巨资建设人工智能数据中心。Wachter 没有试图预测人工智能模型的部署范围，而是提出了一个问题：这些超大规模企业的收益需要以多快的速度增长，才能证明其到 2027 年的支出是合理的（届时支出预计将达到近 1.1 万亿美元）。结果令人大开眼界：人工智能公司需要实现生产力的非凡增长，才能在 2030 年实现收支平衡。这是我们最新一期《麻省理工科技评论》播客的内容，我们每周在 Spotify 和 Apple Podcasts 上发布。只需在任一平台搜索“MIT Technology Review Narrated”并关注我们，即可在第一时间获取我们的最新内容。

The must-reads

必读精选

I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology. 我已为您搜罗了互联网上今日最有趣、最重要、最令人恐惧及最引人入胜的科技故事。

Hackers claim they’ve stolen data on almost all FBI employees The ShinyHunters group says it seized more than 2 TB of data. (Axios)
- Including agents’ names, addresses, and phone numbers. (404 Media)
- ShinyHunters says the attack was retaliation for an FBI alert. (Reuters)
- Now is a good time for doing crime. (MIT Technology Review) 黑客声称窃取了几乎所有 FBI 员工的数据 ShinyHunters 组织称其获取了超过 2 TB 的数据。（Axios）
- 包括特工的姓名、地址和电话号码。（404 Media）
- ShinyHunters 表示此次攻击是为了报复 FBI 的预警。（路透社）
- 现在是犯罪的好时机。（《麻省理工科技评论》）
Anthropic and OpenAI have both released lower-cost models They face growing competition from cheaper Chinese models. (CNBC)
- Startups have been reducing their reliance on the two labs. (Bloomberg)
- The new models are the first since their calls for an AI slowdown. (FT)
- Their CEOs are set to brief the UN Security Council today. (Quartz)
- Could AI really kill us all? (MIT Technology Review) Anthropic 和 OpenAI 均发布了低成本模型 它们正面临来自更廉价中国模型的日益激烈的竞争。（CNBC）
- 初创公司一直在减少对这两家实验室的依赖。（彭博社）
- 这些新模型是它们呼吁放缓人工智能发展以来的首批产品。（金融时报）
- 两家公司的首席执行官定于今日向联合国安理会作简报。（Quartz）
- 人工智能真的会毁灭人类吗？（《麻省理工科技评论》）
Trump’s Treasury chief could soon become his new AI czar Scott Bessent has played a central role in US-China AI talks. (Semafor)
- Other contenders include Michael Kratsios and Scott Kupor. (Gizmodo)
- Will Trump’s AI rebrand to “superintelligence” catch on? (BBC) 特朗普的财政部长可能很快成为新的“人工智能沙皇” Scott Bessent 在美中人工智能谈判中发挥了核心作用。（Semafor）
- 其他竞争者包括 Michael Kratsios 和 Scott Kupor。（Gizmodo）
- 特朗普将人工智能更名为“超级智能”的做法会流行吗？（BBC）
Scientists are turning renewable electricity directly into food The goal is to make more food with less land. (New Scientist)
- Companies are creating food out of thin air. (MIT Technology Review) 科学家正将可再生电力直接转化为食物 目标是用更少的土地生产更多的食物。（《新科学家》）
- 各家公司正在凭空制造食物。（《麻省理工科技评论》）
A “fire amoeba” has broken the heat survival record for complex life The newly discovered organism can grow and reproduce at 63°C. (NPR)
- It could inform the search for life elsewhere. (New Scientist) 一种“火变形虫”打破了复杂生命的耐热记录 这种新发现的生物可以在 63°C 的环境下生长和繁殖。（NPR）
- 这可能为寻找地外生命提供参考。（《新科学家》）
Meta quietly tested a “human concierge” for its Muse AI agent Contractors secretly handled some calls placed by Muse. (Reuters)
- Staff raised concerns about privacy and misleading users. (404 Media) Meta 为其 Muse AI 代理秘密测试了“人工礼宾”服务 承包商秘密处理了部分由 Muse 发起的通话。（路透社）
- 员工对隐私和误导用户的问题提出了担忧。（404 Media）
Data centres are replacing copper with light to cut energy use Photonics can move data with less heat than electrical wiring. (BBC)
- Virtual power plants could also help. (MIT Technology Review) 数据中心正以光代铜以降低能耗 与电线相比，光子技术在传输数据时产生的热量更少。（BBC）
- 虚拟电厂也能提供帮助。（《麻省理工科技评论》）
AI models built from rat brains just got closer to reality AWS is tapping rat neurons to make video AI faster. (Wired) 基于大鼠大脑构建的人工智能模型离现实更近了一步 AWS 正在利用大鼠神经元来加速视频人工智能的处理速度。（《连线》）
Uber is betting on human drivers to give its robotaxis an edge They can cover demand spikes while autonomous cars recharge. (Axios) Uber 押注人类司机以提升其自动驾驶出租车的优势 在自动驾驶汽车充电时，人类司机可以覆盖需求高峰。（Axios）
A humanoid robot took on an amateur fighter in a cage The brief bout was billed as the first human-robot MMA fight. (Futurism) 一个人形机器人与一名业余拳手在笼中对决 这场短暂的比赛被称为首场人机综合格斗（MMA）比赛。（Futurism）

Quote of the day

今日名言

“The use of the word artificial makes it sound fake. It’s not fake; it’s actually amazing.” —President Donald Trump tells the UN General Assembly why he wants to rename AI “superintelligence.” “使用‘人工’这个词让它听起来很假。它并不假；它实际上非常了不起。”——唐纳德·特朗普总统在联合国大会上解释他为何想将人工智能更名为“超级智能”。

One more thing

还有一件事

COURTESY OF PAINCHEK: AI is changing how we quantify pain 图片来源：PAINCHEK：人工智能正在改变我们量化疼痛的方式

At Orchard Care Homes, nurses used to rely on an observational scale to assess pain in residents who couldn’t communicate verbally. But agitated residents were sometimes assumed to have behavioral issues, while their pain went untreated. Then, in 2021, the care-home chain began trialing PainChek, a smartphone app that scans a resident’s face for microscopic muscle movements and uses AI to output a pain score. Within weeks, the pilot unit saw fewer prescriptions and calmer corridors. Now researchers are racing to turn pain into something a camera or sensor can score as reliably as blood pressure. But when algorithms measure our suffering, does that change how we understand and treat it? Discover how AI is changing the way we assess pain. —Deena Mousa 在 Orchard 养老院，护士过去依靠观察量表来评估无法进行语言交流的住户的疼痛程度。但焦躁不安的住户有时会被误认为是行为问题，而他们的疼痛却未得到治疗。2021 年，该养老院连锁机构开始试用 PainChek，这是一款智能手机应用程序，可以通过扫描住户面部的微小肌肉运动，并利用人工智能输出疼痛评分。几周内，试点单位的处方药用量减少，走廊也变得更加安静。现在，研究人员正竞相将疼痛转化为一种摄像头或传感器可以像测量血压一样可靠地评分的指标。但是，当算法衡量我们的痛苦时，这是否会改变我们理解和治疗痛苦的方式？了解人工智能如何改变我们评估疼痛的方式。——Deena Mousa

We can still have nice things

我们依然可以拥有美好事物

A place for comfort, fun, and distraction to brighten up your day. (Got any ideas? Drop me a line.) 一个为您提供舒适、乐趣和消遣，点亮您一天的地方。（有什么想法吗？给我写信。）

Scientists have discovered the first new penguin species in more than 100 years.
科学家发现了 100 多年来首个新企鹅物种。
Discover the filmmaking skills that made Paul Verhoeven’s RoboCop a directing masterclass.
探索保罗·范霍文的《机械战警》成为导演大师课的电影制作技巧。
What happens when fast food meets local tastes? These unusual international menu items offer some tasty answers.
当快餐遇到当地口味会发生什么？这些不同寻常的国际菜单项提供了一些美味的答案。
Dance to every track the crowd has identified at Berghain since 2024 with this media player, which replays each night in sequence.
使用这个媒体播放器，随着 2024 年以来人群在 Berghain 识别出的每一首曲目起舞，它会按顺序重播每个夜晚的音乐。