2026-01-29 Hacker News Top Stories

1. 微软逼我不得不转用 Linux (Microsoft forced me to switch to Linux)

https://www.himthe.dev/blog/microsoft-to-linux

作者回顾了二十多年的 Windows 使用经历，因 Windows 10/11 的非自愿更新、广告推送、隐私与驱动兼容问题，尤其 24H2 导致的 Chrome 闪烁与系统冻结，最终被迫转向基于 Arch 的 CachyOS。作者通过调整配置解决了 NVIDIA 驱动问题，并找到 Ableton Live 的替代 Bitwig Studio，认为 Linux 提供了更可控、可预测的使用体验，决定彻底放弃 Windows。

HN 热度 1543 points | 评论 1233 comments | 作者：bobsterlobster | 10 hours ago

• • Windows 11 在高性能设备上仍出现文件浏览器卡顿、右键菜单延迟等问题，很可能是企业级端点管理软件导致的性能负担。
• • 有用户指出，即使在配置极高的新笔记本上，这些问题依然存在，说明问题不在于硬件，而可能与系统优化或企业策略有关。
• • 一些评论认为，微软的 Windows 11 在功能上并未带来显著进步，例如标签式文件浏览器等特性早已在 Linux KDE Plasma 中实现多年。
• • 企业部署的 Windows 设备普遍受到端点管理软件（如安全代理、日志收集工具）影响，导致系统响应变慢，而同样的软件在 Mac 和 Linux 上对性能的影响较小。
• • 由于上述体验恶化，越来越多专业人士开始转向 macOS 或 Linux 系统，尤其在数字、网络安全和人工智能等高阶岗位中趋势明显。

2. FBI 调查明尼苏达州居民利用 Signal 群聊追踪 ICE 行动 (FBI is investigating Minnesota Signal chats tracking ICE)

https://www.nbcnews.com/tech/internet/fbi-investigating-minnesota-signal-minneapolis-group-ice-patel-kash-rcna256041

FBI 局长卡什·帕特尔表示已对明尼苏达州居民在 Signal 群聊中共享 ICE 车辆牌照、位置等信息的行为展开调查，关注是否妨碍执法或将执法人员置于危险之中。此举由右翼媒体人在社交媒体披露群组内容引发，自由言论倡导者警告调查可能触及第一修正案保护的公民监督权利，事件或对言论自由边界产生重要影响。

HN 热度 914 points | 评论 1470 comments | 作者：duxup | 1 day ago

• • FBI 只是加入群聊并阅读内容，技术上很简单，但其行为在政治背景下引发对权力滥用的担忧。
• • FBI 长期对各类政治团体进行渗透，这是常态，尤其在针对极端分子时，常通过诱导和提供资源来制造事端。
• • 有观点指出，FBI 历史上曾通过监听和施压手段干预民权领袖，如马丁·路德·金，这反映了其滥用职权的先例。
• • 一些人认为，反 ICE 组织虽有协调，但多数活动属于合法的公民监督行为，如记录执法过程、拒绝合作等，属于第一修正案保护的言论自由。
• • 反 ICE 活动中存在部分违法行为，如阻挠执法、暴力对抗、威胁和公开个人信息，这些行为已违反联邦法律。

3. OpenAI 推出全新 AI 原生科研协作平台 Prism，助力科学写作与团队协作 (Introducing Prism)

https://openai.com/index/introducing-prism

OpenAI 于 2026-01-27 推出 Prism，一款基于 GPT-5.2 的 AI 原生科研协作平台，集成云端 LaTeX 环境、文献检索、公式推导、图表处理与实时协作。平台可理解文档结构与公式，支持手绘公式转 LaTeX、与模型实时对话、无限项目与协作人数，并为 ChatGPT 个人账户用户免费开放，旨在降低科研写作门槛并加速团队协作。

HN 热度 757 points | 评论 512 comments | 作者：meetpateltech | 1 day ago

• • 选择“Prism”作为名称令人震惊，因为它关联着斯诺登揭露的 NSA 大规模监控项目，这种历史记忆对国际用户尤其是欧洲市场可能造成负面影响。
• • 尽管“Prism”在技术领域是常见词汇，但将其用于一个备受关注的 AI 产品，容易引发与 NSA 监控项目的负面联想，尤其在当前全球对数字主权和隐私权高度敏感的背景下。
• • 有人认为将“Prism”与“奥斯维辛”类比过于夸张，因为前者在日常语境中更多指向光学或通用产品名称，普通人不会立刻联想到 NSA 项目。
• • 但也有观点指出，尽管“Prism”本身中性，但 OpenAI 作为高曝光度品牌，其命名选择会受到更严格审视，哪怕大多数用户不敏感，社交媒体上仍可能被放大解读。
• • 一些人认为，OpenAI 可能有意使用该名称以示对政府机构的某种致敬或拉近关系，尽管这缺乏证据，但其数据处理行为本身已引发隐私担忧。

4. 自特朗普上任以来，美国政府已流失超 1 万份 STEM 博士学位人才 (U.S. government has lost more than 10k STEM PhDs since Trump took office)

https://www.science.org/content/article/u-s-government-has-lost-more-10-000-stem-ph-d-s-trump-took-office

文章报道自特朗普上任以来，美国政府在职或受雇的 STEM 博士已流失逾一万人。主要原因包括科研经费不确定与削减提案、签证与移民政策收紧、行政不稳定及对外国学生的驱逐或骚扰，导致项目中断、资助延迟与国际合作减少。结果是科研生态被削弱，欧洲合作空缺被中国迅速填补，短期内重建国际信任与人才吸引力面临长期挑战。

HN 热度 560 points | 评论 411 comments | 作者：j_maffe | 1 day ago

• • 美国政府对 STEM 博士的吸引力下降，主要由于科研经费削减、外国学生政策收紧以及行政不确定性导致人才流失。
• • 虽然 NSF 预算 55% 的削减提案尚未通过，但其提出已造成实际影响，导致科研项目申请停滞、资助延迟或取消。
• • 美国政府通过制造不确定性、取消资助、驱逐或骚扰外国学生等方式，间接削弱了科研生态，无需真正削减预算即可造成严重后果。
• • 欧洲科研人员指出，美国科研经费削减导致与欧盟合作项目减少，形成科研“真空”，已被中国迅速填补，中欧科研合作显著增加。
• • 中国凭借充足资金、基础设施和人才优势，正成为全球科研合作的新中心，美国的反智政策反而助推了中国科技发展.

5. 亚马逊宣布裁员 1.6 万人 (Amazon cuts 16k jobs)

https://www.reuters.com/legal/litigation/amazon-cuts-16000-jobs-globally-broader-restructuring-2026-01-28/

亚马逊宣布全球裁员约1.6万人，这是自2022年10月以来第二轮大规模裁员，累计接近3万人，主要针对总部及非仓储白领岗位，约占公司企业级员工近10%。公司称目标是减少官僚、提高决策效率并推动AI应用，部分内部邮件误发（“Project Dawn”）引发恐慌；受影响部门包括AWS、Alexa、Prime Video、广告、物流、Kindle与供应链等。与此同时，亚马逊将关闭Fresh与Go门店并放弃Amazon One，显示其正收缩线下零售，转向自动化与核心电商技术投入。

HN 热度 482 points | 评论 665 comments | 作者：DGAP | 9 hours ago

• • 一名亚马逊中层管理者开发了 AI 工具来替代自己原本的工作，结果发现自己成了被裁的首批目标，这一经历令人感到讽刺和悲哀。
• • 有评论质疑该故事的真实性，认为真正有价值的管理工作远不止信息收集与汇报，因此该管理者可能在夸大其词以显得自己重要。
• • 大量中层管理岗位的工作内容流于形式，缺乏实质价值，这类“虚假管理”在企业中普遍存在。
• • 历史上技术进步始终在替代人力，从工业革命到计算机时代，如今 AI 正继续这一趋势，最终将影响软件开发等高技能岗位。
• • 社会应思考如何通过再培训、基本收入（UBI）或公共服务等方式支持被替代的劳动者。

6. 美国移民局与帕兰提尔合作：利用健康数据追踪非法移民 (ICE and Palantir: US agents using health data to hunt illegal immigrants)

https://www.bmj.com/content/392/bmj.s168

美国移民与海关执法局（ICE）使用帕兰提尔开发的“Elite”应用，整合美国卫生与公共服务部（HHS）提供的数百万医疗记录（包括姓名、地址、照片及“在家中可能性”评分）以定位并突袭疑似非法移民住所。该工具已用于2025年俄勒冈突袭并抓捕多人，引发医学界与隐私专家对侵蚀医疗信任、公共卫生成本及法律与伦理问题的强烈担忧。部分州已提起司法挑战并暂停数据共享，专家呼吁国会修正相关法律以维护公众信任。

HN 热度 423 points | 评论 251 comments | 作者：dberhane | 12 hours ago

• • 私人监控比政府监控更可怕，因为企业有动机开发新监控手段并出售给政府或影响世界的私人实体，导致个人数据可能被任何应用或公司利用。
• • 政府可以利用企业收集的数据进行监控，同时还能推卸责任，声称“我们没有收集数据”，从而逃避监管。
• • 医疗数据如 Medicaid 数据由政府通过官方表格收集，因此政府对这些数据的使用是明确且不可否认的，不应被模糊化为“私人监控”问题。
• • 企业出于利润最大化，会向最高出价者出售数据，这包括外国政府或敌对势力，存在数据泄露和被滥用的风险。
• • ICE 在执行逮捕时经常不获取搜查令，强行破门进入民宅，侵犯了公民的第四修正案权利，有法官因此下令释放被非法逮捕者。

7. 窃取 9000 万美元美国政府控制加密货币的黑客竟是政府承包商之子 (Thief of $90M in seized U.S.-controlled crypto is gov't contractor's son)

https://www.web3isgoinggreat.com/single/lick-theft

2026年1月，一名网名“Lick”的黑客被指窃取了约9000万美元美国政府没收的加密货币。因在屏幕共享炫耀钱包余额而暴露地址，被追踪者zachxbt关联到2024年10月多起盗窃案，并指认其为政府承包商创始人之子，涉事公司正负责管理被没收的加密资产，事件引发对透明度与利益冲突的强烈质疑。

HN 热度 385 points | 评论 83 comments | 作者：pavel_lishin | 1 day ago

• • 一名政府承包商之子涉嫌窃取价值 9000 万美元的美国政府没收的加密资产，因在朋友面前炫耀而暴露行踪。
• • 该嫌疑人通过屏幕共享展示钱包余额以证明财富，被加密货币追踪者迅速锁定。
• • 有评论指出，此类事件反映出政府在承包商管理与网络安全方面存在严重漏洞。
• • 有观点批评美国政府长期依赖关系户承包项目，导致腐败和管理不善问题频发。
• • 有评论讽刺美国政府在处理加密资产时的低效和荒诞，如将巨额资产存放在不安全的系统中。

8. Lennart Poettering 与 Christian Brauner 共同创立新公司 (Lennart Poettering, Christian Brauner founded a new company)

https://amutable.com/about

Amutable 是由 systemd 创始人 Lennart Poettering 与 Linux 内核开发者 Christian Brauner 等人创立的公司，旨在将密码学可验证的完整性集成到 Linux 中，保障构建、启动与运行时三个阶段的完整性，提供从开机到运行的可验证信任基础设施，团队由多位内核与系统安全专家组成。

HN 热度 367 points | 评论 682 comments | 作者：hornedhob | 1 day ago

• • 该公司的成立令人悲观，因其创始人背景和以利润为导向的动机，对用户自主权构成威胁。
• • 远程证明技术可能被滥用，导致用户失去对自己设备的控制权，进而被服务提供商过度管控。
• • 银行等机构若要求使用安全锁定终端，应自行提供设备，而非强制用户改造个人设备。
• • 当前已有 ATM 等物理安全终端，无需通过限制用户设备来保障交易安全。
• • 硬件安全密钥（如 YubiKey）已能有效防范钓鱼和中间人攻击，无需依赖远程证明。

9. 翼型（2024） (Airfoil (2024))

https://ciechanow.ski/airfoil/

本文以翼型为切入点，解释翼型如何影响气流与升力，介绍多种二维可视化方法（速度箭头、追踪粒子、亮度表示速度），并通过稳定流示例说明气流速度与位置关系，奠定后续讨论翼型如何产生升力的基础。

HN 评论 48 comments | 作者：brk | 10 hours ago

• • Airfoil 设计的核心在于提升升力与阻力的比率，而非单纯生成升力，同时涉及失速速度、跨音速性能、层流/湍流流动等多种复杂因素。
• • AeroSandbox 是一个强大的开源工具，可用于空气动力学模拟与优化，内置神经网络模型可快速准确预测翼型气动特性。
• • 空气动力学中升力的本质是飞机通过向下推动空气来实现动量交换，这是理解飞行原理的根本出发点。
• • 翼型并非产生升力的唯一关键，大部分升力来自机翼整体形状，翼型主要优化升阻比，其重要性类似于翼尖小翼。
• • 伯努利原理和迎角理论并非对立，而是同一现象的不同解释视角，实际应用中需结合使用。

10. ASML 人员调整预计导致约 1700 个岗位减少 (Strengthening focus on engineering and innovation)

https://www.asml.com/en/news/press-releases/2026/strengthening-focus-on-engineering-and-innovation

ASML 于 2026-01-28 宣布重组技术与 IT 组织，从项目制矩阵模式转向以产品/模块为核心的专职团队，并保留核心基础技术团队以维持技术深度。公司称此举为提升决策效率与创新文化的战略性调整，预计净减少约 1700 个岗位，主要集中在荷兰并涉及部分美国岗位，受影响多为管理层，同时将创建新的工程岗位并尽量安排内部转岗。制造、客户支持与销售不受此次改革影响，公司承诺与荷兰社会伙伴协商并以透明、公平的方式推进。

HN 热度 327 points | 评论 333 comments | 作者：dep_b | 17 hours ago

• • ASML 计划裁员约 1700 个管理岗位，显示出其试图纠正组织过度管理层化的趋势，与曾经工程师文化浓厚的风格形成对比，此举被视为有勇气的改革。
• • 一些评论指出，公司从原本精益的管理结构逐渐演变为管理臃肿，这在吸收破产企业人员后尤为明显，导致效率下降、创新停滞。
• • 德国大型企业如西门子及其子公司在被分拆后，往往因管理僵化、缺乏战略愿景而走向衰落，即便拥有行业壁垒也难以避免。
• • 德国企业普遍存在过度依赖认证和官僚流程的问题，例如强制要求“认证代码审查员”、频繁的 OKR 追踪等，反而阻碍了软件开发效率。
• • 德国职场中的“工作证明”（Zeugnisse）制度对个人职业发展影响深远，一旦获得负面评价可能长期影响就业机会，且评估标准不透明，存在主观偏见风险。

完整摘要请点击 “阅读原文” 👇👇👇

Hacker News 精彩评论及翻译

Microsoft forced me to switch to Linux

https://news.ycombinator.com/item?id=46798223

My story is simpler. Microsoft dropped the support for Windows 10 and gave me no upgrade path to Windows 11 because my CPU was 5 years too old apparently.

So I installed Fedora on that machine, I learned the process, I went through the hurdles. It wasn’t seamless. But, Fedora never said “I can’t”. When it was over, it was fine.

Only if Microsoft had just let me install Windows 11 and suffer whatever the perf problem my CPU would bring. Then I could consider a hardware upgrade then, maybe.

But, “you can’t install unless you upgrade your CPU” forced me to adopt Linux. More importantly, it gave me a story to tell.

There is a marketing lesson there somewhere, like Torvalds’ famous “you don’t break userspace”, something along the lines of “you don’t break the upgrade path”.

sedatk

我的故事更简单一些。微软停止了对 Windows 10 的支持，并且由于我的 CPU 明显老了五年，没有给我任何升级到 Windows 11 的路径。

于是，我在那台电脑上安装了 Fedora，我学习了整个过程，也经历了一些波折。过程并不顺利。但是，Fedora 从未说过“我不行”。当一切结束后，情况就很好了。

要是当初微软只是让我安装 Windows 11，然后去忍受我的 CPU 带来的任何性能问题就好了。那样的话，我或许就能考虑升级硬件了。

但是，“除非升级你的 CPU，否则你不能安装”这句话，迫使我转向了 Linux。更重要的是，它给了我一个可以讲述的故事。

这其中蕴含着一个营销的教训，有点类似于托瓦兹（Torvalds）那句著名的“你不要破坏用户空间”，或许可以概括为“你不要破坏升级路径”。

Microsoft forced me to switch to Linux

https://news.ycombinator.com/item?id=46797454

I just started a new job where I'm subjected to Windows 11. They gave me a behemoth of a laptop. 64GB of RAM, absolute screamer of a CPU, big GPU, the whole deal.

Windows 11's file browser lags when opening directories with more than 100-ish files. Windows 11's file browser takes a few seconds to open at all.

Context menus take a noticeable amount of time to appear.

I'm getting used to a new keyboard, so I keep hitting Print Screen by accident. Half the time I can smack Esc and Snipping Tool will go away. The other half of the time, I have to mouse over and click the X to close it. There is no pattern to when Esc does/doesn't work.

If my computer goes to sleep, WSL becomes unresponsive. I have to save all my stuff and reboot to continue working.

If Windows 11 struggles this badly on a brand new laptop that I'm certain would retail for $4000+, I can only imagine how miserable it is for everyone else. All my colleagues who have been here for a bit longer got last-generation laptops. oof.

Edit... and besides, what does Windows 11 even do that KDE Plasma 5 wasn't doing a decade ago? How did it take this long to get a tabbed file browser?

ryukoposting

我刚换了份新工作，被迫使用 Windows 11。他们给了我一台巨无霸笔记本，64GB 内存，顶级的 CPU，强大的 GPU，配置应有尽有。

Windows 11 的文件浏览器在打开包含一百来个文件的目录时会卡顿，而且它自己打开时都要花费好几秒钟。

右键菜单出现时需要相当长的时间，让人能明显感觉到延迟。

我还在适应新键盘，所以总是不小心误触到打印屏幕（PrtSc）键。有一半的时候，我按一下 Esc 键，截图工具就会关闭；另一半时间，我则必须把鼠标移过去，点击那个 X 按钮才能关闭它。按下 Esc 键有时管用有时不管用，毫无规律可言。

如果我的电脑进入睡眠模式，WSL 就会无响应，我必须保存好所有东西，然后重启电脑才能继续工作。

如果 Windows 11 在我敢说售价肯定超过4000美元的全新笔记本上都表现得这么糟糕，我真不敢想象对于其他人来说用起来会有多痛苦。而那些比我早来的同事们，分到的都还是上一代的笔记本。真是够呛。

补充一点……话说回来，Windows 11 到底有什么功能是 KDE Plasma 十年前就做不到的？怎么会花了这么长时间才给文件浏览器加上标签页功能呢？

A few random notes from Claude coding quite a bit ...

https://news.ycombinator.com/item?id=46784594

I worry about the "brain atrophy" part, as I've felt this too. And not just atrophy, but even moreso I think it's evolving into "complacency".

Like there have been multiple times now where I wanted the code to look a certain way, but it kept pulling back to the way it wanted to do things. Like if I had stated certain design goals recently it would adhere to them, but after a few iterations it would forget again and go back to its original approach, or mix the two, or whatever. Eventually it was easier just to quit fighting it and let it do things the way it wanted.

What I've seen is that after the initial dopamine rush of being able to do things that would have taken much longer manually, a few iterations of this kind of interaction has slowly led to a disillusionment of the whole project, as AI keeps pushing it in a direction I didn't want.

I think this is especially true if you're trying to experiment with new approaches to things. LLMs are, by definition, biased by what was in their training data. You can shock them out of it momentarily, whish is awesome for a few rounds, but over time the gravitational pull of what's already in their latent space becomes inescapable. (I picture it as working like a giant Sierpinski triangle).

I want to say the end result is very akin to doom scrolling. Doom tabbing? It's like, yeah I could be more creative with just a tad more effort, but the AI is already running and the bar to seeing what the AI will do next is so low, so....

daxfohl

我担心“大脑萎缩”这一点，因为我也有过同样的感受。而且不仅仅是萎缩，我甚至认为它正在演变成“自满”。

比如说，有好几次我想要代码看起来是某种样子，但它总会拉回到它自己的处理方式上。就好像如果我最近提出了某些设计目标，它会遵循，但经过几次迭代后，它又会忘记，回到原来的方法，或者将两者混合，等等。最后，放弃斗争，让它按自己的方式来做要容易得多。

我所看到的是，在最初能快速完成手动需要更长时间工作的多巴胺激增之后，这种几次反复的互动慢慢导致了对整个项目的幻灭，因为人工智能一直在推动它朝着我并不希望的方向发展。

我认为，如果你正在尝试用新的方法来做事情，这一点尤其明显。大语言模型从定义上就受到其训练数据内容的偏见。你可以暂时让它跳出这种模式，这在几个回合里很棒，但随着时间的推移，其潜在空间中已存在内容的引力变得无法抗拒。（我把它想象成像一个巨大的谢尔宾斯基三角形在工作）。

我想说，最终的结果很像“末日滚动”（doom scrolling）。或者叫“末日标签页”（doom tabbing）？就好像是啊，我可以再多花一点努力来更富创造力，但AI已经在运行，看到AI下一步会做什么的门槛太低了，所以……

FBI is investigating Minnesota Signal chats tracki...

https://news.ycombinator.com/item?id=46790186

There seems to be wild speculation about freedom of speech rights or hacking Signal.

The FBI simply joined groupchats and read them. This is trivial stuff.

mw888

似乎有人在捕风捉影，谈论言论自由权或入侵Signal。FBI不过是加入了群聊并阅读内容，这根本就是小事一桩。

Prism

https://news.ycombinator.com/item?id=46792035

I'm dumbfounded they chose the name of the infamous NSA mass surveillance program revealed by Snowden in 2013. And even more so that there is just one other comment among 320 pointing this out [1]. Has the technical and scientific community in the US already forgotten this huge breach of trust? This is especially jarring at a time where the US is burning its political good-will at unprecedented rate (at least unprecedented during the life-times of most of us) and talking about digital sovereignty has become mainstream in Europe. As a company trying to promote a product, I would stay as far away from that memory as possible, at least if you care about international markets.

[1] https://news.ycombinator.com/item?id=46787165

Perseids

我惊呆了他们竟然选择了斯诺登在2013年曝光的那个臭名昭著的NSA大规模监视项目的名字。更让我惊讶的是，在320条评论中，只有另一条评论指出了这一点[1]。美国的科学技术界已经忘记这场巨大的信任背叛了吗？在美国正以前所未有的速度消耗其政治善意（至少在我们大多数人有生之年是前所未有的），且“数字主权”在欧洲已成为主流的当下，这一点尤其令人感到格格不入。作为一个试图推广产品的公司，我一定会尽可能地远离那段历史记忆，至少如果你还在意国际市场的话。

Amazon cuts 16k jobs

https://news.ycombinator.com/item?id=46797370

Amazon axes 16,000 American jobs as it ... relocates to a larger campus in India

https://techcrunch.com/2025/12/10/amazon-to-invest-additional-35b-in-india-by-2030-taking-total-planned-spending-to-75b/

focusgroup0

亚马逊将裁员16,000名美国员工，并……迁往印度一处更大的园区。

https://techcrunch.com/2025/12/10/amazon-to-invest-additional-35b-in-india-by-2030-taking-total-planned-spending-to-75b/

A few random notes from Claude coding quite a bit ...

https://news.ycombinator.com/item?id=46784407

It's so interesting to watch an agent relentlessly work at something. They never get tired, they never get demoralized, they just keep going and trying things where a person would have given up long ago to fight another day. It's a "feel the AGI" moment to watch it struggle with something for a long time just to come out victorious 30 minutes later.

Somewhere, there are GPUs/NPUs running hot. You send all the necessary data, including information that you would never otherwise share. And you most likely do not pay the actual costs. It might become cheaper or it might not, because reasoning is a sticking plaster on the accuracy problem. You and your business become dependent on this major gatekeeper. It may seem like a good trade-off today. However, the personal, professional, political and societal issues will become increasingly difficult to overlook.

einrealist

看着一个智能体不知疲倦地执着于某件事，真是非常有趣。它们从不疲倦，也从不气馁，只是持续不断地尝试、坚持，在人类早已放弃、准备另寻他法的地方继续战斗。看着它长时间地与某个问题搏斗，并在半小时后最终取得胜利，这真是一种“感受到通用人工智能（AGI）”的时刻。

在某个地方，GPU/NPU正在高负荷运转。你提供了所有必要的数据，其中甚至包括你绝不会与他人分享的信息。而且，你很可能并未支付其真实成本。这项技术的成本未来可能会降低，也可能不会，因为推理能力只是为弥补准确率不足而打上的“创可贴”。你和你的业务将因此对这个主要的把关者产生依赖。今天看来，这似乎是一笔划算的交易。然而，其中牵涉的个人、职业、政治及社会问题将变得越来越难以忽视。

Amazon cuts 16k jobs

https://news.ycombinator.com/item?id=46797061

Every large company is updating its standard layoffs announcement press release from "economic headwinds" to "AI".

paxys

各家大公司都在将标准的裁员公告新闻稿措辞，从“经济逆风”更新为“人工智能”。

ICE and Palantir: US agents using health data to h...

https://news.ycombinator.com/item?id=46794654

Private surveillance is so much more scary than regular government surveillance because they have every incentive to invent new ways of surveilling you that they then try to sell to governments, or private actors who want to influence the world. It's like classic government surveillance but every company you interacted with and every app you use may at some point turn on you and use your data against you, just because someone realized "hey, I bet we can sell this data"

We are really seeing the fears of data collection from the 2000s and 2010s come to fruition as privatized surveillance now. Cambridge analytica should have been the warning shot but it wasn't enough.

petterroea

私人监控比常规的政府监控要可怕得多，因为他们有充分的动力去发明新的监控你的方法，然后再试图将其出售给政府或那些想要影响世界的私人行为者。这就像是经典的政府监控，但你与之互动的每一家公司和使用的每一个应用，都可能随时反戈一击，利用你的数据对付你，仅仅因为某人意识到“嘿，我打赌我们可以靠这些数据赚钱”。如今，随着监控的私有化，我们确实看到了2000年代和2010年代人们对数据收集的担忧成为了现实。剑桥分析公司（Cambridge Analytica）本应是一个警钟，但那还远远不够。

U.S. government has lost more than 10k STEM PhDs s...

https://news.ycombinator.com/item?id=46785403

It's harder to recruit PhD students and it's harder to fund them. NSF budget was cut 55% in the first year. The administration is doing everything possible to make it clear that no foreigners are welcome here. America is stabbing itself directly in the brain.

titzer

招博士生更难了，资助他们也更难了。国家科学基金会（NSF）的预算第一年就被削减了55%。政府正在尽一切努力表明，这里不受外国人欢迎。美国这是在直接刺向自己的大脑。

Microsoft forced me to switch to Linux

https://news.ycombinator.com/item?id=46796497

This was me in 2022 or 2023. I have posted on HN about my shift a few times. I gave up with Windows 10 because you needed Windows Pro in order to make an "offline" account, I spent $2000+ for a gaming rig, and I couldn't add new users, one program told me to use the other program which brought me back to the original program... I had to go out of my way, buy a license just to make it work. I just went and installed Linux finally. I was on POP_OS! for a good year, but been on Arch Linux for one year plus now.

I know its a "meme" to talk about how great Arch is, but when you want the latest of something, Arch has it. I use EndeavourOS since it had a nicer simpler installer (idk why Arch doesn't invest in whats standard in every other major distro) and if you just use "yay" you don't run into Pacman woes.

Alternatively, I'm only buying Macs as well, but for my gaming rigs, straight to Arch. Steam and Proton work perfectly, if you don't sell your games on Steam or in a way I can run them on Linux I am not buying or playing them.

giancarlostoro

这就是我在2022或2023年的样子。我曾在Hacker News上发帖几次，提及我的转变。我放弃了Windows 10，因为要创建一个“离线”账户需要Windows Pro版。我花了两千多美元组装了一台游戏主机，却连添加新用户都做不到，一个程序让我去用另一个程序，结果又绕回了原来的程序……我不得不费尽周折，花钱买了个许可证才能让它正常工作。最后，我干脆直接装了Linux。我用POP_OS!挺长时间了，但现在已经在用Arch Linux一年多了。

我知道把Arch吹得天花乱坠是个“梗”，但当你想用上最新鲜的东西时，Arch就有。我用的是EndeavourOS，因为它有更漂亮、更简单的安装程序（我不知道为什么Arch不在其他主流发行版都标配的功能上投资）。而且，只要你用“yay”来管理软件包，就不会遇到Pacman的麻烦。

另外，我也只买Mac，但对于我的游戏主机，我直接上Arch。Steam和Proton配合得天衣无缝。如果你不在Steam上卖游戏，或者没法让我在Linux上运行，那我既不会买，也不会玩。

ASML staffing changes could result in a net reduct...

https://news.ycombinator.com/item?id=46792672

The press release ( https://www.asml.com/en/news/press-releases/2026/strengthening-focus-on-engineering-and-innovation ) seems remarkably to the point, for CEO press release standards.

I'm impressed by their ambition to fire 1700 managers (!) That's a lot of managers! I interviewed with ASML a decade and a half ago and while there was plenty to complain about (eg their tens of millions of lines of absolutely unmaintainable C code), I didn't feel at the time feel like it was a very top-heavy organization. It was very engineer-y, and I loved that about them. This press release (when taken at face value) suggests that this has changed a lot over time and they're now trying to correct it.

I gotta say, if true and not code for general "cheese slicer" cost cutting, I think that this is rather ballsy. Philips (which ASML spun out of) famously never did anything of the sort and gradually cramped into an extremely management-heavy organization where most people just write reports for other people with scary few people actually moving the needle. I think it's cool that ASML has identified that they're risking becoming like Philips and trying to do something about it, even if the method seems rather crude. I think the risk is real. ASML's fast-moving culture formed in a mad multi-decade survival-crunch, but they've been a near-monopolist for a while now and that means those pressures are long gone.

skrebbel

这篇新闻稿（https://www.asml.com/en/news/press-releases/2026/strengthening-focus-on-engineering-and-innovation）就CEO级别的新闻稿而言，可谓切中要害。

ASML打算解雇1700名经理人的野心着实让我印象深刻！那可是相当多的管理者！十五年前我面试过ASML，当时公司虽然问题不少（比如说，他们有数千万行完全无法维护的C代码），但我并不觉得它是一家头重脚轻的组织。它充满了工程师文化，我非常喜欢他们这一点。这篇新闻稿（如果按字面意思理解）表明，情况在过去的岁月里已发生巨大变化，他们现在正试图纠正这一点。

我得说，如果这是真的，而不仅仅是“大刀阔斧”式削减成本的代名词，我认为这相当大胆。飞利浦（ASML就是从其中分拆出来的）就以从不采取此类行动而闻名，并逐渐演变成一个管理层极为臃肿的组织，在那里大多数人只是在为他人写报告，而真正能推动业务进展的人却少得可怜。我觉得很酷的是，ASML已经意识到了自己有变成飞利浦那样的风险，并试图采取行动，即便这种方法看起来相当粗糙。我认为这个风险是真实存在的。ASML那种快节奏的文化是在过去几十年的疯狂求生压力下形成的，但他们如今已接近垄断地位，这意味着那些压力早已不复存在了。

Amazon closing its Fresh and Go stores

https://news.ycombinator.com/item?id=46782592

Doesn’t surprise me. I frequently shop at Amazon Fresh in store and it’s a mediocre experience. It’s a poorly run store with no visible manager making sure things are in order. You constantly have to work around employees fulfilling online orders and they aren’t helpful. I always find expired groceries/produce on the shelf so I have to spend a lot of extra time inspecting each item. The only reason I put up with their nonsense is that some of their prices are insane and they have easy returns, for example $0.85 for a box of Barilla pasta. They actually don’t accept returns in store and just refund you automatically in the app (Returnless returns). It’s pretty silly and rife for abuse.

I also found a loophole with the Amazon.com return grocery credit. The systems are separate for the 40 coupon and you just scan a QR code in the store to get it. It turns out you can just take a photo of their QR code and reuse it over and over again.

Bluecobra

这并不让我意外。我经常在亚马逊生鲜超市线下购物，体验很一般。这家商店管理不善，没有可见的经理来确保一切井然有序。你不得不经常为完成线上订单的员工让路，而且他们一点都不帮忙。我总是在货架上找到过期的食品/农产品，所以我必须花大量额外时间检查每一件商品。我忍受他们这些胡闹的唯一原因是，他们的一些价格低得离谱，而且退货方便，比如说一盒巴里拉意大利面只要0.85美元。他们实际上不接受店内退货，只是会在应用里自动给你退款（无需退货的退款）。这相当愚蠢，而且很容易被滥用。

我还发现了亚马逊.com退货食品积分的一个漏洞。那个满40减10美元的优惠券和这个生鲜的系统是分开的，你只需要在店里扫描一个二维码就能拿到。结果发现，你完全可以拍下他们的二维码照片，然后一次又一次地重复使用。

TikTok users can't upload anti-ICE videos. The com...

https://news.ycombinator.com/item?id=46780791

Allow me to offer some words of wisdom. If you help building weapons to be used against $currently_designated_bad_people, you can rest assured that given enough time, those weapons will be used against you. I am watching all this with a mild sense of bemusement.

iugtmkbdfil834

容我分享几句肺腑之言。倘若你协助制造用以打击“当前 designated 的坏人”的武器，大可放心，假以时日，那些武器终将用于对付你本人。我正以一丝淡然的困惑，静观这一切。

Microsoft forced me to switch to Linux

https://news.ycombinator.com/item?id=46796585

if you don't sell your games on Steam or in a way I can run them on Linux I am not buying or playing them.

So much this. People like to moan about "oh game XYZ doesn't run so it's not reasonable for gaming". More games run on GNU / Linux than any gaming console. There are simply too many games that do run to give a second thought about the ones that don't, and it's been that way for years.

Zambyte

如果你不在Steam上销售游戏，或者我没有办法在Linux上运行，那我就不会购买或游玩。

完全同意。人们总喜欢抱怨“哦，游戏XYZ无法运行，所以用Linux玩游戏不现实”。能在GNU/Linux上运行的 games 比任何游戏主机都多。能玩的游戏实在太多了，根本没工夫去纠结那些不能玩的，而且这种情况已经持续好几年了。

A few random notes from Claude coding quite a bit ...

https://news.ycombinator.com/item?id=46774862

LLM coding will split up engineers based on those who primarily liked coding and those who primarily liked building.

I’ve always said I’m a builder even though I’ve also enjoyed programming (but for an outcome, never for the sake of the code)

This perfectly sums up what I’ve been observing between people like me (builders) who are ecstatic about this new world and programmers who talk about the craft of programming, sometimes butting heads.

One viewpoint isn’t necessarily more valid, just a difference of wiring.

atonse

大语言模型（LLM）编程将把工程师分为两类：一类是主要喜欢编程的人，另一类是主要喜欢构建的人。

我一直都说自己是个构建者，尽管我也享受编程（但享受的是编程带来的结果，而非代码本身）。

这完美地总结了我一直在观察的现象：像我这样的构建者（builders）对这个世界感到无比兴奋，而程序员们则津津乐道于编程的技艺，有时甚至会因此产生冲突。

这两种观点没有谁对谁错，只是思维方式的不同罢了。

Prism

https://news.ycombinator.com/item?id=46787976

GenAI largely seems like a DDoS on free resources. The effort to review this stuff is now massively more than the effort to "create" it, so really what is the point of even submitting it, the reviewer could have generated it themself. Seeing it in software development where coworkers are submitting massive PRs they generated but hardly read or tested. Shifting the real work to the PR review.

I'm not sure what the final state would be here but it seems we are going to find it increasingly difficult to find any real factual information on the internet going forward. Particularly as AI starts ingesting it's own generated fake content.

SchemaLoad

生成式人工智能在很大程度上似乎是对免费资源的一种DDoS攻击。如今，审核这些内容的努力远远超过了“创造”它们的努力，所以，提交这些内容到底有什么意义呢？审核者本可以自己生成这些内容。我在软件开发中也看到了这种情况：同事们提交的是他们用AI生成的巨大PR，但他们自己几乎不读也不测试，从而把真正的工作转移到了PR审核上。

我不确定最终的结局会是什么，但看起来我们将来会越来越难以在互联网上找到任何真实的、基于事实的信息。尤其当人工智能开始吞噬自己生成的虚假内容时，情况更是如此。

FBI is investigating Minnesota Signal chats tracki...

https://news.ycombinator.com/item?id=46786044

With all the predatory tech Palantir has produced, it won't take more than a few minutes for FBI to start taking actions, IF they had anything tangible.

This is just an intimidation tactic to stop people talking (chatting)

hedayet

鉴于Palantir研发了这么多掠夺性的技术，如果FBI手里真有什么实锤的东西，他们根本用不了几分钟就会采取行动。这不过是一种恐吓手段，旨在让人们闭嘴。

I let ChatGPT analyze a decade of my Apple Watch d...

https://news.ycombinator.com/item?id=46773944

Health metrics are absolutely tarnished by a lack of proper context. Unsurprisingly, it turns out that you can't reliably take a concept as broad as health and reduce it to a number. We see the same arguments over and over with body fat percentages, vo2 max estimates, BMI, lactate thresholds, resting heart rate, HRV, and more. These are all useful metrics, but it's important to consider them in the proper context that each of them deserve.

This article gave an LLM a bunch of health metrics and then asked it to reduce it to a single score, didn't tell us any of the actual metric values, and then compared that to a doctor's opinion. Why anyone would expect these to align is beyond my understanding.

The most obvious thing that jumps out to me is that I've noticed doctors generally, for better or worse, consider "health" much differently than the fitness community does. It's different toolsets and different goals. If this person's VO2 max estimate was under 30, that's objectively a poor VO2 max by most standards, and an LLM trained on the internet's entire repository of fitness discussion is likely going to give this person a bad score in terms of cardio fitness. But a doctor who sees a person come in who isn't complaining about anything in particular, moves around fine, doesn't have risk factors like age or family history, and has good metrics on a blood test is probably going to say they're in fine cardio health regardless of what their wearable says.

I'd go so far to say this is probably the case for most people. Your average person is in really poor fitness-shape but just fine health-shape.

chrisfosterelli

健康指标因缺乏适当的背景而完全失真。不出所料，事实证明，你无法可靠地将“健康”这样一个宽泛的概念简化成一个数字。我们在体脂率、最大摄氧量估算、身体质量指数（BMI）、乳酸阈值、静息心率和心率变异性（HRV）等指标上，一遍又一遍地看到同样的争论。这些都是有用的指标，但重要的是，要根据它们各自应有的适当背景来考量。

这篇文章给一个大语言模型（LLM）提供了一堆健康指标，然后让它简化成一个单一的分数，却没有告诉我们任何实际的指标数值，接着再将这个分数与医生的诊断进行比较。我真无法理解为什么会有人期望这两者能达成一致。

我一眼就能看出的最明显的一点是，我发现医生和健身圈对“健康”的看法，无论好坏，通常大相径庭。他们使用的工具不同，目标也不同。如果这个人的最大摄氧量估算值低于30，那么根据大多数标准，这客观上是一个很差的数值，而且一个在全网健身讨论数据库上训练过的大语言模型，很可能会在心肺健康方面给这个人打一个很低的分数。但一位医生，当他看到一个人来看诊，这个人没有特别抱怨什么，活动自如，没有年龄或家族病史等风险因素，并且血液检查指标良好时，他可能会说这个人的心肺健康状况良好，不管他的可穿戴设备显示的是什么数据。

我甚至可以说，对大多数人来说，情况可能都是如此。你身边的普通人，可能身体状况（fitness-shape）非常糟糕，但健康状况（health-shape）却很好。