字号·· | 护眼
techcrunch

这些技术故障让人想起“千年虫”,只不过这次是AI智能体风格These tech outages are giving Y2K, AI agent style

点「原文对照」整页切到原文,或双击某段只看那段的原文。

彩色插画,包含弹出窗口、电脑键盘上的Esc键、表情符号、警告图标、密码、警告和错误提示文字。Colorful illustration with pop-ups, computer keyboard escape key, emoji, warning icon, password, warning and error texts.

正如美国说唱歌手 Yung Joc 可能会说的那样:“系统崩了。”航空公司订票系统、银行平台、生产力平统统瘫痪。在过去两周里,美国头部企业接连遭遇技术故障,导致支撑日常生活运转的系统中断。

It’s going down, as the American rapper Yung Joc might say: the airline booking systems, banking platforms, and productivity platforms. In the past two weeks, the U.S.’s top businesses have faced technical outages that have disrupted the systems that run everyday life.

Apple Pay 瘫痪。NASA 的内部通信网络瘫痪。FAA 不得不让全美所有航班停飞数小时(这是自 9·11 事件以来历史上仅有的第三次),微软遭遇攻击,导致 Outlook、Teams 和 Copilot 瘫痪近两天。

Apple Pay went down. NASA’s i nternal communication network went down. The FAA had to ground all flights for hours (only the third time in history since 9/11), and Microsoft was hit with an attack that took down Outlook, Teams, and Copilot for nearly two days.

人们的第一反应自然是某种 AI 黑客入侵,尤其是在几周前网络黑客成功攻击了法国一家超市连锁的冰箱和一家血库之后。在美国,国家情报总监表示,这些故障似乎与国家资助的恐怖主义无关,而特朗普总统则称这些问题是“个别事件”。许多公司本身也表示,这些故障并非由网络攻击或 AI 攻击引起,而是技术问题。

The first thought, naturally, is an AI hacker invasion of some sort, especially after cyber hackers managed to attack fridges in a grocery store and a blood bank in France a few weeks ago. Here in the U.S., the Director of National Intelligence said these outages don’t seem related to state-sponsored terrorism, while President Trump called the issues “individual events.” Many of the companies themselves have said the outages were not caused by cyberattacks or AI attacks, but tech issues.

只是,它们都在短时间内接连发生,且就在网络安全界所谓的“黑色十月”几周之后,当时一系列同样无关的网络安全事件于 10 月 21 日袭击了 PHP、Node.js、LiteLLM 等。

Only, there’s the fact that they all struck within a short timespan of each other, and just a few weeks after Cybersecurity’s so-called “Black October,” when a slew of also-unrelated cybersecurity incidents occurred on October 21 striking PHP, Node.js, LiteLLM, more.

这确实看起来不像是巧合。坦白说,这给人一种 AI 原生版“千年虫问题”的感觉。

It does seem like more than a coincidence. It’s giving AI-native Y2K, frankly.

在网络安全领域工作了三十多年的投资人戴夫·德沃尔特(曾任麦咖啡总裁兼首席执行官,以及国土安全部 CISA 顾问委员会副主席)也认为这很可能与 AI 有关。

Dave DeWalt, an investor who worked in cybersecurity for more than three decades (including as president and CEO of McAfee and as vice Chairman of the CISA Advisory Council for the Department of Homeland Security), also thinks it’s probably AI-related.

技术故障最可能是“一个正在变得更强而非更弱的安全系统的成长烦恼”,而且“AI 正以我们从未有过的速度和规模,通过 Mythos 等 AI 工具暴露漏洞”,他对 TechCrunch 表示。

The technical outages are most likely the “growing pains of a security system that’s getting stronger, not weaker,” and that “AI is surfacing vulnerabilities at a speed and scale we’ve never had before through AI tools like Mythos and others,” he told TechCrunch.

问题是,许多技术缺陷一直都存在,但由于 AI 以极快的速度将其暴露出来,企业正争分夺秒地修补这些漏洞,以免坏人利用。“短期内,这种修补激增正在扰乱运营,”他说。

The thing is, many of these technical flaws have always existed, but because AI is surfacing them so fast, enterprises are rushing to fix them before bad actors can exploit them. “That surge in patching is disrupting things in the short term,” he said.

软件工程师、安全初创公司 Corsha 创始人 Anusha Iyer 对 TechCrunch 表示,这或许是旧有系统与现代技术首次正面碰撞的时刻。“这些工业系统从未设计为联网,更别提以我们今天推动的方式安全联网,”她说。

This was, perhaps, the first moment when legacy meets modern, Anusha Iyer told TechCrunch, a software engineer and founder of the security startup Corsha. “These industrial systems were never designed to be connected, let alone securely connected in the way we are pushing them to be today,” she said.

参议院情报委员会副主席、弗吉尼亚州参议员马克·沃纳得出了类似结论(尽管调查仍在进行,以弄清可能发生了什么)。“这些中断应敲响警钟。感觉每天都会提醒我们一次:AI 的发展速度快于我们管理其后果的能力,”该参议员在 X 上写道。

Virginia Senator Mark Warner, vice chair of the Senate Intel Committee, came to a similar conclusion (although investigations are still underway as to what could be happening). “These disruptions should serve as a wake-up call. It feels like every day brings another reminder that AI is moving faster than our ability to manage its consequences,” the Senator wrote on X .

AI 正在重塑软件的构建、部署和运维方式,规模之大令每个人都感受到陡峭的学习曲线。Iyer 引用美国联邦航空管理局(FAA)的故障为例,说明旧有技术不如企业试图与其融合的新技术那样复杂或先进。

AI is reshaping how software is built, deployed, and operated, on a scale where everyone feels the sharp learning curve. Iyer cited the FAA outage as an example of legacy tech that is not as sophisticated or advanced as the technologies companies are trying to merge it with.

用不那么专业的术语来说,她说故障可能源于(初步调查结果指向此点)FAA 在尝试现代化其电信基础设施时,过渡过程压垮了部分旧有硬件,新技术很可能以正常的现代性能运行,而旧有硬件却跟不上节奏。

In less technical terms, she said the outage may have been caused (and preliminary findings point to this) when the FAA was trying to modernize its telecommunications infrastructure and the transition overwhelmed parts of its legacy hardware, where the new tech likely operated at a normal modern capacity while the legacy hardware couldn’t keep up.

从这个角度来看,她说,AI 其实是背了黑锅。“导致这些故障的还不是 AI 的快速采用,”她说。“问题在于在缺乏对运营和网络风险的清晰认识、也缺乏守护和保护传统系统能力的情况下,仓促地将传统系统与现代系统混合在一起。”

When looked at this way, she said, AI is actually getting the bad rep. “It’s not rapid adoption of AI yet causing these outages,” she said. “The fault is the rushed mix of the legacy and modern systems without a clear understanding of operational and cyber risk or the ability to gatekeep and protect the legacy systems.”

DeWalt 表示,组织在采用 AI 时需要采取更稳健的方法,包括更多的测试,以及确定修复的优先级和时机。

Organizations need a more steadfast approach to AI adoption, including more testing and prioritizing what needs to be fixed and when, DeWalt said.

Iyer 补充道:“这些工业系统需要更严格地进行分段,并加强访问控制,以便在引入现代系统时能够分阶段实施并降低风险。”

Iyer added, “These industrial systems need to be segmented more tightly and more access controlled, so as modern systems are introduced, it can be phased in and risk mitigated.”

实际上,对于企业(以及风险投资人)来说,这一切或许有一线转机。DeWalt 表示,他宁愿看到这些美国企业以这种混乱的方式发现并修补自身缺陷,也不愿看到恶意行为者发现并利用这些漏洞发动 AI 战争,从而引发更严重的国家安全问题。

Actually, there might be a silver lining to all of this for companies (and for venture investors). DeWalt said he’d rather see these U.S. enterprises find and patch their flaws, even in this messy way, than have bad actors find and exploit those vulnerabilities in AI warfare that could cause even more national security issues.

或许也有一些有抱负的创业者因此受到启发。DeWalt 预计,专注于“AI 驱动的漏洞管理、自主修复、安全补丁测试与部署,以及维持关键基础设施运行的系统韧性”的公司将会增加。

And maybe some aspiring founders were inspired by them. DeWalt expects a rise in companies focused on “AI-driven vulnerability management, autonomous remediation, safe patch testing and deployment, and resilience for the systems that keep critical infrastructure running,” he said.

“正如我们在过去一周所见,这一过渡期确实伴随着真实的阵痛,”DeWalt 说。“但对于那些帮助组织在不破坏现有系统的前提下以 AI 速度前进的公司来说,也蕴藏着巨大的机遇。”

“There’s real pain in this transition as we’ve seen in the past week,” DeWalt said. “But there’s also enormous opportunity for the companies that help organizations move at AI speed without breaking things.”

阅读媒体原文 ↗