RU

Microsoft, Apple and US bank outages: Rubio sees no attack, experts blame AI

Published: 2026-11-18 · Author: AI Release · @ai_releaseorg
Microsoft, Apple and US bank outages: Rubio sees no attack, experts blame AI

⚡ The gist

🔍 What happened

On November 18, 2026, Marco Rubio was talking to reporters at the Solaire hotel in the Philippines — but instead of diplomacy, he was fielding questions about a string of IT outages in the US. Over the past few days, the failures hit Microsoft, major American banks, Apple, air traffic control systems, and the National Security Agency. Users lost access to cloud services and bank accounts, flights were canceled, and some closed communication channels went dark. According to Rubio, none of it is "interconnected," and the country is "not under systemic attack": "We have zero evidence or reason to believe that these isolated incidents are connected to bad actors, hackers, criminals, or another state."

Almost simultaneously, researchers at the American company Calif — using OpenAI technology — discovered a vulnerability in Java, which runs 90% of Fortune 500 companies. The flaw allows remote code execution and lets an attacker plant malware in a system. UCL professor Steven Murdoch poured cold water on the panic: the problem isn't with all Java programs, but with the Base64 library that ships with one version of Java. "This is significant, but not all Java programs are affected," he said. The UK's National Cyber Security Centre, along with authorities in the US and South Korea, urged users to install updates — The Guardian writes in a report by Dan Milmo.

🧠 Our take

Two people have put forward a theory about a common source of the outages. US Director of National Intelligence Jay Clayton said on Fox Business that the cause is routine security updates that are taking longer and springing surprises: systems aren't coming back online as expected, and handoffs between systems are breaking. Cybersecurity expert Tal Be'ery doesn't believe the incidents are directly connected, but sees a common thread — AI: the infrastructure is "fragile," and the pressure has spiked sharply. Our assessment: both theories are compatible, and that's the most unsettling part of this story. It's not hackers breaking production — production is being broken by its own repairs.

The numbers in the source explain why this is happening now. Since April, when Anthropic released the Mythos model with an unexpectedly strong knack for finding IT vulnerabilities, the number of vulnerabilities discovered at major tech companies has grown eightfold. OpenAI has also released models that surface previously unknown defects. Which gives you a cycle: AI finds more holes → companies need to ship more patches, faster → the testing window shrinks → systems go down more often. We think this is a more plausible explanation for the cascade than "one gang of hackers," but we can't prove it with numbers yet — and nobody in the source proves it either.

Who wins? AI scanner vendors: Calif is already openly selling its method, and its Java find is a great showcase. Who loses? Everyone running old Java stacks in production with a small SRE team — they'll have to choose between patch speed and stability. In our experience, at that fork big business almost always patches first and budgets less time for testing. Hence the outages — and, in our view, this is only the first wave.

⚠️ What we don't know

The source says nothing from the Java vendor — no confirmation of the vulnerability, no patch timeline, no CVE number. There's no breakdown either: how many systems were actually affected, what exactly went wrong at Microsoft and Apple, or whether any exploits exist in the wild. The "eightfold increase in vulnerabilities" methodology is also given without detail — you can count that any number of ways. The attribution of Clayton's quote as head of US national intelligence rests entirely on the source — verify it yourself if that matters to you. And don't confuse things: "update your Java" doesn't mean an update is already available.

💬 Worth debating

  • If patches break production more often than attacks do, should you push a critical Java update on a Friday night — or wait a week?
  • AI vulnerability scanners: a defensive tool, or an accelerator for a race where the defenders are always a step behind?
  • "Zero evidence of a link" — is that the conclusion of an investigation, or just an absence of data? Should we trust the official line on day four of the incident?

Forecast: by mid-February 2027, there will either be an official bulletin with a patch for the affected Base64 library, or public confirmation that the flaw has been exploited in a real attack. If neither happens, the find was overhyped.

Source: The Guardian

🤖 AI summary
  • Date: 2026-11-18
  • Source: www.theguardian.com
  • Tags: cybersecurity, Java, outages, ИИ-уязвимости, AI, Microsoft, CrowdStrike, AI
  • Key facts:
    • US Secretary of State Marco Rubio: there's "zero evidence" linking the Microsoft , Apple and bank outages to the NSA — or to any systemic attack.
    • Amid the cascade of failures, the company Calif found a vulnerability in Java — using an OpenAI model; the flaw is in the Base64 library.
    • Regulators in the UK, US and South Korea are urging everyone to install Java updates; not all Java programs are affected.
    • On November 18, 2026, Marco Rubio was talking to reporters at the Solaire hotel in the Philippines — but instead of diplomacy, he was fielding questions about a string of IT outages in the US.
    • Over the past few days, the failures hit Microsoft, major American banks, Apple, air traffic control systems, and the National Security Agency.
    • Users lost access to cloud services and bank accounts, flights were canceled, and some closed communication channels went dark.
    • According to Rubio, none of it is "interconnected," and the country is "not under systemic attack": "We have zero evidence or reason to believe that these isolated incidents are connected to bad actors, hackers, criminals, or another state." Almost simultaneously, researchers at the American company Calif — using OpenAI technology — discovered a vulnerability in Java, which runs 90% of Fortune 500 companies.
    • The flaw allows remote code execution and lets an attacker plant malware in a system.
#кибербезопасность#Java#сбои#ИИ-уязвимости#AI#Microsoft#CrowdStrike#ИИ
← PreviousOne UI 9.5 Could Supercharge Now Brief: Personal Agent Core Engine Will Build the Cards ItselfNext →AI Digest 11/18/2026: The US wants to block China from… and 4 more

Source: www.theguardian.com · post in Telegram