techpolicy.press web signal

Claude Mythos Drives 100,000-Bug Year, Patch Outages Mount

TL;DR

  • Public vulnerability disclosures crossed 100,000 on November 13, 2026, more than double 2025's total, since Claude Mythos Preview launched in April.
  • Microsoft fixed 622 flaws in a single July Patch Tuesday, and Bank of America brought in 1,000 contractors to patch Mythos-found bugs.
  • CISA's Lindsey Cerkovnik warns that non-serious names like Vulnpocalypse make a real threat sound unsolvable and delay response.

Public vulnerability disclosures crossed 100,000 on November 13, more than double the 2025 total, since Anthropic introduced Claude Mythos Preview in April and AI systems began "finding software flaws faster than companies and governments can fix them," according to a Tech Policy Press piece by Justin Hendrix and Ashley Faler.

The article catalogues more than two dozen names the industry has reached for, among them Bugmageddon, Patchpocalypse, Vulnageddon, Software Beirut, Slopdemic, and The Mythos moment, and treats the naming contest as the story. Lindsey Cerkovnik, then a branch chief for vulnerability response at CISA, recalls a presenter at a CVE Program listening session saying, "Let's kill the word 'vulnpocalypse,' because it's just not accurate." Her own line: "Language matters. … It's harder to take action on a serious threat when you give it a non-serious name and you talk about it in a way that makes it sound like it's unsolvable."

The operational numbers are harder to shrug off. Microsoft fixed 622 flaws in a single July Patch Tuesday. Bank of America brought in 1,000 contractors to patch the bugs Mythos found. In the past week, outages hit Microsoft, the Federal Aviation Administration, JPMorgan Chase, the National Security Agency and Apple, which the authors read as evidence that "the rush to patch is itself breaking things."

Jack Cable, CEO of the security firm Corridor, told a House Homeland Security subcommittee in June that "to prevent the bugpocalypse, we need a new path forward." The Cloud Security Alliance has already given the end state a name: the "Bugpocalypse Threshold," the point at which AI-found bugs permanently outpace the industry's ability to fix them. Two researchers we track were circulating the piece by the time we logged it.

Shared on Bluesky by 2 AI experts