
VULN DISCOVERY SURGE
8×
FORTUNE 500 ON JAVA
90%
MAJOR OUTAGES · 6 DAYS
6
EVIDENCE OF ATTACK
0
In the span of six days, Microsoft email went dark, US flights were grounded in the longest ground stop since 9/11, JPMorgan Chase triggered banking ripple effects, the NSA's classified network went offline, United Airlines stalled departing flights, and Apple services including Apple Pay flickered. The official line from Secretary of State Marco Rubio and Director of National Intelligence Jay Clayton is consistent: these are unrelated incidents caused by routine upgrades gone wrong. Zero evidence of hostile actors.
Extraction Map
The more interesting diagnosis comes from cybersecurity researcher Tal Be'ery, who identifies a common underlying driver: AI. Since Anthropic released its Mythos model in April, the volume of vulnerabilities discovered by leading tech companies has soared eightfold. OpenAI's models are finding previously unseen flaws. The result is an unprecedented patching scramble — organizations racing to fix holes they didn't know existed, with upgrade timelines that keep slipping. The Java vulnerability disclosed by Calif illustrates the dynamic precisely. Using OpenAI technology, researchers discovered a remote code execution flaw in Base64, a library bundled with Java — a language used by 90% of Fortune 500 companies and embedded in banking, aviation, telecom, and national security systems. UCL professor Steven Murdoch clarifies the flaw is specific to that library, not universal to Java, but the broader point stands: AI is finding vulnerabilities faster than organizations can patch them. Clayton's explanation is revealing in its specificity. Upgrades are taking longer than planned. Systems handing off to other systems aren't coming back as expected. This is the language of technical debt — decades of layered complexity now being stress-tested simultaneously as every major institution rushes to both integrate AI and patch the holes AI keeps finding. The FAA blamed a router upgrade. Microsoft blamed a network capacity upgrade. Chase blamed a software bug. Each individual cause is mundane. The pattern is not. The political response ranges from dismissive to cautious. Trump waved it off: "It's life. When you're big and powerful and you have all the money, things like that happen." Democrat Mark Warner, vice-chair of Senate intelligence, backed the no-hostile-actor assessment but framed the upgrades as defensive. Former FBI cyber official Cynthia Kaiser raised the possibility that the upgrade race was fuelled by specific intelligence warnings about planned attacks — a more alarming explanation hiding inside the reassuring one. The structural concern is straightforward. AI has created a vulnerability discovery engine that runs faster than human institutions can respond. The eightfold increase in discovered flaws since April is not slowing down. Every patch is a potential outage. Every outage cascading through interconnected systems — Chase's bug hitting Zelle, Microsoft's email outage hitting millions of workers — reveals how little redundancy exists. The question is not whether these incidents were coordinated. It is whether the upgrade cycle itself has become a systemic risk. Kaiser's warning about opportunistic criminals and scammers capitalizing on the confusion points to the extraction layer. Each outage creates a window where phishing, fraud, and social engineering become more effective. The costs of brittle infrastructure are borne by ordinary users — workers locked out of email, travelers stranded, bank customers unable to access funds — while the AI companies whose tools are driving the discovery-and-patch cycle face no corresponding liability for the disruption their capabilities create.
Worth reading?
SKIP
Score
24
/100
Extractive — AI labs generate vulnerability discovery capabilities that impose massive upgrade costs on every institution, with outage costs borne by end users while discovery tools are monetized privately.
What Is GRIN?
GRIN stands for Generativity, Resilience, Innovation efficiency, and Novelty openness. These four dimensions capture the essential characteristics of any memetic intelligence — whether a nation, a corporation, a religion, or a movement.
About GRINGenerativity
1/5
Resilience
1/5
Efficiency
1/5
Novelty
2/5
Extractive
4/5
Inverted
Does this create new capability or just redistribute existing value?
No new capability for the public. AI finds flaws faster, but the upgrade scramble destroys availability. Net effect is value destruction for end users, value creation for cybersecurity vendors.
Verified Facts (8)
Marco Rubio said the US was 'not under systemic attack' and a cascade of network problems was 'not inter-related'
Microsoft, major US banks, Apple, American air traffic control systems and the National Security Agency have all been affected by technical failures and outages
Since April, when the AI startup Anthropic released its Mythos model, the amount of vulnerabilities found by leading tech companies has soared eightfold
Java, a programming language used by 90% of companies in the Fortune 500
Flights across the US were grounded the next day in the longest 'ground stop' since the September 11 attacks
JPMorgan Chase caused ripple effects throughout the financial system. Chase said it was caused by 'a software bug' and was 'not cyber-related'
Steven Murdoch, a professor of security engineering at University College London, said the flaw did not apply to every Java user and was specific to a library of code – Base64
Cynthia Kaiser, a former deputy director in the FBI's cyber division, said it was 'plausible' that the race to upgrade systems was fuelled by warnings that specific cyber adversaries were planning attacks
More latest news
Read all