In recent weeks, the global digital infrastructure has appeared increasingly fragile. From widespread Microsoft system outages to the disruption of critical air traffic control networks and service interruptions at major financial and tech institutions like Chase, United, and Apple, the digital backbone of the modern economy seems to be buckling. As these high-profile failures have cascaded through the headlines, a wave of speculation has surged alongside them. Conspiracy theories regarding the role of artificial intelligence—and whether these outages are symptoms of a more sinister, automated sabotage—are permeating public discourse.

However, according to CrowdStrike founder and CEO George Kurtz, the reality is far more pragmatic, albeit no less concerning. In a candid conversation on the Rapid Response podcast, hosted by former Fast Company editor-in-chief Robert Safian, Kurtz explained that these outages are not the result of a science-fiction scenario, but rather the manifestation of a fundamental "bill" coming due for the cybersecurity industry. As organizations race to secure their networks against increasingly sophisticated threats, they have collided with the physical and operational limitations of software maintenance.

The Mythos Moment and the Speed of Vulnerability

To understand why so many systems are failing simultaneously, one must look back to the industry’s recent history. Kurtz points to what he calls the "Mythos moment"—a period roughly nine months ago that fundamentally shifted the landscape of software security. When the capabilities of advanced AI-driven tools were unleashed into the public and research spheres, it triggered a massive, industry-wide scramble to identify and catalog vulnerabilities.

"It really set in motion everyone looking for these vulnerabilities," Kurtz noted during his interview. The result was a race against time. Cybersecurity firms, internal IT departments, and independent researchers began identifying flaws at an unprecedented rate. This surge in discovery necessitated an equally massive, urgent push to create patches. However, the industry has now hit what Kurtz describes as the "sound barrier" of deployment. The sheer volume of vulnerabilities being discovered has outpaced the collective ability of organizations to roll out, test, and integrate the necessary patches without causing collateral damage to the systems they are meant to protect.

This cycle of intense vulnerability discovery followed by a frantic, often error-prone remediation process has created a recurring pattern of instability. Kurtz characterizes this transition as a shift from the vulnerabilities identified in "Black October" to the consequences witnessed in "Darker November." The systemic stress is not the result of a single catastrophic failure but the cumulative pressure of attempting to patch thousands of interconnected, legacy, and cloud-based systems simultaneously under the threat of aggressive exploitation.

The "Breach or Break" Paradigm

The modern IT landscape has effectively been forced into a high-stakes dilemma that Kurtz describes as "Breach or Break." Organizations are currently caught in a vice: they are either faced with the high probability of being breached because their systems are not patched fast enough, or they risk breaking those very systems by rushing to deploy updates that have not been adequately stress-tested.

"That is, are we going to get breached because we can’t patch fast enough, or are we just going to break something and hopefully fix it if we roll it out?" Kurtz explained. "That’s kind of a new norm now because the industry can’t keep up."

This is not a new problem in the history of computing—the industry has previously scrambled to plug gaps in foundational code like PHP or Java—but the stakes have shifted. Today, AI agents are capable of turning a theoretical vulnerability into a functional, working exploit within an hour. This compression of time has stripped away the "grace period" that IT teams once enjoyed. In the past, a vulnerability might remain theoretical for days or weeks, allowing for orderly patching. Now, if a patch is not deployed almost immediately, the window of exposure is dangerously narrow.

This reality has forced CEOs and infrastructure leads to make agonizing decisions. In some cases, the risk of a cyberattack is so imminent that organizations are opting to take entire services offline rather than risk the catastrophic fallout of a breach. By proactively severing connections or shutting down servers, companies hope to avoid the exploitation of unpatched systems, even at the cost of immediate service disruptions.

The Complexity of Legacy Infrastructure

A significant portion of the current instability stems from the collision between cutting-edge AI-driven exploitation and aging, "legacy" software architecture. Many critical systems—particularly in the banking and government sectors—are built on layers of older code that were never designed to be updated at the velocity required by modern cybersecurity standards.

"Technology is messy," Kurtz observed. "Things happen, and we’re going to get better and better as an industry at dealing with this sort of AI speed that we’re facing. When you think about all the legacy systems that are out there, and you think about even banking-type systems, there’s a lot of legacy-type code out there, and it’s not as easy just to create a fix while everything is in midstream."

When these legacy systems begin to melt down under the pressure of updates or attacks, the recovery process is often chaotic. Teams are relegated to "24-hour bridges"—extended, high-pressure troubleshooting sessions where engineers and executives attempt to diagnose the root cause of an outage in real-time. In these moments, Kurtz describes a state of "fog of war," where the complexity of the digital environment makes it difficult to distinguish between the effects of an attempted breach, the unintended consequences of a patch, or a simple hardware failure.

The Future of AI in Security

Despite the current atmosphere of anxiety, Kurtz remains focused on the evolution of security tools. He highlights that CrowdStrike is actively developing new mechanisms, such as an "AI Kill Switch," designed to provide a more controlled response to these high-speed threats. However, he warns against the notion that slowing down the development of AI in frontier labs will effectively mitigate the threat from malicious actors.

The reality, as Kurtz sees it, is that the attackers are already utilizing these tools to their advantage. "Slowing down the frontier labs won’t slow down the attackers," he emphasized. The genie is out of the bottle, and the speed of exploitation will continue to be a primary challenge for the foreseeable future. The industry’s focus must therefore shift toward creating more resilient systems that can withstand the pressure of rapid patching and the immediate threat of AI-driven exploits.

As the dust settles from recent outages, the conversation among business leaders has moved beyond mere technological concern to a broader discussion of operational resilience. The lesson of the past few weeks is clear: in an era of AI-accelerated threats, the traditional, manual approaches to software maintenance are no longer sufficient. Organizations are being forced to rethink their entire architecture, weighing the cost of downtime against the existential threat of a security failure. For now, the "breach or break" dilemma remains the defining challenge of a digital economy trying to find its footing in a new, high-speed reality.