This past Sunday, my poor coworkers were subject to a Sunday Rant by me - meaning I woke up full of sass and fire, and decided to share my views with the team. My boss pointed out - probably rightfully so - that I needed to share my opinion beyond our internal chat.

It is my personal opinion that currently AI is a technology that transforms efficiency, but it isn’t a replacement for human creativity and innovation. It can run through scenarios faster than a human, but it doesn’t create novel ideas without a prompt.

I have been observing the reactions to the blog by Anthropic of the Mythos Preview . The media noise around Anthropic’s Mythos model often leverages “FUD”—Fear, Uncertainty, and Doubt—to make it seem like an existential mystery.

But let’s be clear: While leveraging AI to increase the speed, and in some cases find unknown vulnerabilities - to me Mythos is highlighting years of vulnerabilities that have not been addressed. This information is not new —not the missing patches, the lack of MFA, or the unsecured, internet-facing infrastructure. Mythos is just proving cybersecurity professionals were right – except it is way faster, and offers a distinct warning about what could come in the future.

The Speed of Adaptation What Mythos has highlighted is the incredible speed at which the industry now needs to adapt and adopt cybersecurity principles. We have moved into a world where the time between a vulnerability existing and causing disruption has collapsed from hours to minutes .

Per the blog - Mythos can autonomously find critical vulnerabilities across every major operating system and browser, generating working exploits without human guidance. It identifies complex, “chained” vulnerabilities that were previously too expensive or skill-intensive to execute. This isn’t just a temporary spike; it is a permanent acceleration of the threat landscape.

Project Glasswing: A Defensive Head Start To address this, Anthropic launched Project Glasswing, a multi-party vulnerability coordination effort. By providing early access to Mythos, Anthropic is allowing selected critical infrastructure providers and open-source maintainers to patch their products before adversaries gain similar capabilities. While this coordination is a vital defensive step, its impact depends on whether the patch and disclosure pipeline can keep pace with the speed of AI progress.

Principles at Machine Speed This isn’t about the industry scrambling; it’s about finally executing on principles we’ve established over two decades. The new “ Mythos-ready” Security Program authored by the Cloud Security Alliance (CSA) focuses on moving from reactive cycles to a permanent Vulnerability Operations (VulnOps) function. This means using LLM-powered agents to find our own weaknesses before attackers do.

The National Institute of Standards and Technology (NIST) has also spent considerable time and effort developing frameworks and guidance around the implementation and use of AI over the past 5 years.

The Bottom Line: Accountability to the Basics We can build coalitions like Project Glasswing all day, but they are only as effective as our willingness to adopt the fundamentals as the CSA has highlighted.

The principles haven’t changed, but our timeline to implement them has. If we are not willing to adopt and enforce basic cybersecurity principles with the same urgency that attackers are adopting AI, then we are simply waiting for the next wave of a storm we already know how to survive.

Mythos isn’t a myth. It’s a reality check. The game is the same, but it’s being played at machine speed.