New York, New York
— April 14, 2026
Anthropic’s New AI Model Is Changing Cybersecurity and Software Faster Than the Market Realizes
Perhaps the biggest tech story of the year? Anthropic’s new AI model, Claude Mythos Preview, is changing the conversation around cybersecurity in real time. Its ability to uncover vulnerabilities in software appears to be so powerful that it is raising serious questions about the security of some of the internet’s most important systems and how to approach said vulnerabilities.
We are not talking about obscure edge cases. We are talking about major operating systems and web browsers. In other words, foundational software that millions of people and organizations rely on every day.
According to Anthropic, the model is surfacing bugs that have gone undetected for decades. Take a look at a few examples Anthropic just published:
- Mythos Preview found a 27-year-old vulnerability in OpenBSD—which has a reputation as one of the most security-hardened operating systems in the world and is used to run firewalls and other critical infrastructure. The vulnerability allowed an attacker to remotely crash any machine running the operating system just by connecting to it
- It also discovered a 16-year-old vulnerability in FFmpeg—which is used by innumerable pieces of software to encode and decode video—in a line of code that automated testing tools had hit five million times without ever catching the problem”
- The model autonomously found and chained together several vulnerabilities in the Linux kernel—the software that runs most of the world’s servers—to allow an attacker to escalate from ordinary user access to complete control of the machine.
These are major developments! And says a lot about where AI is headed next. These models are no longer just helping people write emails, summarize meetings, or generate code. They are beginning to operate much closer to the underlying infrastructure of the digital world.
Given the potential for malicious use, Anthropic is not releasing Mythos Preview to the public. Instead, the company launched Project Glasswing, a program designed to give access to organizations responsible for some of the world’s most critical code. That reportedly includes major technology companies and large enterprises such as CrowdStrike, JPMorgan, and Cisco, with a group totaling a little over 40 organizations.
The goal is straightforward: find and patch vulnerabilities before this type of capability ends up in the wrong hands.
This is also another interesting development in the software versus AI debate unfolding across the tech world and Wall Street. The Market still seems unsure how to value the shift and where things net out.
Even with the Nasdaq rebounding 8% so far in April, software names have not exactly been treated like clear winners in this new AI landscape. For example the Software ETF, IGV is still down at about $80 per share. The disconnect suggests investors are still trying to figure out whether AI will lift the software industry, disrupt it, or force a complete rewrite of how the space is valued.