Because software code is a language, LLMs are becoming superhuman coders. This makes them incredibly effective at finding system vulnerabilities for hacking (offense). However, this exact same capability makes them equally powerful for identifying and fixing those flaws (defense), leading to a rapid escalation in cybersecurity.

Anthropic's new AI model, Mythos, is so effective at finding and chaining software exploits that it's being treated as a cyberweapon. Its public release is being withheld; instead, it's being used defensively with select partners to harden critical digital infrastructure, signifying a major shift in AI deployment strategy.

Advanced AI cyber tools like Anthropic's Mythos don't create new vulnerabilities; they excel at discovering existing, dormant bugs in human-written code. Their proliferation will catalyze a one-time, industry-wide upgrade cycle, ultimately hardening global infrastructure and leading to a more secure equilibrium between AI-powered offense and defense.

Anthropic's new AI, Claude Mythos, can find software vulnerabilities better than all but the most elite human hackers. This technology effectively gives previously unsophisticated actors the cyber capabilities of a nation-state, posing a significant national security risk.

The same AI models that can exploit system vulnerabilities are also the most effective tools for identifying and fixing those weaknesses. This duality creates a policy paradox: restricting the technology to prevent its misuse as a weapon also prevents its use as a defensive shield, leaving systems vulnerable.

The greatest cybersecurity risk is not powerful AI, but an imbalance where attackers possess capabilities that defenders lack. Open-sourcing models ensures defensive tools can evolve alongside offensive ones, creating a more resilient ecosystem. It empowers defenders to react faster and make the entire system safer for everyone.

The long-term trajectory for AI in cybersecurity might heavily favor defenders. If AI-powered vulnerability scanners become powerful enough to be integrated into coding environments, they could prevent insecure code from ever being deployed, creating a "defense-dominant" world.

A government policy that prevents US AI models from finding security bugs would be counterproductive. To write secure code, an AI must first understand what a vulnerability looks like. Such a ban would force American developers to rely on uncensored foreign models and would paradoxically result in the creation of less secure American software.

Advanced AI models capable of finding complex code vulnerabilities are expected to be publicly available within months. This puts enterprises in an urgent race to find and patch their own security holes before malicious actors use the very same tools to exploit them.