Warfare has evolved to a "sixth domain" where cyber becomes physical. Mass drone swarms act like a distributed software attack, requiring one-to-many defense systems analogous to antivirus software, rather than traditional one-missile-per-target defenses which cannot scale.

The primary strategic advantage of an AI like Claude Mythos is not launching destructive attacks, but finding unique vulnerabilities for quiet, persistent intelligence collection. Its power lies in the slow, insidious shaping of the information environment rather than overt, 'whiz bang' effects.

For the military, the toughest AI adoption challenge isn't on offense, but defense: overcoming institutional resistance to granting AI the autonomy needed to defend networks at machine speed. A human-alert system is too slow, creating a major bureaucratic and command-and-control dilemma.

While AI enhances offensive capabilities, its greatest potential may be defensive. Easterly posits an optimistic future where AI models become so adept at finding and fixing vulnerabilities in code that they effectively 'end cybersecurity as we know it,' leading to a new era of inherently secure software, much like modern cars are inherently safer.

The cybersecurity landscape is now a direct competition between automated AI systems. Attackers use AI to scale personalized attacks, while defenders must deploy their own AI stacks that leverage internal data access to monitor, self-attack, and patch vulnerabilities in real-time.

Kevin Mandia predicts that within two years, all cyberattacks will be AI-driven. The sheer speed of these threats makes human-in-the-loop defense obsolete. The only viable response is a fully autonomous, AI-powered defensive system to counter AI-born threats.

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 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.

While AI gives attackers scale, defenders possess a fundamental advantage: direct access to internal systems like AWS logs and network traffic. A defending AI stack can work with ground-truth data, whereas an attacking AI must infer a system's state from external signals, giving the defender the upper hand.