The increasing sophistication of AI models is driving the speed of cyber-security efforts to secure America's critical infrastructure, including water, electricity, health care, financial services, and utilities. The rise in AI capabilities has led to concerns that they could disrupt systems that power daily life. Anthropic is limiting access to its Mythos Preview model and OpenAI is planning to do the same for a forthcoming release. However, researchers have found open-weights models capable of exploiting many of the decades-old bugs that Mythos preview targeted in testing. Security experts warn that these new models could potentially find and weaponize vulnerabilities in as little as a day, giving companies no room for error in their defenses. Cyber defenders argue that limiting the rollout of highly cyber-capable AI models will give them an edge against attackers, particularly in finding and fixing long-standing security vulnerabilities and building more secure software from the start. However this could also hinder maintainers of open-source projects that might need these capabilities the most.