The history of export controls on sensitive technologies, particularly in the realm of cybersecurity, is a long and often frustrating one. For the past 30 years, governments have grappled with the challenge of restricting the flow of powerful encryption and security tools, often citing national security concerns. Yet, as history has repeatedly shown, these efforts have proven remarkably ineffective. The very nature of software, its ease of replication and global dissemination, makes it incredibly difficult to contain. Once a technology is developed, it tends to find its way into the hands of those who seek it, regardless of regulatory hurdles. Now, as we enter an era dominated by advanced artificial intelligence, the debate around export controls is resurfacing. Anthropic's recent development of Mythos, a sophisticated cybersecurity model, brings this issue into sharp focus. Given the past failures to effectively control the spread of previous generations of security software, it is highly questionable whether current or future export controls will succeed in limiting access to powerful AI-driven cybersecurity tools. The global nature of research and development, coupled with the rapid pace of innovation, suggests that containment is an increasingly elusive goal, raising fundamental questions about the efficacy and relevance of such policies in the modern technological landscape.