From PGP to Mythos: a brief history of export controls that didn't stop anyone

The article delves into the recent White House directive compelling Anthropic to restrict the export of its powerful AI models, Mythos and Fable, outside the US. This decision, driven by national security concerns following a perceived security breach and alleged ties of a partner to China, serves as a modern test case for the efficacy of technology export controls.

• Anthropic had previously marketed Mythos as a highly dangerous AI, limiting access to a select group, but reports of a 'jailbreak' and concerns over a South Korean telecom partner's purported China connections precipitated the government ban.
• Historically, efforts to control technology proliferation have met with limited success; the "Crypto Wars" of the 1990s saw the US government fail to suppress PGP encryption, whose source code was famously published as a book.
• Similar challenges arose with the Wassenaar Arrangement, an international agreement intended to control dual-use technologies like spyware, which has been undermined by non-participating nations and lax enforcement by some member states.
• The article concludes that government-mandated export controls are unlikely to be an effective long-term solution for preventing malicious actors from misusing powerful dual-use cyber technologies, creating a dilemma for American AI companies' global competitiveness.

This incident underscores the complex interplay between national security, technological advancement, and global commerce, raising significant questions about the future of AI regulation and the practical limits of government control over information and innovation.