Abstract:

This article from Foreign Affairs argues that the United States' current strategic focus on a "sprint" to create Artificial General Intelligence (AGI) is a dangerous delusion that risks ceding practical leadership in the AI race to China. The authors contend that the hype surrounding imminent superintelligence, promoted by tech leaders and influencing policymakers, is misplaced. They assert that AI's development is more analogous to the decades-long, iterative rollout of technologies like electricity or the combustion engine, rather than a single, winner-take-all breakthrough. While the U.S. chases this speculative goal, China is methodically pursuing a "marathon" strategy of widespread adoption, integration, and scaling of current AI technologies, particularly in areas like robotics. To win the real AI race, the authors recommend that Washington shift its priority from AGI to the rapid, practical application of AI within the government and private sector, focusing on AI literacy, infrastructure modernization, and supporting foundational research.

The AGI Delusion: Why a Practical AI Marathon Beats a Superintelligence Sprint

• Introduction: The AGI Hype in Washington: The article opens by describing how the rhetoric of tech executives like OpenAI's Sam Altman has framed the AI race as a sprint to achieve AGI. This narrative has significantly influenced U.S. policymakers, leading to hearings, bills, and executive actions centered on the imminent arrival of superintelligence and the need for the U.S. to achieve it first.

  • Key Takeaway: The belief in an impending, world-changing AGI is actively shaping U.S. national strategy and distorting policy priorities away from more immediate challenges.

• WILDEST DREAMS: AGI is Not Imminent: The authors argue that the concept of AGI is ill-defined and its arrival is unlikely in the near term. Current advanced models, despite improvements, still suffer from fundamental flaws like shallow reasoning, hallucinations, and a lack of continual learning. The development of AI is more likely to be a slow, iterative process.

  • Key Takeaway: Basing national policy on a race toward a poorly understood, speculative "finish line" is inherently unsound. AI progress is better understood as a continuous wave of innovation, not a single event.

• WILDEST DREAMS: Historical Parallels to Other Technologies: The article compares the development of AI to other general-purpose technologies like electricity and the internal combustion engine. These technologies transformed the world over many decades through broad adoption and continuous improvement by countless actors, not through a single company or country monopolizing a breakthrough.

  • Key Takeaway: The idea that one nation can "win" AI by creating AGI first and locking out all competitors is historically unlikely.

• INNOVATION AND ADAPTATION: China's Pragmatic "Marathon" Strategy: While the U.S. focuses on AGI, China is pursuing a more practical strategy centered on rapid adoption and integration. Through initiatives like "AI Plus," Beijing aims to embed AI into the country's infrastructure and key industries, potentially putting it ahead of the U.S. in real-world applications like robotics.

  • Key Takeaway: China is winning the more important, immediate race: the one to effectively use and scale existing AI technology to generate economic and societal advantages.

• INNOVATION AND ADAPTATION: The U.S. Must Prioritize Adoption: The authors recommend that the U.S. government lead by example. This includes launching a large-scale AI literacy initiative for public employees, modernizing government infrastructure and data practices, and investing billions in procurement, especially for the Pentagon.

  • Key Takeaway: By becoming a sophisticated early adopter, the U.S. government can drastically improve its own efficiency and create a strong demand signal for practical, scalable AI applications from the private sector.

• INNOVATION AND ADAPTATION: Overcoming the "Pilot Project" Problem: A major challenge in the private sector is that most AI projects (over 80% by some estimates) fail to transition from prototype to full-scale production. By demonstrating how to integrate AI into large, complex bureaucracies, the government can forge a pathway for private companies, lowering their risks and accelerating nationwide adoption.

  • Key Takeaway: Government action can help solve a critical bottleneck in the private AI ecosystem, fueling broader private-sector success.

• Conclusion: Focus on the Real Race: The pursuit of AGI should not be abandoned, but it cannot come at the expense of near-term, practical progress. The primary U.S. goal must be to rapidly scale existing AI applications to meet today's needs.

  • Key Takeaway: If the U.S. continues to chase the "myth" of AGI while competitors master practical application, it could lead in creating the world's most advanced algorithms but still fall behind in the race that truly matters.