Rethinking the "A.I. race"

An opinion: target implimentation

Excerpt: "The U.S. is winning an AI race—but it’s the wrong one. American policymakers have assumed that artificial general intelligence, or AGI, is achievable relatively soon, and so the aim has been to maintain an 18-month lead over China in reaching it. Washington has restricted Beijing’s access to advanced chips, built AI energy infrastructure, and imposed export controls on other components. This has kept the U.S. ahead in the sprint for AGI, but it’s a contest that can’t be won. America should turn to a different, winnable AI race.

Experts shift the goal post for AGI, or “true intelligence” such as you’d see in a person, with each AI advance. Mastering chess and writing a coherent essay were once held out as AGI benchmarks. AI can now do both, but clear, obvious gaps with human capabilities persist. AGI is a philosophical goal—a perpetually receding horizon—rather than a practical target for strategic victory.

But even if experts could arrive at a stable definition for AI technological supremacy, trying to be the first nation to hit that goal isn’t a smart policy priority. Because of how AI advances, foreign competitors will quickly catch up and likely using far fewer resources.

Model capabilities increase logarithmically with the hardware resources used to train them. In effect, this means you can make a model 90% as good as the model on the current frontier of AI performance with only 10% of the hardware. This is why limiting access to graphics processing units won’t stop America’s competition.

Foreign companies and governments, even those with a fraction of the resources, will still be able to push neck-and-neck with U.S. companies. It was inevitable that a Chinese model like DeepSeek—open-source, cheaply trained—would come along to challenge American pre-eminence in AI, regardless of how tightly Washington controlled chip exports."