The eternal complement

This essay, authored by OpenAI employees on an independent platform, argues that intellectual progress depends not only on frontier intelligence but on "institutional intelligence" — the capacity of institutions, infrastructure, and bureaucracy to execute ideas. Through examples ranging from Galileo's telescope to the James Webb Space Telescope, the authors demonstrate that each leap in knowledge requires exponentially more support infrastructure. They frame intelligence and execution as economic complements: more of one raises the value of the other. A great idea must survive a long chain of correct local actions across institutions — laws, funding mechanisms, supply chains — before it becomes reality.

The authors outline two potential futures shaped by AI. A "civilization of depth" envisions superintelligence achieving breakthroughs through superior reasoning, simulation, and targeted experimentation, requiring relatively little physical expansion. A "civilization of width" posits that AI will generate ideas faster than infrastructure can execute them, making physical expansion and institutional capacity the critical bottleneck. Notably, while AI currently relieves execution scarcity — enabling ambitious individuals to pursue what once required whole organizations — it may eventually make us more execution-starved than ever by producing ideas faster than supporting infrastructure can absorb them. In a civilization of width, most machine intelligence would be deployed not to do the brilliant, but to do the boring.

Finally, the authors argue that human curiosity remains essential regardless of AI advancement. Humans may hold a comparative advantage in creative frontier intelligence over institutional work, contribute irreplaceable creative diversity, and retain an obligation to understand the world. The essay concludes that the pace of future progress depends on whether great minds need more of the world, or whether they can do more with less.

(Source：OpenAI)