OpenAI’s Intelligence Age publishes the essay “The eternal complement” – Unite.AI

OpenAI published The Eternal Complement on October 1, 2026, an essay by Hemanth Asirvatham and Elliott Mokski arguing that genius machines could demonstrate their greatest value by doing monotonous work. It is the first essay in the authors’ series on the next economy, published through Intelligence Age, a platform launched by the company in August 2026.

An author’s note states that the essay belongs to a new platform that hosts independent voices exploring the future of AGI, and that it reflects their opinions rather than those of OpenAI or their peers.

The platform for the intelligence age

Intelligence Age launched on August 20, 2026, with an introductory post by Dean Ball, who described it as the blog of OpenAI’s new Strategic Futures team. Ball wrote that the small team asked one overarching question: How free society should be restructured to preserve individual rights and free will while accommodating the emergence of transformative artificial intelligence. His post states that blog entries convey only the opinions of the authors, not OpenAI’s organizational positions, and that the team will also share their work in other formats, including articles, videos, and podcasts.

A note attached to Ball’s post says the blog has been renamed Intelligence Age to distinguish it from the nonprofit AI Futures Project. The platform’s index page describes its mandate as perspectives on the intelligence age and the next economy, and lists The Eternal Complement under the date of October 1, 2026.

The growing cost of progress

The authors open with a discrepancy: Humanity’s equations give a mostly consistent account of the universe from its earliest moments, yet no humans have traveled beyond the Moon. This gap, they write, could mean that deeper truths can be plumbed from a single planet, or that human minds have surpassed humanity’s capacity for execution.

To illustrate the second reading, they compare Galileo’s telescope, built with two lenses and a tube, with the James Webb Space Telescope, which the essay describes as a ten billion dollar observatory whose eighteen mirror segments are designed to fifty nanometer precision and whose construction involved three hundred organizations in fourteen countries.

The essay cites research by Nick Bloom and his co-authors on research productivity in the American economy: More than eighteen times more researchers are now needed to support Moore’s Law than in the early 1970s, economically effective research effort has increased twenty-three times since the 1930s, and measured research productivity has declined by a factor of forty-one. It also cites findings that the workforce of technicians is growing twice as fast as that of scientists, that the use of specialized equipment in science has doubled in four decades, and that a chip factory today costs five times more than it did thirty years ago.

From that paper, Asirvatham and Mokski argue that genius should be understood as an input to a production process. Borrowing economists’ term for inputs that increase mutual value, they describe frontier intelligence and the ability to execute on one’s ideas as complementary, and coin the term institutional intelligence for what they call unsung execution intelligence: the laws, bureaucracy, financing mechanisms, and supply chains that an idea must survive on its path to reality. “The genius designs the monument; the institutions lay the stone,” they write.

Execution, taste and two civilisations

The essay argues that AI is already making execution less rare by writing code, searching unknown literature, and turning sketches into working prototypes, so that ideas that once required an entire organization can increasingly be pursued by one person. In the authors’ account, the scarce input then shifts from execution to taste: deciding what is worth doing and what questions are worth asking. They warn that the relief may prove temporary, because as AI systems start arriving at insights of their own, ideas may abound faster than the supporting infrastructure can absorb them, leaving civilization hungrier than ever for execution.

The authors outline two directional paths. In a deep civilization, superintelligence would overcome the need for more physical capital and bureaucratic orchestration; in a scale civilization, the complexity of nature would surpass the ability of any intelligence to advance without ever-widening experiments in the real world. Which path prevails, they write, depends on how far thought can travel before having to make new contact with reality.

The essay distinguishes three ways in which intelligence advances knowledge: by reasoning from principles, drawing new insights from existing evidence, and gathering new evidence. The first two can travel far only through thought, the authors argue, while the third is stifled by physical processes that cannot be rushed, from chemical reactions and the growth of organisms to cosmic speed limits on spacecraft and signals. They compare the path in depth to a puzzle that is more difficult in the middle, citing Mendeleev’s description of the elements to be discovered and the Standard Model’s anticipation of the Higgs boson decades before its observation. Biology, they write, is the clearest harbinger of the breadth path, because in silico drug simulations remain unfaithful enough to the real world to still require large-scale human trials.

For the width scenario, the essay offers a Dyson sphere as an extreme case: designing one would require extraordinary breakthroughs, but the bulk of the project would involve construction and logistics on an astronomical scale. In such a civilization, the authors conclude, almost all artificial intelligence would be used “not to do brilliant things, but to do boring things,” and major changes could occur millennia or even millions of years later.

A place for human curiosity

Even if future AI systems outperform humans in both producing frontier insights and coordinating the resources needed to make them useful, the authors provide three reasons why human curiosity retains its importance: an enduring obligation to enlightenment, a possible human comparative advantage in frontier intelligence over institutional intelligence, and the creative diversity brought by human variety.

The essay ends with the question that, according to the authors, will determine the future history of civilization and its pace of progress: whether great minds need more space in the world, or whether they do more with less.