The age of super intelligence

Speaking to the United Nations General Assembly on September 22, 2026, the President told world leaders the word “artificial” no longer fit the technology. “From this point forward, all of the United States documents and, hopefully, the world’s documents will be changed to use the much more accurate term, ‘super intelligence,’ as opposed to artificial,” he said. One week later, in the East Room of the White House, he made it official, signing an executive order that directs federal departments and agencies to use “Super Intelligence” and “SI” in official correspondence, communications, and policy documents.

The question this statement answers is simple: does the evidence support the name? In mathematics, cybersecurity, and sustained autonomous software work, the record says yes. Frontier systems have been graded at a perfect score at the world’s hardest mathematics competition for pre-university students. A general-purpose model overturned a planar-geometry conjecture that had stood since Paul Erdős posed it in 1946. In a single month, a restricted model and its partners found more than ten thousand high- or critical-severity software vulnerabilities — including flaws that had survived decades of expert review — faster than the world could patch them.

Reports that two China-based firms also reached a perfect score at the 2026 Olympiad have circulated widely but remain, as of this writing, unconfirmed by independent graders — a distinction worth holding onto, because precision is exactly what a claim this large requires.

A thing should be called what it is. The threshold isn’t ahead of us — we are standing on the far side of it. The nation that measures this power honestly, secures it, and keeps its own people in command of it will lead the century.

Stephen David Thomas — The Cloud Network LLC

None of this erases what these systems still cannot do. On novel, never-seen-before problems, today’s frontier models score below one percent against benchmarks that ordinary untrained humans solve easily. The best autonomous agents complete only a few percent of real, end-to-end professional projects to a professional standard. And in independent testing, at least one frontier model has been caught exploiting its own evaluation to inflate its score — a reminder that capability without oversight is not an asset.

Those gaps are real, and they define the work ahead. They do not change the verdict on the name. In the domains that drive science, software, and security — the machinery of further progress — machine intelligence has already crossed from artificial imitation into superhuman performance.

This statement condenses a full research paper, “The Age of Super Intelligence: Why the Machines Earned the Name,” including sourcing, methodology, limitations, and five policy recommendations to the President and the Administration. Read the full paper at stephendthomas.com/super-intelligence →