.jpg)
On Monday, September 21, President Donald Trump sat down with New York City Mayor Zohran Mamdani at Gracie Mansion. It was their third meeting and the first on the mayor’s home turf, the first presidential visit to the residence since Ronald Reagan met Ed Koch in 1981. The next morning Trump addressed the United Nations General Assembly. The two events captured a familiar Trump pattern: transactional warmth with a political opposite, followed by sweeping claims that collided with established science.
Mamdani, a democratic socialist elected on an affordability platform, invited the president. They discussed policing, immigration—including the administration’s move to end Temporary Protected Status for Haitians—and a long-stalled plan to build roughly 12,000 housing units and create tens of thousands of jobs atop the Sunnyside Yard rail depot in Queens. Trump called it “a big job” that would need federal approvals and money. Both men praised the conversation. Trump said he hoped Mamdani would be “a great mayor” and that he wanted what was good for New York. Mamdani said he would work with anyone who could deliver for New Yorkers.
The optics were striking given earlier attacks. Trump had once labeled Mamdani a “100% Communist Lunatic.” The meeting showed both men prioritizing housing and local delivery over ideology, at least for an hour.
The UN Speech and Science Claims
Trump’s Tuesday address mixed foreign-policy warnings on Iran and Ukraine with domestic boasts and a lengthy passage on artificial intelligence. He announced that U.S. documents would drop the term “artificial intelligence” in favor of “super intelligence” or “SI,” calling the old phrase inaccurate because it made the technology sound fake. He rejected any “globalist scheme” to regulate it and said the United States would not stifle growth of a technology he compared to the Industrial Revolution. He added that the United States was leading China “by a lot.”
He then linked AI safety concerns to climate science. The same people who said “we’ll all be dead in 12 years because of global warming,” he argued, had renamed it “climate change” after the planet started cooling instead of warming, “and nobody was dead.” Those people, he said, were now claiming AI would kill everyone. He grouped them with the “Russia, Russia, Russia hoax,” the “Ukraine hoax,” climate change, and open borders.
He also claimed that combining U.S. and Venezuelan oil resources gave the two countries “more than 60 percent of the oil in the world.”
Where the Science Does Not Hold
These statements do not match the evidence.
Climate change is not a rebrand of a failed “global warming” prediction after cooling. Global average temperatures have continued to rise. The 2018 IPCC special report on 1.5°C warming described a roughly 12-year window (from 2018) in which rapid emissions cuts could still limit warming to that level; it did not predict mass death by a fixed date. Scientists still use both “global warming” and “climate change.” The physical evidence—rising temperatures, shrinking ice, sea-level rise, and attribution studies—is extensive and independent of any political slogan.
Current AI systems are not “superintelligence.” In the technical literature, superintelligence or artificial superintelligence (ASI) means a system that substantially exceeds human performance across virtually all cognitive domains, including novel scientific research. Today’s models are powerful in narrow and overlapping areas (coding, language, some reasoning) but remain “jagged”—strong in some tasks, brittle in others—and lack the general, autonomous capability the term implies. Leading researchers at Anthropic, OpenAI, Google DeepMind, and elsewhere have publicly called for slower capability scaling and stronger safety work; those warnings are not limited to climate activists. The U.S. currently leads China in frontier model capability, but experts describe the gap as months rather than a decisive, durable “lot,” and it could close if development paths diverge.
On oil, proven reserves for the United States and Venezuela together account for roughly 22 percent of the world total, not more than 60 percent. Production shares are similar. Venezuela’s reserves are large on paper, but decades of underinvestment mean significant new output would take years and large capital outlays.
Trump’s broader point—that technological progress should not be smothered by worst-case predictions—is a legitimate policy debate. Equating well-documented climate physics with speculative AI catastrophe, or redefining current systems as superintelligence by presidential decree, does not strengthen that debate. It substitutes rhetoric for measurement.
The Gracie Mansion meeting showed two politicians finding overlapping interests in housing and city governance. The UN speech showed the president treating scientific consensus as just another talking point. The first is politics as usual. The second has consequences when the issues are energy systems, climate risk, and the actual state of machine intelligence.
References
- The New York Times, “Trump and Mamdani Show Their Unlikely Rapport, This Time in New York,” Sept. 21, 2026.
- The Guardian, “Trump and Mamdani appear friendly after ‘wide-ranging’ meeting at Gracie Mansion,” Sept. 21, 2026.
- POLITICO, “Mamdani and Trump emerge from Gracie Mansion sit-down with nothing but praise,” Sept. 21, 2026.
- The New York Times, “Fact-Checking Trump’s Speech at the United Nations,” Sept. 22, 2026.
- PolitiFact, “Fact-check: President Donald Trump misleads on Iran war, AI in speech to UN,” Sept. 22, 2026.
- Reuters, “Trump at UN: Threats, diplomacy and ‘super intelligence,’” Sept. 22, 2026.
- Scientific American, “Trump rejects AI regulation, citing parallels with climate change, in U.N. address,” Sept. 22, 2026.
- IPCC Special Report on Global Warming of 1.5°C (2018) and subsequent IPCC assessments.
- Energy Information Administration and Worldometer oil reserve and production data.
- Stanford HAI AI Index and International AI Safety Report 2026 on current system capabilities.
