There is no A.I. race

The obsession with beating China at A.I. ignores what actually matters: deploying the technology across the economy.

By ·

Sep 24, 2026

19 cards · 4m · tap to begin

From · · · 4m

There is no A.I. race

The obsession with beating China at A.I. ignores what actually matters: deploying the technology across the economy.

In brief

The obsession with beating China at A.I. ignores what actually matters: deploying the technology across the economy. Winning with A.I. isn't about inventing it first, it's about diffusing it widely across society, which requires safety, trust, and institutional change. Originally reported by The New Yorker.

Britain invented electricity, lost the race

Charles Parsons revolutionized power generation in 1884, but America built the infrastructure. By 1912, the U.S. produced five times more electricity per person than Britain.

“our countrymen have been among the first in inventive genius, but its practical development was in a backward condition.”

The invention myth misses the real story

Historian Nathan Rosenberg showed that what matters isn't the eureka moment but diffusion: how technology spreads into firms, municipalities, and people's lives.

America pioneered fewer than a third of innovations

America pioneered fewer than a third of innovations

In steel, electricity, chemicals, and automobiles, industries that made America dominant, American firms invented fewer than 33 percent of the key innovations.

The real A.I. race is about diffusion

Political scientist Jeffrey Ding argues that winning means spreading A.I. across the entire economy for productivity gains, not just building the fanciest models.

“the lead is about who can diffuse A.I. across the entire economy and gain a productivity boost.”

Safety and trust enable faster adoption

A.I. researchers and Anthropic CEO Dario Amodei are calling for industry-wide safety commitments. Prioritizing reliability over raw power could actually accelerate America's competitive advantage.

“pacing the frontier is actually very aligned with a strategy to win the race.”

Three Mile Island stalled nuclear for decades

The 1979 meltdown destroyed public trust in nuclear energy. Since 1996, the U.S. completed three reactors while China broke ground on nine last year.

Americans already distrust A.I.

A majority of Americans are pessimistic about A.I. The Hugging Face hacking incident could trigger the same pattern that derailed nuclear energy.

China is focused on diffusion, not dominance

race robots china america blue red c

China's A.I.+ initiative aims to integrate the technology into factories, hospitals, and local administration. Chinese leaders rarely discuss runaway superintelligence.

State-directed economies win sprints, not marathons

The Soviet Union launched Sputnik but America built an economy on satellites. China's A.I.+ program is talk; cloud computing adoption shows the U.S. still leads.

America needs to invest in the middle

Community colleges, vocational schools, and state universities can train an A.I.-literate middle class to deploy technology in banks and city halls.

Tech companies invoke China to avoid regulation

Mark Zuckerberg and OpenAI have used China as justification for unchecked growth. OpenAI claimed it needs copyright access or the A.I. race is over.

“If the PRC's developers have unfettered access to data and American companies are left without fair use access, the race for AI is effectively over.”

Philosophers imagined doomsday scenarios

Nick Bostrom's 2014 thought experiment about a decisive strategic advantage has been treated as near-term forecast by technologists imagining killer robots and ethnic bioweapons.

Military A.I. integration takes decades

Former Pentagon A.I. director Jack Shanahan said Project Maven took nine years to reach widespread use. Fully integrating today's A.I. with the military could take another fifteen.

“The gears didn't match. We're trying to turn it at a thousand r.p.m. and the rest of the bureaucracy is moving at fifty r.p.m.”

The Soviet anthrax weapon never worked

The Soviet Union couldn't replicate an anthrax weapon design even with four hundred pages of instructions and the original designers present. Tacit knowledge can't be written down.

“Replicating past work using scientific documents alone cannot be achieved without access to the corresponding tacit skills and the related communal knowledge.”

Cyberattacks require more than A.I. advantage

Chinese hackers already breach American networks. Moving from lurking inside systems to actually disabling them requires far more than winning the A.I. race, including diplomatic breakdown.

China regulates A.I. more strictly than the U.S.

China forces A.I. companies to register algorithms, targets misinformation and deception, and suspended a ChatGPT competitor for calling Russia's invasion a war. The U.S. has no comparable framework.

Both sides lack trust for cooperation

American officials cast Chinese A.I. as an incalculable menace. China interprets U.S. calls for a mutual slowdown as Cold War containment. Mutual distrust blocks safety agreements.

The irony: China isn't chasing superintelligence

While Silicon Valley and Washington imagine China barrelling toward superintelligence, Chinese policy focuses on diffusion. Chinese leaders rarely mention runaway self-improving A.I.

Get your own house in order first

To win the A.I. race, America needs to focus on deploying technology safely and widely, building public trust, and training workers. The real competition is with ourselves.

The essence

Winning with A.I. isn't about inventing it first, it's about diffusing it widely across society, which requires safety, trust, and institutional change.

Read the original on The New Yorker

React

Sign in to react and comment.

Comments (0)

Life is short. Keep it sweet. Respect others' opinions and be kind!

    Recommended next

    More decks on artificial intelligence and related topics.