Amodei, Trump, and China - The Last Human Argument
by Frank Kaufmann
Chinese researchers recently published a model for how an artificial intelligence system could repeatedly improve itself on its own without human input or oversight. It is referred to in shorthand as L5 (Level 5). Another name for this is recursive self‑improvement.
AI at L5 no longer gets upgraded by humans. Rather it redesigns its own process of improvement, passes that redesign on to future systems, and begins a chain of self‑directed progress.
Suddenly and recently frontier lab celebs have suddenly adopted a voice of alarm, calling this step a safety boundary: a point where control might be lost. This was once popularly called “the singularity.”
In Beijing, such an eventuality is not seen as a danger or a threat at all, but rather is seen as a breakthrough to aim for.
But there are not just these two views, the caution of the frontier magnates, and the aggressive advance posture of the Chinese. There is a third view expressed by President Trump and millions of America First supporters. This third group, like the Chinese, do not see an existential or imminent threat posed by AI progress, even progress approaching L5.
This asymmetry derives from the different governance of these two countries. America is a free and open society, China is totalitarian in its concept of the state. Because America is free and open, divergent views arise. Presently in the US, one view is expressed by the president and his supporters, and a second almost opposite view is arising from the AI frontier CEOs.
This phenomenon called L5, or recursive self‑improvement is seen by Trump and the America First populace as a prize in a global race that America must win, just as the Chinese do for the ascendency of their own globalizing ambitions.
That the frontier‑lab leaders have seized upon in the narrative battle in the US is what has long been warned of, namely , what is called the “x-risk,” (extinction risk), when machines function independently of human guidelines.
Anthropic co-founder Dario Amodei, has become a poster boy for the caution movement. Elon Musk, though a Trump-supporting patriot, has long expressed similar concern. Matters are tricky. Questions are not simple. Solutions are not necessarily obvious.
Because of the presence of the “x-risk” factory, this issue naturally rose to the tip of the global relations iceberg. The difference between China and the US comes into highest relief because of what is at stake on two fronts. 1. Can AI become an existential threat? 2. Can AI give a nation like China sufficient advantage in its openly expressed determination to destroy the United States?
One of the main differences between the open society basis of America versus the domestic totalitarianism of PRC is its impact on the respect two ranges regarding the extent of humanitarian care and concern.
The US is built on a view that God is related to every human being in the same way. Each person alive is created by God to be sovereign, free, equal, served by governments we permit, which are subject to the same laws as us.
China in the exact opposite is built on the view that the state defines human viability. Dissent is not permitted, and deviation from state mandates by definition deprives humans of rights, including the right to life should the state see fit.
The US then must approach the question of L5 for the sake of all people. China on the other hand approaches the question purely on its potential benefit for China alone, the range of concern is narrow and limited only to Chinese state power and control.
This is why in the US there are two reactions, whereas in China there is only one. The American approach is concerned with all people alive and in the future. The China approach is for the controlling power of the party elite. The American creed speaks at the scale of the entire species. The American instinct to “slow down” is not, at its root, anti‑progress. (Though there is a great deal of well researched and convincing observation that the frontier AI leaders are pursuing private power advantage to create a “duopoly” over the infinite wealth tied to AI development.)
China is well known and on the record on meddling in US institutions and manipulation of US public opinion (China’s Three Wars Doctrine). This is another reason why Trump and American patriots are suspicious of the caution, slow down movement. It is seen by the right as infected by America haters both China, and the Soros’s, Singhams, and the Hansjörg Wyss’s of the world.
Trump’s view is that if a free country loses control of the most powerful technology of the coming age to a party‑state like China, then the US will not be able to protect our God-given mandate to promote human freedom, equality, and dignity.
That deeper belief is the real subject.
If Level 5 really does carry extinction‑level risk then it triggers the American creed to speak at the scale of the entire species. Then the American instinct to “slow down” is not, at its root, anti‑progress.
The Chinese Communist state on the other hand holds that L5 is nothing more than a multiplying instrument of national power. This narrowness is consistent with a closed view of the human person.
A utopia of genuine Chinese and American collaboration for the good of all would be the best of all worlds. Each side could provide civilizational wisdom to bear on the serious challenges of AI progress. But no such possibility exists. China has long openly declared and acted aggressively on its purpose to destroy America and Western Civilization. Today’s Chinese Communist system is not a shared seminar. It is a rival form of rule skilled at exploiting the openness of our free society.
So the United States faces a dilemma it cannot escape with a slogan. It must not be naive about international collaboration, but it must not betray our human universalism making us no better than any national hegemon.
Level 5 makes the dilemma sharp because it is both an “arms” race and a species‑wide problem. Win in the wrong way and you keep the flag but lose the founding idea. Lose in the name of moral purity and you keep the sermon but lose the country
There is no simple solution. But we must at least seek a standard for how to talk about the problem. Avoid the shallow, irresponsibility of social media narratives so as instead to pause, think, and study these issues conscientiously.
