A special report recently published in "The Algorithm," a prominent AI newsletter, has revealed that a sharp and unusually public clash broke out over the past weekend among several of President Donald Trump’s current and former AI advisors. During the heated exchange, the advisors lobbed severe public insults at the leading and largest AI companies operating in the United States. In doing so, they exposed deep internal divisions within the administration and among policymakers regarding the strategy required to face the growing technological competition from the People's Republic of China.
Among the central figures participating in this public dispute was David Sacks, who served as President Donald Trump's official AI and crypto "czar" until last March. Sacks did not hold back his criticism, branding models developed by the AI safety firm Anthropic as "lobotomized" and "woke." Simultaneously, Emil Michael, a senior Pentagon official, launched a sharp attack against Dean Ball, who was recently appointed as OpenAI’s head of strategic futures, calling him the "supreme village idiot."
The current public uproar among the president's advisors was ignited by widespread disagreement surrounding the launch of a new AI model named Kimi. Kimi is a completely free, open-source model released last week by the Chinese AI startup Moonshot. The new model apparently showcases capabilities that—at least according to initial impressions—rival the intelligence levels of the highly sophisticated, flagship models of American giants OpenAI and Anthropic. Unlike the free Chinese offering, accessing these American models involves substantial financial costs and is not provided free of charge to users.
The Economic and Political Challenge of Chinese Models
The Chinese model Kimi, alongside other similar Chinese models being launched at a rapid pace, presents a tangible economic and political problem for President Donald Trump. These developments are causing a deep split among the most senior AI strategists in his orbit, dividing them into rival factions. Every time a smart, free, open-source model from Chinese companies like Kimi is released to the global market, US businesses and enterprises see fewer and fewer economic incentives to pay high, recurring fees to access the closed, expensive models of Anthropic or OpenAI.
Given that the immense enthusiasm and investments surrounding these AI companies constitute a central engine and a disproportionate share of total US economic growth, these Chinese models present combined economic and political challenges for President Trump. Anton Leicht, a research fellow at the Carnegie Endowment for International Peace, addressed the issue on his X social media account, writing that these models represent "a threat for an administration that really doesn’t want more economic bad news." The impact of these launches has already been felt in the financial markets, causing fluctuations and shaking major US technology stocks.
The Trump administration faces the difficult question of how it should act in light of this situation. To understand the background, one must keep in mind that all these events are unfolding just a week after New York State imposed the country’s first state-level ban on the construction of new data centers. This step reflects growing public distrust among citizens toward large AI companies. According to various estimates, a non-insignificant portion of the American public feels little sympathy for the struggles of OpenAI or Anthropic as they contend with cheaper foreign competitors, and many would argue that it is not the federal government's job to protect the business interests of these companies.
Open Source Approach vs. Increased Government Intervention
At this specific juncture, there is a minor and highly limited overlap between the public sentiment and the stance of David Sacks. On July 19, Sacks sharply criticized top US AI companies, arguing that they "want the government to eliminate their open source competition." Sacks added that Chinese AI models have become highly popular globally precisely because they are released with far fewer restrictions on how people can use them—putting aside, of course, the built-in state censorship mechanisms within China.
However, Sacks no longer holds any official role and no longer serves as an active advisor in the Trump administration. His position, which argues that more open and free AI is superior, has been largely replaced within the administration's circle by an opposing approach that advocates for a much broader and more active role for government intervention and oversight. The leading rationale driving this current approach is that AI models have reached such advanced levels of capability that they pose a direct and tangible threat to national security, meaning the federal government must control how they are utilized.
This perspective has driven and fed the White House's new vetting process, which aims to review and test the security and safety of AI models before they are released and made available to the general public. Dean Ball, a former Trump AI advisor who recently transitioned to OpenAI, criticized this new review process over the weekend, labeling it a "de facto licensing regime for frontier AI."
Ball predicted that Trump might attempt to resolve the Chinese open-source challenge through the application of soft power, such as by fostering fear and deterring American companies from using Chinese models like Kimi. This prediction triggered a sharp and furious reaction from Emil Michael, who, alongside Secretary of Defense Pete Hegseth, serves as one of the Pentagon's primary liaisons and official contacts with AI companies. Michael called Ball the AI industry's "supreme village idiot," expressing strong resentment at the idea of the government using quiet, coercive pressure on private firms instead of operating through open, democratic channels. Michael emphasized that actions must be taken strictly through "the democratic process not some Deep State scheme."
The Source of China's Tech Capabilities and the Use of "Distillation"
Aside from the heated debate, the question remains how a model like Kimi managed to reach such a high and advanced performance level in the first place. Throughout much of the Biden administration, and even during the beginning of the second Trump administration, blocking China's access to the market's leading and most advanced chips was defined as a top national security priority. However, these export controls subsequently loosened. Trump made a controversial decision to allow Nvidia to sell more chips to China, in exchange for the US government taking a direct cut of these revenues and transactions. Alongside this, the federal government has alleged that instances of illegal chip smuggling into China have occurred. Nevertheless, China still possesses relatively limited computing power, and it remains entirely unclear exactly which chips Moonshot used to train its new Kimi model.
Another possibility raised in this context is that the training process of the Chinese model involved "distillation"—a technological practice in which AI models are trained and fine-tuned based on the pre-existing outputs of other sophisticated, established models. Companies like OpenAI and Anthropic have long complained that Chinese AI firms use this method to cut corners, and have even turned to the US government for official assistance to halt the practice. Last April, they received their desired response when the Trump administration announced a series of federal efforts designed to limit and curb this practice.
Despite these efforts, the Chinese model Kimi is already available to the public and offered completely free of charge. Its performance and capabilities are highly close to those of Anthropic's advanced model, which the US government previously deemed so powerful that it decided to briefly shut it down and suspend it over concerns of a direct threat to national security. The exchange of accusations, insults, and fierce debates recorded this past weekend indicates that many among policymakers and Trump's advisors view the launch of Kimi as an urgent wake-up call—though at this stage, none of them can agree on the exact meaning of this call or the practical steps that should be taken in its wake.