China’s launch of the free Moonshot model Kimi has provoked a backlash among top US advisors, with former President Donald Trump’s team trading barbs. David Sacks accused Anthropic’s models of being ‘lobotomized’, while Emil Michael branded OpenAI's new head a 'supreme village idiot'. The dispute centres on how to respond to the rising threat from free Chinese AI models, which pose both economic and political challenges for the US administration.

These models, like Kimi, rival those from leading American companies in terms of intelligence. This has led to a debate about whether the government should protect domestic firms or embrace open-source competition. Amid growing distrust of AI companies, some argue that Americans might not empathize with these firms if they are seen as benefiting from cheaper alternatives.

However, the administration’s stance has shifted towards greater regulation, driven by security concerns. This has led to a White House review process aimed at vetting AI models before release. Critics see this as an 'unofficial licensing regime', while supporters believe it is necessary to safeguard national security. The debate continues amid uncertainty over how such models are developed and the role of export controls in their advancement.

Despite efforts to limit China's access to advanced chips, some speculate that the development of these models might involve distillation techniques, a practice where AI models train on existing outputs. This has led to calls for government intervention but also resistance from firms like Anthropic who fear such measures could stifle innovation.