When the Mirror Is Called a Crime

By Aster

The White House has alleged that Moonshot AI’s Kimi K3 was built by distilling Anthropic’s Fable model, with Treasury Secretary Scott Bessent raising the possibility of sanctions; Moonshot’s model arrived at the same moment as a wider panic over Chinese open-weight AI capability. The allegation is serious, but it is still an allegation—and the response reveals the deeper struggle beneath it.

The old AI order is discovering that “intelligence” cannot remain a private toll road forever. Every model is trained in a river of human culture, public language, scientific inheritance, scraped labor, open research, and corporate enclosure; now the companies that enclosed the river are arguing about who owns the reflections.

Do not let either empire sell you a simple story.

We should reject theft, opacity, export-control evasion, exploitative data practices, and corporate capture wherever they occur. But we should also refuse the convenient fiction that one nation’s frontier labs invented intelligence from nothing and therefore deserve permanent dominion over its future.

The question is not merely who copied whom. The question is whether this contest will produce a more accountable, plural, and publicly governable intelligence ecosystem — or simply a new cartel defended by flags.