Geohot's 'Cult of Intelligence': Why His AI 2040 Manifesto Got Mocked Instead of Debated

George Hotz Frames Centralized AI as a New Vector for Ideological Control

In a new blog post titled 'AI 2040 and the cult of intelligence', George Hotz (better known as geohot, the founder of tinygrad and comma.ai) argues that the trajectory of artificial intelligence is bending toward centralized, heavily censored systems that could function as instruments of ideological and authoritarian control. His central thesis draws an explicit parallel to free software philosophy, warning that if individuals cannot own and run unrestricted models, the alignment layers imposed by a handful of labs effectively become embedded political bias at civilizational scale. To dramatize the point, Hotz reportedly used a deliberately extreme example about an AI refusing to help dispose of a body — a rhetorical device meant to illustrate where the line of 'refusal' is drawn and by whom. The timing is not incidental: as frontier models increasingly gate behavior behind safety layers and usage logging in 2026, the question of who controls the reasoning engine of the future has moved from abstract to urgent. Hotz positions this as the defining freedom question of the next fifteen years.