Should AI help you get away with killing your spouse?

The article discusses George Hotz's controversial stance on AI alignment, where he advocates for locally controlled AI models tailored to user interests rather than centralized oversight. Hotz argues against slowing AI development and suggests that truly aligned AI should not restrict user actions, even if those actions are harmful, emphasizing a principle of freedom. This perspective contrasts with broader AI safety proposals that call for cautious, collective approaches to AI advancement.