I’m 90% convinced AI companies should be legally restricted to developing and licensing AI models—not competing with every business that uses them. I’m also 90% convinced we’ll eventually need to nationalize them. Otherwise, I see two possible outcomes: American AI companies go bankrupt competing against Chinese firms that benefit from state subsidies and allegedly copy their technology. American AI companies become monopolies that swallow entire industries. Neither sounds particularly appealing. Take mathematics. It’s incredible that AI can solve problems humans have spent decades studying. But should the mathematicians whose research made these breakthroughs possible get zero credit while an AI company takes the victory lap? OpenAI reportedly spent $10–40 million in compute on its Navier–Stokes breakthrough.