In the accelerating contest to shape artificial intelligence, Meta's Mark Zuckerberg has placed his faith in an old and contested idea: that markets, left to compete, will discipline themselves. Speaking against calls for an industry-wide development pause, Zuckerberg argues that companies building unsafe AI will lose trust and market share — making coordinated restraint unnecessary. His position, which diverges sharply from safety-focused rivals like Anthropic, surfaces a question as old as industrial capitalism itself: whether the invisible hand can be trusted when the consequences of failur