America must lead AI development, opinion argues

Tech CEOs who usually embrace change now seek to throttle AI, with complaints likened to a Borscht Belt joke about poor food and small portions.

Anthropic voluntarily held back its Mythos model to allow partners to patch cybersecurity flaws, showing value of America's lead but questioning government's role.

Elon Musk reportedly moved from assailing frontier labs to selling them compute at a markup and got his SpaceX IPO out as OpenAI may delay theirs.

Fortune 500 companies are turning to unregulated Chinese AI models to save money and avoid exposing data to domestic labs, accepting risks of snooping and sabotage.

Washington threatens to monopolize superintelligence for safety and geopolitics, turning labs' billion-dollar bets into lose-lose propositions.