AI Firms Warn of Doom, Then Go Public. Coincidence?

Anthropic, the company behind the chatbot Claude, has a warning. It says its AI can now write 80% of its own code and design its own successors. The company calls this "recursive self-improvement," a loop where AI gets better without human help. This sounds like the plot of *2001: A Space Odyssey*, where the AI HAL 9000 decided humans were the problem. The warning came just before SpaceX launched its big stock market debut. It also came as Wall Street values AI firms at almost one trillion dollars. Anthropic and OpenAI are both preparing to sell shares to the public. This creates a strange situation. The companies that could make the most money from AI are also the ones shouting the loudest about its dangers. Some believe the warnings are real. Many AI researchers have long worried about keeping AI aligned with human interests. But others see a pattern. Anthropic recently took part in a Vatican meeting on AI. *The Guardian* called this "popewashing," a way to look responsible by borrowing the Church’s moral authority. Critics, like former AI official David Sacks, say the message is: "You want the government to save us... from you." The debate is now central to finance and politics. Public opinion favors limits on AI, and even the Trump administration is starting to write rules. Maybe Anthropic is truly scared. Or maybe this is a smart marketing move before a giant stock sale. Either way, when the people building the machines start warning about them, it is worth paying attention. The business model seems to be: sell fear and hope in the same package.