Politicians on both sides are suddenly talking about AI, and what a shock — none of them have any idea what they’re talking about. They’re being led by the nose through a conversation that’s opaque and has a long history of hyperbole. This latest round began when a 25-year-old pre-training engineer left Anthropic, saying the company was hurtling toward an unsafe world in which AI could kill humanity. That’s new. Anthropic believing there’s a risk of AI killing humanity isn’t. It’s why the company exists: Its founders left OpenAI because they believed OpenAI didn’t take that danger seriously enough.
Now Anthropic CEO Dario Amodei says the industry needs to “pace the frontier.” OpenAI and Anthropic are primarily racing each other to build the best model, while Google, Meta, and SpaceX AI could enter the conversation. Pacing the frontier means slowing that race while also trusting China to slow down. Buried within the argument is the idea that competing companies should be allowed to coordinate and that the government should back them because the stakes are so high. One might say this is where Anthropic’s messianic complex comes from: If you believe you’re saving humanity every day, then your logic can’t be questioned.
