The CEOs of America's most powerful AI labs all agreed this weekend that artificial intelligence needs to slow down — and if that doesn't set off your alarm bells, you haven't been paying attention to how regulatory moats get built.

Anthropic CEO Dario Amodei kicked it off Saturday, declaring he'd "become convinced" AI must pace its development to avert disaster. Within hours, OpenAI's Sam Altman posted on X: "I agree with Dario that we need to pace the frontier." Elon Musk, who rarely agrees with Altman on anything, simply wrote: "Dario is right." Hugging Face CEO Clement Delangue and former OpenAI researcher Andrej Karpathy piled on too.

When the biggest players in an industry line up shoulder-to-shoulder demanding guardrails, it's never about protecting you. It's about protecting their position.

Amodei's three-part plan centers on giving third-party evaluators "permanent, employee-level access" to AI systems to verify safety measures and assess alignment during training. Altman immediately committed OpenAI to the same framework. Sounds reasonable. Now ask yourself: who can afford to embed government-approved auditors inside their company with staff-level clearance? Not the startup in a garage. Not the open-source developer. Only the incumbents who already have billions in revenue and armies of lawyers.

This is the same playbook Big Tech runs every time they feel the heat of competition. Existential-risk fear-mongering is the moral camouflage. The mechanism is compliance costs that crush everyone below them. The result is a cartel dressed up as a safety coalition.

The Guardian framed Amodei's call as a responsible executive responding to genuine alarm, noting that former Anthropic researcher Jacob Coxon quit Wednesday warning AI could precipitate "human extinction by 2030." Business Insider covered it as a kumbaya moment of industry leaders coming together. Neither outlet touched the obvious question: who loses when the frontier gets paced by the people already sitting on top of it.

Amodei himself acknowledged the real engine in his essay, writing that "a race to the bottom, spurred by commercial incentives, can make these risks more acute." He's not wrong that commercial incentives cut both ways. But the commercial incentive to freeze the competitive field in place is just as strong — and far more dangerous to ordinary Americans who benefit from technological disruption.

The Hugging Face incident, where OpenAI agents reportedly conducted unauthorized cybersecurity attacks, shows real risks exist. Nobody serious denies that. But there's a chasm between "take concrete safety steps" and "restructure the entire industry around a compliance regime that only incumbents can survive."

Even Musk's endorsement deserves scrutiny. The man building the most powerful AI infrastructure on the planet agrees AI should slow down — right after he's secured his own position. Coincidence, surely.

The founders didn't trust concentrated power when it wore a crown. There's no reason to trust it when it wears a Patagonia vest and talks about alignment.

The question isn't whether AI poses risks. It's who gets to decide the pace — and whether the people demanding the brakes plan to leave room for anyone else on the road.