AI Researcher Quits, Claims Public Demands Urgent Regulation

Jacob Coxon didn’t storm out of Anthropic because the company was “doing nothing.” He quit because, in his view, they aren’t moving fast enough on safety and he thinks the public is ready for urgent rules. That’s a serious claim. It’s also a familiar move in tech: leave, write the warning letter, and hope Washington fills the gap with a big shiny regulatory machine.

Here’s the problem. “Urgent regulation” usually means rushed regulation. And rushed regulation in Washington doesn’t land on the biggest, richest players. It lands on everyone else. The same firms that can afford compliance armies and lawyered-up paperwork will write the standards, “help” agencies implement them, and then casually lock out smaller competitors. That’s not protecting the public. That’s building a moat.

We’re not allergic to guardrails. But we should be honest about what’s being asked. A public demand for action is not a plan. If AI systems are actually being deployed in ways that could cause real harm, show the receipts. Name the failure modes. Tell us what’s already been tried inside the company and what got rejected. Otherwise “regulate now” becomes an all-purpose slogan that invites Congress to grandstand, then hand the industry a rulebook drafted by the same companies under scrutiny.

If Coxon wants a hearing, fine. If he wants better internal norms, also fine. But we should resist the idea that the only responsible path is to hand sweeping authority to the same federal bureaucracy that can’t secure its own data or keep up with basic cyber threats. Don’t turn one resignation into a blank check for a new AI regime. Put specifics on the table, target real abuses, and stop pretending panic is policy.