We should pause the race to build superintelligence and only proceed under strong, enforceable safety guarantees and oversight. The point is not to ban AI, but to stop an uncontrolled sprint toward systems no one knows how to make safe.
We are in a competitive race that trades safety for speed, and the potential downside is existential. Coordination and verifiable standards can let us capture the benefits without gambling the future.
Enforceable global safety standards plus technical assurances that frontier systems are controllable — at which point continuing would be defensible.