I ask not because the answer isn't obvious, but because I think it exposes the way they are accelerating AI progress in ways that they agree poses risks, which they say needs to be slowed. Sure, unilateral disarmament isn't a viable solution, but as frontrunner, this seems bad.