I mean, I agree, but the AI labs clearly don't even if they sometimes pretend they do to achieve their goals. And we're talking about "incidents" caused by the very same people here.