agreed. we should normalize this level of skepticism/scrutiny for all claims from big AI labs