I realized. probably a lot of the reason people who say "warning about ai risk is just advertising for capabilities" say that... is because sam altman seems to straightforwardly do exactly that thing.