Not to get back into the AI safety debate from last week, but I really, really want a team of human experts to understand what the AI is suggesting before we administer new medical treatments or synthesize new lifeforms in the lab.