“Grand schemes like nationalizing AI labs or creating public-private partnerships are risky, as the current landscape of AI lab leaders is a relatively good outcome compared to many counterfactuals.”