"This is the most aligned model OpenAI has ever had. So most aligned with broadly acceptable human morals, but it's also the least open about its reasoning." Axios' Senior AI Reporter @MadisonMills22 on Astra: "It's less likely to tell you why it's doing something or be open about what's called its chain of thought." "That obviously raises a lot of eyebrows about what these models are going to be able to do without telling us." "It's great that we're getting closer to the model being more…