I think any amount of 'paternalistic/controlling' turns out to have been justified when, after dismantling the safety teams and pretending not to know what safety is, OpenAI had the HuggingFace series of scandals. You can dislike the idea of safety and people talking about safety, but not only is the evidence right there, but OpenAI came out shamefacedly and literally agreed with Amodei's statements, including that they agreed to pace the frontier.