One of the things I most respect about OpenAI is how (relatively) little it constrains employee voices. OpenAI employees can post whatever they want, argue with one another on X, and shape discourse more publicly. I wish other labs would allow the same for their employees