Excited for this! While OpenAI has generated plenty of overhyped results (imo as an AI researcher), their focus on large scale empirical research is pretty different from most of the field and had yielded some great discoveries. And with this being started by many of the safety and policy people from OpenAI, I am pretty optimistic for it.