One thing that strikes me as odd now about the US Senate AI Oversight hearings is the focus on _new governmental agencies to create_ rather than _what parts of AI development should be regulated_. Shouldn't you start with the latter to inform the former?