Pretty much. They banked on "if we can solve FSD, we can partially solve humanoid robot autonomy, because both are robots operating in poorly structured real world environments".