I guess we risk double-correcting: presumably Eliezer has already thought about the impact of biases on his forecast, and adjusted it accordingly. (Ok, there is no need really to presume, as he has written at length about doing so).