Earlier quoted context omitted.
Or at least: someone writing in Lesswrong, a not particularly reliable site (to put it mildly, but I don't want to get into an argument about it), claims this.
It was extensively cited, and written by a professional researcher using their real name who has no particular axe to grind - what more could you possibly ask for?
From passing familiarity and a bit of a browse around LessWrong is a great place for pretentious airs, florid prose and lengthy posts but none of that makes things inherently more accurate. And for this kind of reporting accuracy of the result is pretty important.