The part they’re misremembering is that if you used 39 digits of pi as pi (not 45), it would be enough to calculate the circumference of the observable universe with a forward error of less than the width of a hydrogen atom (not the distance between 3)