That's because he didn't. Cory published DaOitMK in 2003, meaning it was in the publisher's pipeline in 2002 and written probably 2000-2001. This essay was published in 2016. He's had 15 years of cumulative experiences upon the basis of which to change his opinion on reputation economies -- is it surprising that he's done so?
(Similarly: I disagree violently with the Charlie Stross who wrote Accelerando, circa 1998-2004, on the subject of the singularity. But hey, it's been close to 20 years; younger Charlie had less data and experience to go on, is all.)
((Disclaimer: I have no insight into Cory's mind on this topic other than that which comes from having known him for about 15 years and written a book with him.))