[fiatjaf]

Somebody will be tempted to defend him saying: "Well, he takes risks, it's expected that he will fail sometimes."

The problem is not that he fails. The problem is that he never acknowledges his failures nor questions his assumptions behind the failed predictions -- he just mixes them up and say something in the lines of "I was right all the time".

[Chathamization]

That's not true. In fact, he's even brought up underestimating the impact of the internet multiple times as an example of when he was wrong. For instance:

"And what about me personally? I don’t have a turban and crystal ball hidden in a back room, and I’ve made some pretty wrong predictions in my time. The thing, however, is that all of these bad predictions involved stepping outside the models I understood. I underestimated the payoff to the internet; well, I was speculating about technology, and while that’s my right, I don’t have any special expertise."[1]

[1] http://krugman.blogs.nytimes.com/2013/06/02/non-prophet-econ...

[fulafel]

He does talk about "getting it wrong" on his blog fairly regularly. Example: http://mobile.nytimes.com/blogs/krugman/2015/06/10/the-decad...