More difficult, I have found, is to adapt to the "build vs buy" choice. As a developer, I generally was asked to build something that could solve the problem. As an architect I am more often asked to decide whether we should build it or we should buy something instead. I have had to train myself to ignore that first instinct which screams "Hey... I know how to do this!!" and replace it with a cold evaluation of whether it is WORTH it to build ourselves.
Overall, though, I disagree: I think good programmers can be good architects. But the article raises some interesting points.
I also think that programmer to architect could go either way. Some programmers are very well versed in their domain and current application and make excellent architects while the application remains the same. Once enough changes are made, they are no longer an expert in the code base, and don't know the best ways to extend and improve on it. Others are good at envisioning the current solution and know, from doing it with the programmer hat, how to best architect a change, no matter how it evolves.
