Very True - but that ends up being a critical reason FOR architecture, not against it.  The more distributed things get to be, the more you need to have that "higher view" to help steer the various implementations in a common direction.

Keep in mind that in this setting architecture tends to be advisory, and helps the various teams and constituents first "find" each other, (as simple as "hey - that looks a LOT like that team X did a few years back"), and work on pooling resources so as to not reinvent the same wheel each time.  It's also not very reminiscent of the doers vs thinkers as described:  we're very much embedded with others as part of a common framework, not thundering down judgement from on high.  It also doesn't mean that it has to be a full time occupation, but it IS critical for someone on any given team to have the responsibility to take the "long view" and understand what the series of short-term decisions we keep making during our busy days will do to the technologies that we support. 

Taking that extra step back to look out and think whether you could automate something, or dramatically improve the usability of a feature , or avoid trouble in the coming future by steering a slightly different direction, makes you a thinker.  As many folks on here have said in other ways, you could save yourself a LOT of doing if you were to just think a bit more.