Hehe, to me these contrarian comments are strongly reinforcing my point about overzealous pedantry. :) Of course the specific string '6174' is specific to base 10, but the idea itself can be applied to other bases.
Here are some additional examples:
dec, 3 digits: 495
hex, 3 digits: 7F8
hex, 30 digits: EECCAA88664421FFDDBB9977553312
dec, 30 digits: 988766544332209987766554332111
hex, 100 digits: FFFFEEEEDDCCCCBBAAAA9998888776666554444333222210FFFFEEDDDDCCCBBBBAA999988777766655554433332211110001
dec, 100 digits: 9999998888888877666666665544444444332222222210999999998877777777665555555544333333332211111111000001
> And I'd say the fact so many people in this very thread don't understand it is exactly the proof that the GPs comment actually has some merit.
Finally someone noticed the irony, thank you.
(On the other hand, at least some people also criticize my tone rather than the point I'm making, which I guess is fair as well.)
So, yes, the described "special" thing about 6174 is actually a special thing about the string 6174 (representing a number in base-10). And I'd say the fact so many people in this very thread don't understand it is exactly the proof that the GPs comment actually has some merit. People kinda mix up properties of numbers and properties of some other mathematical objects — like their representations in base-10. Most of numerological games are concerned with the latter. Which is why it's especially interesting, when something like that happens to hold in other bases, which sadly just isn't the case with Kaprekar's constant.