Right, but if the entire site really was about a week's worth of work (or anything like that order of magnitude) then I wouldn't say this was insanely risky, I'd say "why the hell would I not try this?" ;)
Of course he was solving his own problem in the process - so worst case scenario : An educational programming experience that leads to a useful personal tool.
I'm guessing he didn't just write it then sit back and let it run. Things tend to need tweaking/maintainence/upgrading, especially if you're scaling to the numbers he has...
But I agree, why not try stuff like this... Might just work.