The original gigaom article was about neophytes (people who only took coursera course) beating out people with much more experience. This has occurred many times on kaggle. All the preprocessing etc. you're talking is not a new thing, a lot of database developers have been doing that for decades (and that's in addition to my original claim that pretty much anyone with a science or engineering degree having those skills as well). As I said elsewhere, this data science is hard meme is just a ploy to inflate salaries (which every career does anyway, so there's nothing particularly sinister about it).