Excellent commentary, highly recommended read. I think he concludes with some interesting, if slightly vague advice: get into the job market, see what it's like, give yourself some stability, and start your "thing" on the side. If it gets to the point where it becomes more clear that it may take off, then maybe you can quit your day job. But if it never gets there you still have your day job stability. It does mean most likely sacrificing your nights and weekends for a year or something, but it's worth it to maintain both stability (and possibly earn startup funds of your own) and afford yourself comfortable time (due to stability) to work on your thing. Now this does have the unwanted side effect of possibly not providing as much motivation since the startup is not your sole source of potential support. However if you invest significant portions of your day-job earnings into it that can easily be taken care of.
So, the hybrid approach. It seems to work at least as well as the "balls-to-the-wall" startup from scratch approach, and it has much less of a potential crash if it fails. Risk is good, but unnecessary risk is just stupid. There are no rewards for stupidity.