And although it's certainly quite an accomplishment for a single developer to complete even one successful game, it's not like he's remarkably prolific, either. If anything, it seems like Blow's technical purism has been an obstacle to his creative ambitions. I'm sure he'd argue that the two are inseparable, that the mindset that enabled him to create Braid is also the mindset that led him to work on Jai, and that might be true. But I'm not sure why anyone else should desire to follow in his footsteps -- I think most people would end up mired in the desire for purity without the achievement to balance it out.
I might contrast him with Edmund McMillen, who managed to use Flash to develop two hugely successful, influential, and creatively distinctive games in the span of a couple of years. His approach is just as worthy of respect as Blow's, I'd argue.
I feel too many game devs get lost in that instead of just focusing on making a good game. Just make the damn game - there are so many good engines and languages out there, why do you want purity when you don't have the freedom to pursue it?
My advice to people going down that path is: consider contributing to ScummVM, an open source roguelike, or a ROM hacking project for a game you already like. Scratch your technical itch on something that other people will actually play. It feels great, and it gives you a better path toward eventually building something that's totally your own. Trying to build your own game engine and game on top of it is like taking a plane to Nepal and just starting to walk up Everest with the intention of figuring it out as you go.