I have a friend. He wrote his own programming language. Then he proceeded to embed said language into his own webserver. He found someone crazy enough to host his mishmash. Now he does all types of server side application development in this stack he cobbled together.
I know building things to understand them motivates my friend. In fact, I think his motto is "certum quod factum". We believe it, if we build it. Except this friend gets very invested into his tools. Its understandable isn't it? If you build something you have to eat your own dog food for awhile. Then, oh then, :) you start to like the dog food. It tastes so yummy.
That is what has happened to my friend. He built his entire stack and he is very productive with it. He explains that most common stacks are unnecessarily complex (although much less work than inventing a new one from the ground up!) and so he is unwilling to switch.
I worry for my friend. Isn't he hurting his marketability? There aren't many jobs advertising "We need a developer familiar with Bob's Stack". Of course Bob could open source his stack, market it a little, and if it survives peer review maybe he'll become a rockstar. (I prefer ninjas).
I'm directing this at the Bob's out there. Do you ever experience an unmarketable innovator's dilemma? Is the unmarketable innovator a myth? Does your proven ability to take things apart show you could learn anything you needed to? Or are you doomed to keep hacking, waiting for the day you can market your brand of dog food?
I suspect my friend doesn't really care. He just loves learning.