Sometimes all you have to do is shave the yak and give the wool away, and a nice yak-wool coat will come back to you. :)

Which is to say: sometimes, if you build a low-level library/infrastructure service and release it, someone else will take advantage of it to write the high-level thing you were originally aiming to create.

I'm speaking from experience — my company's DBaaS is built on the premise of just building a really good database domain-model to hold a certain type of large, public-available datasets, for efficient, flexible querying of them; loading those datasets into it; and then giving people access to it (through SQL or various use-case-shaped APIs.) It's a really thorough yak-shave of the DB+ETL layer of what was originally planned to be a higher-level B2C product.

But it turns out I have a pretty unique view of what should be considered "fun", because my yak-shave was everyone else's schlep. Everyone turned out to be dying to get their hands on this thing, because they wanted to spend all their time on a product, rather than getting good data to feed their product. Once we realized that, we stopped trying to build a B2C product at all, and just started selling the yak wool directly.

I never thought about the implication that shaving a yak is the most fun thing you want to do all day. It's so great you would rather shun all your other responsibilities to shave some yaks. I guess I'd better try it.