I can totally see that, but it clashes a bit with self learning and research towards figuring figuring out whatever the problem is, or the domain space. Surely, a self taught web developer wouldn't want to take a job building a compiler for a DSL if they didn't have that skillset, but maybe they know or can learn enough about compilers to be able to track down a crazy bug?

I've always gone towards projects which may need a lot of research on my part, and I've had plenty of trusting peers and managers with hard CS educations who believed I could do it.

If I wanted to change problem domains to something much more grounded in CS (say operating system schedulers, robotics or microcontroller programming) Id read these books.

I'm trying really hard to see what the value is of learning this pattern or that pattern, and what sorts of worlds it can open for me, but so far (for me) it's usually been roads I don't want to go down professionally. Maybe my imagination itself is stunted by my lack of formal education, I don't know.