Interesting post! It's almost exactly what I've found that has worked well for me. Putting in the three or four hours at the beginning of the process of learning something to find the best resources is essential.
That being said, I do think the importance of mastering one or two things should not be overstated, and was kind of ignored in your post. Let me elaborate: for me my deepest skill has been 3D modeling and rendering: it's been about six years of pretty consistent practice, doing motion graphics, general modeling, animation, color theory, and really just trying to absorb all the knowledge I can about this specific passion. This process has been super useful for a couple reasons, and kind of worked to snap me out of the "generalist" learning mindset. For example, I've found the learning curve over time for 3D modeling is pretty generalizable to other skills I learn: because I've sunk so much time into it, and have had to push through plateaus, and grind over hills, and realize when I'm just coasting, whenever I am learning other things, I recognize the emotion swing and know roughly where in the learning process I am. Similarly, the fluency I have with my 3D modeling tools (pretty much all the hotkeys burned into muscle memory) gives me something to strive for when learning other tools and skills. I know what fluency feels like now, and I know when I'm lacking in it.
For me though, the most important thing about pursuing mastery of one or two things instead of full generalization, is that it teaches you patience. Genuine patience. Something I feel is becoming a lost art in our modern time. Seeing how far you've come after six years of working at something, and then looking at your inspirations and seeing how far you still have to go, is a feeling that can either break you, and make you give up, (if you are impatient), or can help you, as you have the choice to just keep grinding away and choosing to be patient. 3D modeling and animation is a ten year (minimum) goal for me, and it's served as a wonderful learning experience for me in many, unforseeable ways.
To tie it back to your post, I take this "ten years" mentality to almost everything I choose to learn nowadays: I try to stay away from the fancy new frameworks and toys, and try to learn fundamental things that I can spend ten years learning and still have space to grow.