It's more like going "breadth first, depth later" vs "depth first, breadth later". RH argues for the second.
I 100% with OP since if you start "breadth first" you can just exhaust yourself before you've made any significant progress on the depth dimension (because there's so much variety in technologies, a lifetime is not enough to learn them all, and one you try to "spread out" you never know when to stop"). Also, going "depth first", you get clear feedback about when you've went deep enough and need to branch out a bit: things just start feeling too hard!
Otherwise you can end up "repeating your first year for a decade" and end up with a much smaller total area.
I 100% with OP since if you start "breadth first" you can just exhaust yourself before you've made any significant progress on the depth dimension (because there's so much variety in technologies, a lifetime is not enough to learn them all, and one you try to "spread out" you never know when to stop"). Also, going "depth first", you get clear feedback about when you've went deep enough and need to branch out a bit: things just start feeling too hard!
Otherwise you can end up "repeating your first year for a decade" and end up with a much smaller total area.