I would say they still do have ways of figuring these things out - even incredibly complicated things. It just takes a little longer. It also depends at what level of abstraction you want to look at things anyway. You probably didn't know the physics behind the stuff you built in the 70s; it was just an abstraction of it using capacitors, diodes and resistors. Or the biology of the wood you used for the tree house, at a cellular level. In my opinion it's the same premise.
I'm on the younger end of the HN spectrum, so I grew up playing Minecraft pretty regularly. I remember stumbling into low-level hardware design with a thing in the game called Redstone. You could combine it in numerous ways to make various Boolean logic gates. That progressed into making memory, ALU's, and the like in said game. Fast forward 8 years and I'm a hardware engineer IRL. Once you understand the basic 'building blocks' of how something works you'll develop an interest in the rest of the abstraction generally.
I suppose the best way to see it would be to think of it like a program. You can build a program from bunch of functions, not knowing how those individually "work", and have a finished product at the end. That's still an enjoyable process. You might enjoy it so much that you jump into the library and look at how the functions work. You might go further and look at how the code works at a lower-level language. Then down to machine code. Then at a part level (graphics card, processor, RAM etc). Then a component level. Then at an electro-dynamics level. You can keep going further and further into deeper levels of phsyics.
A kid is still going to have a grand old time plugging in all the parts of their computer and having it turn on, as much as they do building tree-houses and go-karts.