The primary reason is that the technology is designed in such a way that large distributed teams of people can build them without anyone needing to understand the entire structure, because the entire structure is beyond the understanding of a single mind.

A software developer wouldn’t even try to read all the source code of all the libraries their app relies on, nor the machine operations and logic operations and character encodings and chip design and the chemistry and physics of computation.

We’ve consciously decided to abstract things down to reliable interfaces, and as long as the thing behind the interface works, we can understand the interface and build on top of it.

These other reasons are secondary to this one: people don’t understand fully because we’ve gone beyond what a human can fully understand, and deliberately and consciously, decided to adopt this system of abstractions and interface contracts to allow ourselves to operate in the space beyond where the human mind can go.