IANAP, but I think using "lazy evaluation" as an analogy is missing one of the most interesting aspects of quantum physics: it's not just that the particles (or what have you) are in an indeterminate state. The fact that they're in an indeterminate state changes the way they behave. The reason it's possible to build up interference patterns using single electrons is because the various probabilities for its path between the electron gun and target are interfering with each other. I'm pretty sure a typical "lazy evaluation" system will not exhibit that type of behaviour :).
It's not. When you're in a superposition of two states, it only appears to be "indeterminate" under measurements which are not orthogonal to the state. To measurements which are orthogonal, the state is 100% determinate.
What's more, measuring with a non-orthogonal measurement makes the state orthogonal to that measurement, and not orthogonal to the other measurement!
This is totally different to how lazy evaluation works.