How is this, then. If I can model everything as a calculation, such that all things are computers. What isn't a computer?

Exactly my point :)

And my point is if your definition for a thing is so expansive that everything can fit, it seems a poor definition.

I think "description" is more apt than "definition", however still both are different from "distinction". What I mean is, a distinction (by definition) draws a boundary between "that" and "not-that", but descriptions don't need to do that, they are merely conceptual framings that can be applied to whatever is relevant.

For instance, an exponential decay is a bad model of a stepwise function. So, while everything is a computer, not everything computes what you want it to.

For purposes of human life, only a device easy to control is a computer.