I’ve been thinking about this as an academic-design science question for a while, actually!
My sketch-theory of “power use” is that it’s a relationship between the design of a tool and the extent to which a user can leverage that design to gain advantage in the completion of tasks. The “power” of tool use is therefore a relative measure of the advantage granted to a user by a tool.
A tool’s power depends on (1) the tool’s design, (2) the tasks the user wants to use it for, and (3) the user. Factors like functional fixedness come into play, because a power user who deeply understands a tool can think of ways to use it effectively far beyond the tool designers’ intentions.
For a terrible example … say someone owns a conventional claw hammer.
If they use their fist to bash in a nail, they’re not gaining a lot of power from that hammer.
If they use the hammer to bash in a nail, they’re using the hammer for ~1/2 of its features.
If they use the claw on the hammer to remove nails, they’re using all of the features of the hammer as they were designed.
If they use the hammer to split lumber or as a measuring stick, they’re using their knowledge of the hammer’s design to gain advantage beyond the basic features.
By extension of the above, a power user is probably also good at knowing when to stop trying to use the hammer to do things it wasn’t intended, and to go select the right tool for that task instead. So a fourth factor in power use is the other tools available to the user.
More on-topic: I thought of myself as a Windows power user before I’d ever even touched a Mac! But, as others have noted, it was easier to be a Mac power user then than it was to be a Windows power user. macOS’s features and rich app ecosystem afforded me more flexible ways of getting more out of the OS.
These days I think it’s different: Windows has a unix shell, Powershell exists, Power Automate exists… I feel like I’d like the OS more if I used it now.