Add comment explaining NonCallableMock's Any subclassing

[Avasam]

Brought back and reworded a 5yo comment explaining why NonCallableMock is subclassed with Any.

Feel free to wordsmith.

Ref: #9491

[github-actions]

Diff from mypy_primer, showing the effect of this PR on open source code:

cwltool (https://github.com/common-workflow-language/cwltool) - tests/test_cuda.py:3: note: ... from here:

[srittau]

The problem here are not so much the false positives, but that mocks are explicitly designed to stand in for other types, something that can't be expressed with our static type system.

[AlexWaygood]

I'd also rather avoid introducing unused type: ignores -- I'd rather introduce type: ignores for the false-positives all together as part of #9491.

[Avasam]

Closing in favor of #9491