You need logic to understand computers.
So thinking types would understand.
I believe I understand the level at which you're intending to pin down `computers'.
However, the shifting semantic of computer qua `computer' has shifted so much in recent decades that with `smart phones', `tablets', and all *many* think of a `computer' as something more like a laptop or desktop even though so-called `smart phones' and `tablets' either `are' or `contain' computers.
Nowadays I imagine `understanding computers' is closing the gap with `understanding a car', or understanding a abacus.
Most people don't have much interest in understanding cars; they are content to use them, drive, them, operate them.
So too with their programmable digital devices;
most have low to no desire to `understand' them any more than their cars.
And -- if truth be told -- they don't use them for `computing' so much as communicating with others; in this regard they are used more like the CB radios of the 70s and early 80's.
NTs over the developmental course of what lay persons have regarded a `computer' have been both instrumental in their development and their application.
The huge difference between primitive barely-computers -- the first computer I programed had ferrite core memory -and no compiler or interpreter for higher-than-machine-code programming -- and modern smart phones, tablets, laptops, desktops, clusters, and such is several layers of abstraction.
So an NT's `understanding' may be extend over a broad range or be narrowly scoped.
Two NTs may each have `an understanding' of computers or `work with computers' and yet not have enough overlap to hold a coherent discussion if/when they do both have sufficient interest to bridge their co-engendered gap.
A firmware engineer and someone developing apps for GUIized systems may work so many levels of abstraction apart that they each may each experience what people speaking different `natural' languages do when they have no common language to intermediate.
And even if/when both NTs are working and playing at the same level of abstraction the tools and programming languages and paradigms they have-used or are-interested-in may put them at odds.
One's `understanding' of domain-specific problem when looked at through an OOP lens may be so experientially different than the view one gets when looking at the same domain space through a functional paradigm lens where a
domain-specific language may be crafted via
metalinguistic abstraction.
All this said, it seems that as computers have evolved from the days of tubes, relays, and flip flops made from discrete transitors the nature of any conversation about `understanding computers' has become increasingly multi-meta leveled.
So much so that individuals sufficiently well-intentioned may either talk past each other OR end up pinning down their mutually-agreed-upon notion of `computer' to the extent they co-manifest
Folie à deux
In recent decades the OOP paradigm was so entrenched in the thralls of CS departments around the world that it's (mal)practitioners couldn't -- or wouldn't -- even hold a conversation which didn't beg the paradigm; they could only talk in paradigmatic terms of the prejudicial cognitive framework they collectively manifested via Folie à plusieurs
No! I don't want to talk about `methods' which beg the paradigm; I want to talk about `algorithms' and `data structures' which don't favor any one paradigm over an other AT THE OUTSET.
With every additional degree of nested complexity the so-called `natural' language used to talk about arguably MORE-complex languages and the computer subsystems in which they apply becomes more strained and inadequate to the task.
How can one employ mere `English' to convey the depth and breadth of NT's `understanding' or `advantages' vis-a-vis `computers' to whatever extent we/they do?
In recent years I've noticed that such conversations take on an air of developmental psychology in that for me to communicate with someone effectively I have to become aware of which paradigms, languages, concepts, and preferential biases they hold.
My skills and abilities in linguistics, psychology, and philosophy have to improve above and beyond those one might imagine sufficient to `understanding' computers to a level which advantages myself over mere-mortal non-NTs.