Throwing my hat into the ring (many month too late 
) :
Honestly I think OpenSCAD is totally worth learning. I started fiddling around with in in ~2014 or so and i cursed the hell out of it, because the learning curve is very steep in the beginning (comparable to learning linux on the commandline for the first time). However, think of physicists or biologists or chemists: Many of them can do programming to a certain extend (e.G. in a simpler language like python). Not because they want to be programmers, but because the ability to create software from scratch is a tool to them. The same applies to OpenSCAD. While the standard CAD-Approach has numerous benefits, especially in terms of Speed or for “just quicky designing a simple part”, classic CAD only very rarely is used in a truely parametric fashion. Sure, some variables here and there, but most models (in my opinion) do break very easily. In OpenSCAD in contrast, i basically use absolutely no numeric values at all. In my design-style they are all(!) derived from variables. this seems very slow and inefficient (and it often is), but the flexibility for changing things later, and the level of how “unbreakable” the parts can become is absolutely unmatched compared to classic CAD-modelers.
There are also things that are easy in OpenSCAD, that are hard-to-impossible otherwise. Think like a programmer: you can use (nested) loops, functions, inline-includes from other files, use other files as libs, do fancy arithmetics, and so on. You can also import 3d (STL) and 2d(SVG/DXF)-geometry and continue working from there (You can directly use them as geometry, which is much more complex in classic cad-software). It is basically a programming language, and that comes with some awesome benefits. You can manage your models as code in git, you can build your whole stack of parts from a CI/CD-Pipeline, do proper versioning, branching and so on. You can automatically render images and animations from the code for documentation, and so on. You can apply and enforce coding-standard, do automatic testing-routines and many, many more. Many of these things are used in the OFM-Project.
For the same reasons, other things, that are simple standard-operations in CAD-Modelers, are incredibly hard to pull off in OpenSCAD, like shell-/pipe-/hollowing-Operations and especially chamfers and fillets. While possible, they cannot be used in the same way as with other CAD-Software.
If you are somewhat comfortable with reading and writing code (you don’t need to be a programmer), you should give it a try. As long you can “think in basic, geometric primitives and how to combine or subtract them from each other to get to a finished part”, OpenSCAD can be quite a powerful tool (just use a recent version, which can do multithreading, not the stable from 2021).
I learned it this way:
Constrain yourself to a very basic set of Operations, like “cube(),sphere(),cylinder(),difference(),translate()” and try to build one or two basic models with that. after some cursing and swearing it will work. For the next model, extend the things you already know, e.G. with “rotate()” and “scale()”, and when you’re comfortable using those, add loops, functions and finally value-arrays. The OpenSCAD-Documentation is very good, and some super-complex geometry can sometimes generated in just a few lines of code cleverly combined (just look at the fan-model on the openscad-startpage. it is just that few lines of code). I think this is the easiest way to learn OpenSCAD. When you’re already a programmer, use functions from the beginning.
