Welcome John!
Dependency control is still a big motivator...though Ren-C aims to solve fundamental problems of fitness-for-purpose of the language.
This has led to some pretty crazy-seeming inventions, like Generalized Isotopes (including antiform FRAME!, which you've already met 
), whole new ways of looking at binding, and experimental designs for solving problems in really new ways.
So it's kind of an "as simple as possible, but no simpler" situation. There are those who say Ren-C's inventions aren't necessary...but that's merely because they're willing to blind themselves to just how poor the language is without them.
Glad you liked it! That demo hasn't been patched up lately due to other priorities. There is a GitHub Workflow for running the process, though:
ren-c/.github/workflows/linux-tcc-build.yml at master · metaeducation/ren-c · GitHub
I'll see if I can get it up to date and working again.
I'd heard of the APE format a while ago. Sounded like a neat trick. But at the time there wasn't much of a library, and no X86 on ARM emulator. Seems to have gotten further along.
Another avenue of attack here are WebAssembly runtimes. We do have a WASI build:
Ren-C is Now on the Clouds ☁️ with WASI
There's a lot of interestingness to the medium. It leads you to look at problem solving in a very different way... to kind of "paint with code". So things other languages would attack with libraries (or mixing the languages you use) you just attack by bending Rebol itself.
I'm not really an evangelist for using Ren-C (or any Redbol) to solve "important" problems, at least not in the current state. But I do think it's a language family worth knowing, and tackling a little task or two in it is worthwhile.
If that generates enough interest to want to shape the future of it, then I welcome participation in the discussions here!