Yeah, NKS clearly came after Mathematica, but I think there's a strong through-line there that goes back to his early work and connects directly to this meta-math stuff he's doing now.
The core language is a pattern matching substitution system, and that's what he eventually distills into his 1d automata and then generalizes to meta-math.
"at the core of Mathematica is the notion of storing collections of rules in which each rule specifies how to transform all pieces of data that are similar enough to match a single Mathematica pattern. And the success of Mathematica provides considerable evidence for the power of this kind of approach." https://www.wolframscience.com/nks/p627--human-thinking/
It's been a while since I read NKS but I recall him saying he started playing with automata earlier and getting some of the ideas for NKS then, but then did Wolfram Research/Mathematica for some time while he was noodling. I think that's kinda supported by his timeline:
https://www.wolframscience.com/nks/notes-1-4--timeline/
7,833 repos using Mathematica [1]
363 repos of that containing "physics" [2]
Did find this though. Looks neat!
https://github.com/FeynCalc/feyncalc maintained by: https://github.com/vsht
[1] https://github.com/search?q=language%3AMathematica&type=Repo... [2] https://github.com/search?q=language%3AMathematica+physics&t...
I don't think GitHub will be able to tell you much about how physicists are working one way or another. Physicists in general, and also mathematicians and researching engineers who I suspect make up the bulk of most Mathematica users, don't use GitHub.
Mathematica predates any kind of statement of NKS by over a decade (and SMP nearly a decade before that). Moreover, I don't see any evidence Mathematica is becoming more "NKS-y" over time; it's a usual sort of moribund enterprise tooling, just for a highly selective enterprise. I don't see any real connection, other than some possible post-hoc handwaving.
Coming from the other side, Mathematica's relevance to modern physics seems at a 10-20 year nadir. (E.g. from what I can tell it missed the boat entirely on the data-crunching nature of many modern experiments, and its proprietary nature makes it difficult to integrate into increasingly-complex pipelines.) So if NKS - a theory of physics - is being researched, it's either having little success or it's not related to Mathematica use.