Domain expertise is hard but not that hard compared to the insane mental discipline required to write efficient scalable code.

Not all code has to be efficient scalable code. I know some domain experts that were not programmers. They picked up enough Python or Swift + UIKit (this was pre-LLMs) and their applications are now widely used in their domain. In some cases, they contracted some software companies in the past to do the work. However, they did not understand the problem domain well, requiring the domain expert to iterate with them over and over. In the end it was more efficient to let the domain expert to learn enough programming do it with some guidance from experienced programmers.

Also remember that a lot of domain-specific apps do not need to be huge. E.g. I was once at a factory in the 90ies and they needed to do time-consuming calculations for a particular expensive machine. So, someone who worked at the factory wrote a small program in Excel + I think VBA. This accelerated their work extremely and probably saved them 10,000s or 100,000s in manual labor. It was easy for a non-programmer to write, because they knew all the details of the machine and the calculations, since they had done it over and over again. But it's not a enterprise web app or anything and it does not matter if the program runs in 1 or 4 seconds.

By the way, I also consider efficiency to be a specific knowledge domain, or architecting a very large project. So in the end it really depends on how narrow or wide your definition of 'programming' or 'coding' is.