The problem is, so much of is BS that it's physically painful to keep up. Look, someone reinvented a basic lisp-ism in their mediocre language! Yes, new buzzwords for each part that was a function parameter to the original system! Yay! Now you can reduce your boilerplate by 10%!
The system is still an unmaintainable ball of mud, but at least your for() loops can now have ridiculously high jitter and concurrency landmines thanks to multicore use on a system bottlenecked by disk IOPs.
Part of you wants to play the curmudgeon and say "I like my old simple ways" instead of yelling "you idiots are spinning wheels on useless APIs instead of learning to do any actual engineering analysis on maintainability or performance." It's tough to balance!
Where I currently work, I'm being mentored (even at my age and experience!) by a developer who is in his early 70s. He really is a curmudgeon, but what he actually does is perform. His code is some of the best I've ever seen, and consistently outperforms that of the younger crowd. Then he just points out when he's using an "obsolete" technique to run circles around them.
I think a lot of reasons older people encounter resistance is they don't want to play along with the BS.
Lots of 20-somethings would have no trouble working with a graybeard in his 50s if he was patient, played along, and acted as a mentor. But they don't want to work with him if he's rude and dismissive, and there's more of them than him.
When faced with a ball of mud or spaghetti I recommend putting all what can be factored out in a micro-service and connect them via RPC to REST endpoints. Make your services are stateless so you can scale effortless and put an API gateway in the center so you can manage. As the service business logic isn‘t really stateless put your state in a caching server. Actually you can‘t trust a cache so also put in an opaque field into your client interface. Any problems you encounter can be solved however considering the number of services to be build a fully dynamic CI/CD pipeline is a good investment for potential developer demand first.
The system is still an unmaintainable ball of mud, but at least your for() loops can now have ridiculously high jitter and concurrency landmines thanks to multicore use on a system bottlenecked by disk IOPs.
Part of you wants to play the curmudgeon and say "I like my old simple ways" instead of yelling "you idiots are spinning wheels on useless APIs instead of learning to do any actual engineering analysis on maintainability or performance." It's tough to balance!