Connor here - so what follows is my opinion not Toms :-)
For me, I think the philosophy has if anything been strengthened over the years. A friend once coined the term "YAFET" (Yet Another Front End Technology) representing the fact that today's front-end tool is tomorrow's legacy piece of junk. And its never been more true than today - all that has changed is the names are now more funky :-)
It seems to be the case that we're stuck in this endless cycle of:
"Oh... you built your app in JingleBells ? You should move to the FerrisWheel framework".
(10 mins later)
"Oh... you built your app in FerrisWheel ? You should move to ChocolateHamster".
and so on, and so on.
Dont get me wrong, I've little doubt that each of these new frameworks/technologies are an evolution and improvement of the previous, but businesses run the risk of a giant soup of technologies to look after in the long term. That's ok though - we developers will claim we can glue it altogether with microservices and charge on regardless :-)
Now I'm also a realist, so I know that the rapid fire arrival of new development technologies will not cease. So for me, the principles of SQL and PL/SQL remain unchanged, in that, when I'm dealing with the data, if its nicely encapsulated within SQL and PL/SQL, then I'll be much better equipped to ship out/ship in changes to app dev technologies as they appear. I'd much rather that stability than having to go hunting in 'ChocolateHamster' for data access code so I can refactor it to 'PumpkinSandwich 2.0' or whatever tomorrow's cool tech is.
And even though SQL and PL/SQL are within my realm of expertise, I'm not particular bigoted toward them. It's more about what they offer - great data access language, code that is close to the data, near seamless transition between to two etc etc... For example, I think NodeJS related technologies could easily fall into similar territory, ie, you can run it close to the data (ie, on the database server), you can consolidate data access through it etc. It's the principles via which SQL and PLSQL succeed, rather than them in their own right, if that makes sense.
With the world becoming cloud-cloud-cloud-cloud-cloud-cloud :-) such consolidation and close-to-data code is going to get all the more important, because as servers continue to get insanely fast, latency ultimately becomes king of response time.
Bryn Llewellyn also has a paper on the topic which you might find interesting
https://blogs.oracle.com/plsql-and-ebr/entry/why_use_pl_sql Other welcome to share their thoughts.