Programming paradigms are like salt and sugar... and other seasonings and condiments.

There's a time for one, and then there are times for others.

You don't choose one and use it in large doses without question. That's not engineering, that's an assembly line.

With that said, declarative programming kickstarted the whole IT industry, because when we started writing programs for machines, the machines didn't have enough storage (broadly construed) to store states.

When that's solved, people left functional programming in droves and embraced OOP and imperative programming, because many real-world problems could be solved more rapidly and flexibly with new programming paradigms. And now, some functional programming is making a come back because the scale of information could once again benefit from not having gigabytes of data overloading single servers.

At no point did we as engineers completely real out one programming paradigm and embrace full-heartedly another one, because real-world problems don't come with the same constraints and requirements every single time.

Dismissing a programming paradigm with a tiny set of specific cases misses the point of having programming paradigms in the first place.