It's interesting to watch the wonderfully meandering train of thought(s) triggered by a success report of using a non-mainstream programming language. So let's meander some more...
The strong reactions are understandable. New programming languages are added to the zoo faster than ever before. And within each language there's a steady influx of new frameworks (angular1, angular2, react, vue, ...) each with a learning curve. It's a challenging environment, and it feels like it's getting exponentially harder as time progresses. One must resist the trap of following every silly new idea.
So why bother spending even an afternoon with Haskell? Spend an afternoon playing with Python and you'll be writing code the next morning. Spend a day with Haskell and you'll have more questions than answers.
The fundamental problem is, most likely, the new paradigm. Functional programming takes some time getting used to. Sure, the functional style is creeping into mainstream languages, so most of us have seen and used it already. But pure functional programming is another level. Broadly speaking, everything is a function, a function in the mathematical sense: Same input, same output. Everything is immutable, it's all about chaining pure functions. Crazy, right? Why would you subject yourself to such rigour?
I'll argue that pure functional programming has something to offer that is more relevant than ever.
1. Pure functions are easy to test. Same input, same output, that is a strong guarantee. It'll still be a challenge to thoroughly test every single function, but at least each test is meaningful because there are no interactions with anything outside the function.
2. Pure functions are easy to compose and parallelize. Single-thread performance is hardly ever a relevant benchmark anymore. Sure, C is faster, but good luck trying to parallelize a complex C program. (The worst thing I can imagine is having to debug messed-up, multi-threaded C++ code.)
3. There is structure in how to compose pure functions, and it's structure in the mathematical sense, not in the fuzzy Gang of Four pattern sense. Much of the 'arcane', 'academic' jibber jabber is exactly about these mathematical structures. The power of having a rigorous method for structuring programs is immense. It really is. The hurdle is not a high as it may seem (unless you have a low affinity to maths - im which case you probably shouldn't be programming computers anyway).
4. Pure functions are easy to refactor. (See below.)
I'll also argue that strong typing has a lot to offer and will become more mainstream.
1. Make the compiler your friend, not your enemy. Strong types glue the functions together in just the right way and prevent me from making silly mistakes.
2. Business logic can be encoded in a compact and transparent way. Changes to the business logic (and it always does change, doesn't it?) are also compact and transparent.
3. Refactoring is just ... wonderful. Strong types combined with pure functions, that is a magical combination. Change the type and let the compiler tell you what the implacations are. As there are no side effects code can be refactored with confidence. It is a vastly different experience from any OO language.
Now, finally, to Haskell: It can be characterized as a strongly typed, lazy-evaluated, garbage-collected, pure functional programming language. (Lazy evaluation hasn't been mentioned, yet. It's something that needs to be understood and considered in order to write efficient code.)
In "language design space", something like Haskell must exist and will always exist. (I say 'always' because maths has something eternal about it and pure functional programming and strong typing have solid mathematical underpinnings.) And in this "language design space" Haskell dominates. Contrast this with the compiled, weakly typed, imperative, low-level languages like C, C++, Golang, Rust. I have a hard time picking amongst those, but I believe everyone should know at least one of them.
So is it worth learning Haskell? I tried to argue that it's not a fad and the concepts are fundamental and increasingly important. To me, personally, it looks like a an investment I can benefit from for decades to come.
As a business, do the advantages outweigh the costs and risks? We all know that corporate culture often stands in the way of change. One root cause is the asymmetric payoff: A thank you if it goes well, and job-loss if it goes badly. In such an environment inertia is the rational choice. - But if your company truly believes to be an innovator, or leader in the field, or seeks to gain an edge over the competition, then Haskell should be on the list of things to try.