The above answers didn't rely resonate with me. I will try to explain why i started to learn a constructed language called lojban.
1. Better Meaning extraction for human-computer interaction
I want to build AI's and talk to them like real humans. The problem i'm facing is how to convert my English from speech to - text form - then into meaning - so that the computer can take different programmatic actions depending on that meaning..
And i want to do this with complex thoughts - not "Cortana close the light" type of commands. That is easy to do it in English already.
English and most other languages is not well suited for easy parsing of meaning. The amount of ambiguity in any non trivial phrase is enormous. Exponential complexity is a big problem. We humans solve it by putting things into context - and also trough lot's and lot's of experience.
And even so - we sometimes misunderstand each other.. Keep in mind that we are very smart. Computers are kind of dumb at this point.
Building this context and this experience in an AI is almost an inapproachable task. Building the context is even a bigger problem then parsing the relations between words. It's called an AI hard problem. Only big companies like Google and Microsoft have the resources to approach a problem of this size. Cortana is mostly pre-scripted answers at this point (feb 2018). Is not the real deal.
Now if you have a constructed language like lojban - you have at least 2 big benefits:
the speech to text conversion happens very easily - because lojban has no phonetic ambiguity. Words are typed the way they are pronounced. Is a 1 to 1 mapping. I don't need to understand what you meant in order to know for certain what you said.
the grammar is very congruent / precise - and is said to have zero exceptions - this makes parsing and meaning extraction easier by orders of magnitude. Ambiguity is still an issue - but a manageable one.
In English - we extract meaning and solve ambiguity by relying mostly on statistics. We garter billions of phrases - and based on that we infer what is the most probable meaning of a phrase. We do this with models based on neural networks - that are black boxes for the most part. Point is - grammar rules are not that useful in English. When you debug your program - you can't clearly see why a certain chunk of text was summarized to a certain phrase. You can use your human intuition - and you can agree or disagree with the summary - but if you disagree - you can't easily inspect the decision that the computer took at each step.
Introduction to the parsing problem is a good place to start. The production systems are more advanced then this.
In lojban, by contrast things are simpler: there are only verbs(selbri) and nouns(sumti) - actually is more like - functions and arguments to those functions.
One can heavily rely on rules - and that is enough to get the meaning out with a very high accuracy. And in principle is scalable .. phrases, paragraphs, pages .. hole speeches, hole books, the entire Wikipedia - a web of meaning..
So this is why i'm interested in it. A web of meaning - as opposed to a web of data. Right now computers can see the data - but they can't see the meaning.
2. Answer to questions the way an Expert would do it
Another even more interesting use-case for a constructed language is it's ability to provide a sufficient and universal representation of knowledge that is easy to query by computers. We are interested in getting useful answers from big amounts of unstructured data.
Google is trying to do this for ages. You still get bad results even for easy to answer questions. The way an expert answers a question - is different from google.
We need more of that expert answer. This is almost always true for complex questions.
You want the answer alone not 10's of articles that might or might not contain the answer - articles that you need to read yourself.
And this situation is primarily because 99.9% of human knowledge is represented as text - mostly in English - which is very hard to parse, and query in a meaningful way.
So one way to solve this - is to have all this knowledge in a system that is easy to query by design. And this is what constructed languages can potentially offer. And i don't mean sql databases - is way better - is in the computation itself. The way we humans do it internally - the computers will be able to do it as well. We are still a far cry form that - but i think this direction is promising..
3. On a personal level. Language might affect the way you think.
This is best express by watching an awesome movie like Arrival. or by reading articles like this.
If lojban were to be our default language instead of English - i think our entire civilization we would have been centuries more advanced that we are now - not just technologically - but in general.
I'm just learning this stuff myself. But hopefully this allows you to see that there are much more practical applications - with very serious consequences then just: "you look more smart in front of your friends" - that is not something would convince me to get trough the struggle of learning a new language.
Yet there are people that don't struggle that much - for them the other reasons might make more sense.