Your question is not a matter of knowledge. Nobody can tell you what will happen in a few months or years, and even if somebody told you so, it would be a highly questionable answer.
What I can tell you is that the universal language you seek does not yet exist. It is a curious thing however the fact that all programming languages can pretty much do everything. The difference is that some can do it faster, some can do it safer, and some can do it easier.
My advice to you is to study what you want to work on and find out the knowledge required. If you only need Python, then just learn Python. If you only need Julia, then just learn Julia. And if you conclude that you need both Python and Julia, learn both of them.
Tip: If you are working on an AI project, then I highly recommend Python. There are libraries that will be of great use for your project.
Julia is quite strong as a language, and holds a great potential on machine learning applications, but depending on the field, older languages, such as C++ (to put an example) still has enormous influence. It's not easy to trascend from a programming language to another when working on huge collaborations where lots of people put their hands on a code for a prolongued period of time. It's worth following the progress though.
Your question is not a matter of knowledge. Nobody can tell you what will happen in a few months or years, and even if somebody told you so, it would be a highly questionable answer.
What I can tell you is that the universal language you seek does not yet exist. It is a curious thing however the fact that all programming languages can pretty much do everything. The difference is that some can do it faster, some can do it safer, and some can do it easier.
My advice to you is to study what you want to work on and find out the knowledge required. If you only need Python, then just learn Python. If you only need Julia, then just learn Julia. And if you conclude that you need both Python and Julia, learn both of them.
Tip: If you are working on an AI project, then I highly recommend Python. There are libraries that will be of great use for your project.
Julia aims to solve the named: "two language problem": for example several Python libraries relies on legacy C/Fortran code for numerical methods that need more performance. Python itself has been evolving to try to resolve this "two language problem" with options like PyPy or Numba, but no one has already consolidated, and Julia has emerged as a solid alternative.
My personal opinion is that Julia has now a good ecosystem of libraries to be an alternative but just it will not eventually replace Python or R, at least for sure in a short-term. It might eventually replace progressively matlab/octave as it has more posibilities as a general purpose language, and at the same time it has proven its advantages for scientific computing.
I am heavily invested in Python since 2013. The only language that I would consider learning to replace Python at the moment, is Julia. I believe it holds greater potential.
The Julia language developers are doing many things right to get the language into use, specifically interoperability with Python and R. Better performance is of course very appealing as well. It would eliminate the need to work in something like Cython. Julia's parallel computing capabilities is another reason why I may start using it in the near future.