Julia is the Plan9 of data science-y languages. It's better than python and R, but I don't see it being better enough to cause a meaningful shift away from those languages.
My experience working in a large research University is that nobody is really using it. We have like 2 people loading the Julia module in Slurm out of 8000 researchers who use our cluster.
Same. I'm one of a handful of people that have used Julia in my circles. It's an enthusiastic handful, but the ranks don't seem to be swelling that fast.
The language is stable. Julia code you write now should not break after future updates. This has been the case for quite some time.
The situation with packages is of course more complicated, but there isn’t much problem with breakage, especially with commonly used packages. Perhaps more importantly, Julia’s package system manages dependencies for each of your projects, which pretty much eliminates any upgrading danger:
