I’d actually argue that if Julia took off as more of a general purpose language, it likely would not be nearly as interesting as it is today. It’s hard to please everyone, and I’d much rather have a language that is truly excellent for numerical/scientific computing than one with better development tooling.
That said, I think people also are too negative on the state of Julia and it’s package ecosystem / tooling. JETLS is pretty much usable day to day at this point and Flexible Julia is a big improvement in terms of IDE support. Package management keeps getting better version after version. gRPC used to be a blocker for me, but we now have a gRPC client that is being used for production applications and the server gap will likely be closed soon as well. Reactant has gotten to the point where I’m using it for all of my ML training and serving, often seeing large speedups relative to PyTorch (2x to as high as 10x on the same hardware). All of this while being an absolute joy to read and write, especially when trivially doing things that would not be straightforward in a glue language like Python. I don’t measure the success of Julia by the number of people using it but by the ever increasing capability of the language and it’s ecosystem. There is no shortage of things to be excited about.