My views align pretty closely to yours, but I am a bit more religious about Julia, even. I think languages deserve objective criticism, but as someone who has used Julia for years, there certainly are some small things to critique but the overall product is insanely robust.

I think a significant problem is that many people coming into Julia also have absolutely no idea how to actually write Julia -- for instance, you touched on type annotations -- one of THE MOST important things for Julia performance. The programmers coming in from Python are used to type " annotations" only being for arguments, and not meaning anything at all.

But overall, I think it is one of those things -- I firmly believe Julia will absorb the Scientific domain, and probably even more domains as a general-purpose language. It is kind of a " only time will tell" scenario. As you touched on, the ecosystem really needs time to mature, especially with the way Julia has exploded since 2019, and ideally will continue to explode it seems like we make leaps and bounds everyday.

Luckily, Julia is a great language for iterative development, and to be honest -- usually ends in a lot less code as well. That being said, I really think we develop things a lot faster and with a lot less maintainers.