I think Julia solves the two-language problem within a range of project complexity. If I am writing a simple machine learning model to classify MNIST digits, Py

I think Julia solves the two-language problem within a range of project complexity. If I am writing a simple machine learning model to classify MNIST digits, Python and its libraries is still the easiest way to go (although Julia is fine too). If I'm writing a large language model and processing large amounts of data, perhaps with some fancy plotting downstream, Julia makes more sense. If you're building a large, stable production application then you still might be more inclined to use Rust or C/C++ or the like.