https://www.youtube.com/watch?v=3-L9ENuw-ec might be interesting (especially the last 10 or so minutes). It's a talk by one of the primary Modelica developers who is now working for JuliaHub on Dyad. The TLDR is that Modelica is pretty solid (and there's https://openmodelica.github.io/OMJulia.jl/which is Julia OpenModelica integration), but it very much is a project from the 90s with all the baggage that entails. Modelica invented their own vector graphics format (SVG wasn't stable), and were/are very focused on the electrical and mechanical domains somewhat to the detriment of other domains.

A lot of what Dyad is trying to do is see what a Modelica-like language can look like when build for 21st century tooling, and possibly tackling even harder types of problems (e.g. synchronous systems, PDEs, SDEs, etc).

That got its own whole post: https://juliahub.com/blog/why-dyad-a-perspective-for-modelic...