I started using F# late last year in a *nix environment and it is really well done. There's .NET Interactive which allows for F# in a Jupyter Notebook. You can create .fsx scripts that are interpreted at runtime and handle inline dependencies from NuGet. There's the Fable which compiles F# to JS, TS, Python and Rust.
And of course it's a full on functional language with some great syntactic sugar, especially computation expressions.
The stomaching of .NET basically means getting a much larger set of third-party libraries than OCaml or Haskell can offer.
The part about C++ is true, when you need to use a combination of SIMD/CUDA/MPI there is no game in town but C++. The ecosystem is so vast, we rely on too many scientific/numerical libraries that took tiny armies of grad students to create.
But for new projects in the future, I am looking towards Mojo and Chapel.
Mojo has its foot in the door through compatibility with Python.
Otherwise, there is no reason to pick either when you can do `dotnet add package System.Numerics.Tensors` and access pure C# BLAS that runs at optimal hardware speed that lets you easily implement inference without ever reaching for C or C++. And for GPU-related scenarios there is ILGPU. Now, either is extremely niche compared to Python, but the overall choice will give you great support everywhere, unlike with Chapel or Mojo.