Contemporary academia especially in the West has a massive surplus of staff.
Many people pursue academic careers solely for a comfortable lifestyle, doing minimal or even no research for long period of time. With extra lack of oversight that allows researchers to isolate themselves they create circles which cover each other.
Occasionally, folks outside of the circle come in and they start finding ton of fraud in the research with multiple big cases in past few years on top universities like Harvard for example.
Thanks to ireland all big US corporations saved hundreds of billions of dollars past few years so now they can get back to US with this massive cash for anything they want (ofc. nothing will get back to EU as long as they ignore tax heavens)
But Triton is just abstraction over CUDA for Python (same like cupy, numba etc.).
If you need low lvl access you still will use CUDA if you want high level you can use Triton or numba even higer you'll just use wrappers like pytorch/jax.
Jai is a lot slower atm than at 2018, and feature-wise it is closer to C than to Swift with 2 different purposes.
Swift compiler time is spend mostly on LLVM stuff and it isn't much slower than other modern LLVM languages (Rust/Julia etc.) <ofc. this don't count this very specific complex type checks that are super slow>