Nim seems to have adopted a lot of the trappings of rust in the last few years. Sort of weird he didn't see Julia first, as it's explicitly aimed at this group.

Speaking as someone who has built things in python, rust and Julia,I have to say that while I love multiple dispatch and the possibilities the Julia ecosystem offers, I'm not surprised he went to Rust over Julia: it's not much slower to hack in (as in, change some code, try running things, see what breaks, fix) because cargo check will catch a lot of errors before actually running things and you can stub things with "todo!()", but it manages complexity much better (if you lean on the typesystem and your code compiles, it's usually almost done), integrates easier into other ecosystems (cbindgen, pyo3)and the finished product will start up faster

Yeh, that tracks with my experience. I'd imagine nim to be a somewhat nice learning experience if one already is used to python and the indentation. And with nimpy the integration into existing python codebases is pretty smooth sailing.

Rust seems to get serious corporate support, which future-proofs it and guarantees ecosystem growth. Not sure about Nim.

Edit: others mentioned Julia, and it indeed looks like a valid choice since they don’t seem to program embedded or safety-critical stuff.

Well, maybe, but IME expecting scientists to have high level thoughts like that about programming language ecosystems is expecting too much.

Nim is very niche and doesn't have nearly the reach or possibilities that Rust does. In terms of expressiveness it's not all that different in practice.

You'd need to expand on possibilities because the only factor I could see where that's remotely true would be the current size of the ecosystem.