I'm a PhD student currently writing my thesis in Typst. On paper this is an absurdly risky decision: it's a new technology without a huge user-base, it's not totally stable yet, etc. But I tried Typst and I had no choice. It was obviously the right thing to do, even though I'm going to have to make a pixel-perfect clone of my university's LaTeX template.
I've been using LaTeX for over ten years and I still wouldn't say that I "know" TeX in any meaningful way. I was not only productive but proficient in Typst in a day or two. If there isn't a package for something that I need (and, surprisingly often, there are packages for what I need, and excellent ones!), I find that I can just do it myself. Quickly. Things that never would have been possible for me with LaTeX are within easy reach. In ways I'd be happy to talk about if anyone's curious, it's been a huge enabler of my productivity as a researcher. I owe this tool so much.
Interestingly, I spend far, far less time yak-shaving in Typst than in TeX, since I can just Do The Thing That I Want. I'm actually focusing on writing instead of figuring out why package A conflicts with package B, and then how to install a different version of just a single package to override one of them, and then... [this sort of thing doesn't/can't happen in Typst; it's a real programming language with real modules] I could go on and on about the (relative) quality of the ergonomics and devex: fast compilation times, as others have said, but so much more than that. Try it out. Just try it out.
I should also say that it's not perfect. There are some funky design decisions for sure. There are some missing features, like including pdfs as images in your document (that one's on its way, I believe). Critically, the quality of the typesetting is maybe... 95% as good as TeX, which is perfect every time. Sometimes things don't kern quite right, and you have to adjust them by hand. It's a work in progress. I'm optimistic that it will achieve parity on that front.