LaTeX typesetting is a solved problem. Memoir or Classic Thesis, paired with microtype, provide outstanding results and you need to spend zero time on tweaking stuff.
Typst is interesting, but it doesn't yet support all microtypography features provided by microtype. IMHO, those make a big difference.
I’m going to have to disagree with you there. The compile times are long, the error messages are worse than useless, and tikz diagrams are almost always unreadable messes.
Large swathes of mathematics, computer science, and physics involve notations and diagrams that are genuinely hard to typeset, and incredibly repetitive and hard to read if you don’t make heavy use of the macro system. Integrating some actual programming features could be a game changer.
What in microtype makes "a big difference"? I don't recall using it (my LaTeX years are long behind me), but all of the examples on https://www.khirevich.com/latex/microtype/ seem incredibly minor. I don't think I'd notice any of them as the reader.
Sure, I don't like creeks like in your last example. But I absolutely prefer paragraphs, where the final line would be considered 'too short'. It also makes an appreciable impact for me, in how easy a text is to read.
Which is to say, half of these things are pretty subjective.
It depends what you're typesetting—if you're using letter/A4 paper with 1" margins, then you're unlikely to notice any difference; but if you're using narrow columns, then it will vastly reduce the number of paragraphs with ugly huge spaces between words. Margin kerning is the other big feature, but you probably won't notice that unless you're fairly picky.
I'd also not overemphasize the significance of microtype features. They might help with narrow columns but on wider columnds the difference is very small and most people will never notice them at all.
Typst is interesting, but it doesn't yet support all microtypography features provided by microtype. IMHO, those make a big difference.