I’ve thought about this a lot and could post my longer thoughts later, but to start with the short version is that you can increase a sound’s complexity either vertically or horizontally.
Vertical complexity basically involves layering and we see that in a lot of synths. More and more layers, more voices, more oscillators, etc. But at a certain point, I think it’s more interesting to layer different musical parts in your arrangement rather than have a massive layered sound beneath a single keypress. And digital recording with unlimited track counts means we don’t really need that kind of massive layered complexity within an individual instrument anymore.
Horizontal complexity is changes over time which means modulation, more lfos, more envelopes, etc. But when you think about acoustic instruments, those changes over time happen through performance and expression. Maybe it’s more interesting to get additional movement and detail through things like aftertouch, poly at, and other expressive performance controls rather than programming a lot of lfos and envelopes. And there’s so much you can do with post processing to continue to add complexity over time even after the part has been recorded. So again, maybe having that complexity available at the patch level isn’t entirely necessary.