So, the question of exponential growth gets into the specifics of different technological mechanisms, and specifically, that seems most associated with network effects, where (at least early-stage) growth approximates n^2. (The bigger story is more complex, as it is with biological systems. Exponential growth is rarely sustained very long.)
Anything that functions as a network, and this includes information itself, can see periods of exponential growth. Which itself doesn't mean "very fast", but as we all (re-)learned last year, "imperceptibly slowly at first, then very suddenly".
The other elements tend to have other sorts of interactions, and a key reason for splitting up the mechanisms as I did was around not only the technical operations, but the economic implications. In particular, hygiene factors, unlike (but also intrinsically tied to) network mechanisms, behave very differently. That's explored somewhat in this post (principally in comments): https://joindiaspora.com/posts/e045d2304aec0139cee5002590d8e...
(That's very preliminary, I'm developing thoughts, and need to do a proper write-up. The fact that I was thinking along those lines spurred a draft version after one line of responses in the thread.)
And you could add all kinds of other elements. So long as you're discussing nonlinearity, there are chaotic systems which can suddenly transition between states, or collapse entirely. There are long-latent potentials waiting for a trigger or opportunity which suddenly spread forth. There are interactions effects. There are subtle accidents of space and time. (Historians and anthropologists tend to haet haet haet heavily on geological determination, but to ignore it entirely strikes me as a peculiar fetish.)
Prediction is hard, especially about the future. But there do seem to be some trends which emerge. And the role of negative unintended consequences does seem to me a significant one.
Markets and financial systems ... that's a whole 'nother game.