3 ms·

Interesting thoughts, thanks. > - Benefit / harm > - Near term / long term > - Clearly evident / non-evident I still think that in addition to these things it

by mojuba 6y ago

Interesting thoughts, thanks. > - Benefit / harm > - Near term / long term > - Clearly evident / non-evident I still think that in addition to these things it's the growth function, the linear segments of it, then sudden non-linearity that we fundamentally always get wrong in our predictions. Think of the 1960-1970s that was considered the era of space exploration or the beginning of it, and how people wrongly extrapolated it up to the year 2000 (A Space Odyssey, yep). Or when we try to predict non-linearities, we get it wrong too. Taleb predicted the end of Google in 2007, also made fun of e-scooters back then. But if I'm not mistaken he also praised segways. Well, everybody did. In other words, we are wrong most of the time and we are wrong about the non-linearities in the growth functions. So much so that you could probably even say: if a prediction is a linear extrapolation then it's definitely wrong. If it's a prediction of a breakage then the chances of it being correct are purely random. Look at all the predictions of the next market crash made in the past 10 years. It's as if the market is listening and doing the opposite!

dredmorbius 6y ago
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/e045d2304aec0139cee5002590d8e506 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.

by mojuba 6y ago

dredmorbius 6y ago

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/e045d2304aec0139cee5002590d8e506 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.