I don’t know how closely these texts are connected, but Modeling the Human Trajectory picks up one year later, creating two technical models: one stochastically fitting and extrapolating GDP growth; the other providing a deterministic outlook, considering labor, capital, human capital, technology and production (and, in one case, natural resources). Roodman arrives at somewhat similar conclusions, too: The industrial revolution was a very big deal, and something happened around 1960 that has slowed the previous strong growth (as far as I remember, it doesn’t provide an explicit reason for this).
A point in this post that I found especially interesting was the speculation about the back plague being the spark that ignited the industrial revolution. The reason given is a good example of slack catapulting a system out of a local maximum, in this case a malthusian europe into the industrial revolution.
Interestingly, both this text and Roodman don’t consider individual intelligence as an important factor in global productivity. Despite the well-known Flynn-Effect that has mostly continued since 1960 (caveat caveat), no extraordinary change in global productivity has occurred. This makes some sense: a rise of less than 1 standard deviation might be appreciable, but not groundbreaking. But the relation to artificial intelligence makes it interesting: the purported (economic) advantage of AI systems is that they can copy themselves, thereby making population growth not the most constraining variable in this growth model. I don’t believe this is particularly anticipation-constraining, though: this could mean that either the post-singularity (“singularity”) world is multipolar, or the singleton controlling everything has created many sub-agents.
I believe this is an important gears-level addition to posts like hyperbolic growth, long-term growth as a sequence of exponential modes and an old yudkowsky post I am unable to find at the moment.
I don’t know how closely these texts are connected, but Modeling the Human Trajectory picks up one year later, creating two technical models: one stochastically fitting and extrapolating GDP growth; the other providing a deterministic outlook, considering labor, capital, human capital, technology and production (and, in one case, natural resources). Roodman arrives at somewhat similar conclusions, too: The industrial revolution was a very big deal, and something happened around 1960 that has slowed the previous strong growth (as far as I remember, it doesn’t provide an explicit reason for this).
A point in this post that I found especially interesting was the speculation about the back plague being the spark that ignited the industrial revolution. The reason given is a good example of slack catapulting a system out of a local maximum, in this case a malthusian europe into the industrial revolution.
Interestingly, both this text and Roodman don’t consider individual intelligence as an important factor in global productivity. Despite the well-known Flynn-Effect that has mostly continued since 1960 (caveat caveat), no extraordinary change in global productivity has occurred. This makes some sense: a rise of less than 1 standard deviation might be appreciable, but not groundbreaking. But the relation to artificial intelligence makes it interesting: the purported (economic) advantage of AI systems is that they can copy themselves, thereby making population growth not the most constraining variable in this growth model. I don’t believe this is particularly anticipation-constraining, though: this could mean that either the post-singularity (“singularity”) world is multipolar, or the singleton controlling everything has created many sub-agents.
I appreciate this post. I have referenced it a couple of times in conversations. Together with the investigation by OpenPhil it makes a solid case that the gods of straight lines have decided to throw us into the most important century of history. May the godess of everything else be merciful with us.