On last week’s EconTalk, Professor of economics Robin Hanson presents a what-if scenario of the technological singularity, singling out a version he believes is most predictable—and most likely—in a vast quadrant of unknowns. By way of introduction, Hanson first examines the three great shifts of world production in history: The emergence of tool-using, knowledge-sharing, producing animals such as Neanderthal and humans; the advent of agriculture and the population growth thereafter; and, the industrial revolution, which effectively increased world production at a rate far out-stripping population. This last shift is important because the heart of Hanson’s prediction of the post-singularity economy is the relationship between world production and per capita wealth.
Hanson sets up his argument for a technological singularity based on a purely numerological projection predicting a doubling of world production every two weeks (compared to fifteen years as at present, or one thousand years before the 1800’s). He examines a few technologies offhand beginning with surveillance, then space colonization, and finally nanotechnology, dismissing each of these as incapable of doubling the economy on the scale of his numerological projection. Hanson states that the next great economic shift that could change the market drastically enough to double production every two weeks would probably have to revolutionize labor, which currently averages slightly less than 70% of costs in products. Hanson singles out general artificial intelligence as the only disruptive technology that could bring about this sea change, and his bet as the most likely scenario where AI emerges is whole brain simulation. He acknowledges that the term AI is code for a vast spectrum of possible minds, all of which cannot be fathomed or predicted, which presumably is why he has little interest in going into other scenarios.
In Hanson’s future, should whole brain emulation become not only technologically feasible, but also economically viable, then wages will begin to bottom out as people who are willing to source copies of their brains for labor at say ten cents per hour. This means that labor costs for anything this emulated brain can achieve will drop proportionately with the cost of hardware running the software. Of course “irrelevant” information will be scrapped as part of the optimization process. The result should be, Hanson continues, a world that becomes unimaginably wealthy, especially for those who gain even the tiniest fraction of profit from this technology, but that per capita wealth will plummet. If that sounds good so far, it gets much, much worse. The assumption expressed is the lowest bidders on the labor market will have their minds emulated and sent out to work for next to nothing. That sounds to me like a roundabout way of saying the some sweatshop workers will be exploited for the benefit of all leisure class citizens. And this isn’t even the slavery version of AI. Hanson’s cavalier attitude toward this outcome is more than disturbing. He also predicts a Malthusian catastrophe where every person in the world whose asset is only their ability to work will have no means of earning a living. We’re talking about hundreds of millions—if not billions—of people. And yet, he maintains that this is still a world to look forward to.
To translate Hanson’s projected post-singularity into an analog version, rather than emulating consciousness in hardware, let’s assume through some unpredictable emergence of biotech, it becomes possible to make memory-intact clones of people at the same economic viability Hanson proposes. Now humans have the power to replicate humans in form, consciousness and memory. The same trial-and-error optimization process is applied to keep this labor cheap: Eliminate any irrelevant information, expand or reduce physical growth as necessary for the labor type, consuming as little nutritive resources as possible. As the costs of replication decrease, and these copies become engineered to function on fractions of saline-vitamin solutions than their predecessors, the supply-demand of labor in many fields maintain a perfect and cheap equilibrium. For those of us fortunate to have a nest egg of assets, we’re okay, Jack. Anyone with no other assets than their ability to work will have no means of living. When the invisible hand of the market deals you cards like these, what can you do?
It is easy to imagine brain-in-a-boxes becoming conditioned like the lower caste epsilons in Brave New World. Hanson is willing to accept an economic genocide at an unprecedented scale, brought about by the digital extension of labor exploitation, mass produced at margin-level operation. I wonder how many other economists welcome the manufacture of this ideal ultra-poor labor class at the cost of incalculable human sacrifice?
You can download or stream the podcast from the EconTalk website.