Robert Gordon just published a paper that is very pessimistic about long-term economic growth. It essentially argues that the big wave of innovation from IT is over. Infotech may radically reorder our lives in the coming decades, but it won’t allow us to make more stuff. Moreover, the big important inventions have been made. His example: would you rather give up all the electronic and IT stuff invented after 1975, or keep all that but have to abandon indoor plumbing?
As a metaphor for the state of play, that nails it. If it doesn’t for you, that is because you have never had to live without indoor plumbing. At night. In the rain.
There are three ripostes to this thesis. The first is that Moore’s Law ain’t over yet. (More to the point, Koomey’s Law ain’t over yet: it states that the energy efficiency of computers doubles every 18 months.) If Moore and Koomey hold for another decade or two, then there is the possibility that AI will emerge seemingly out of nowhere, sometime in the 2030s. Economic growth will surge, but prepare yourself for the robot riots. (I will be joining the rioters, and training my young son to lead them.)
There are reasons to think this will not happen, possibly ever. (I can hope!)
The second riposte, then, is that even without singularian magic, better computers will improve productivity the old-fashioned way. Consider construction robots. My co-blogger and I went back-and-forth on this with our friends Will Baird and Carlos Yu. We concluded that construction is a super-conservative fragmented industry, so there will be no revolution (or robot riots) ... but there will be a slow trickle out of construction (and vehicle operation) just as there was out of agricultural labor, elevator operation, and telegraph delivery. That could keep growth going for a while yet. (Matt Yglesias points out some other sectors ripe for growth here.)
But there is an even more convincing riposte than the above two. It comes from Carlos Yu, and it is based on work by Alexander Field, an economic historian who has pointed out that the Depression-wracked 1930s were likely the most technologically-productive decade of the past century.
What then is the third riposte?
SHORT VERSION: Stop focusing on the damn computers!
LONGER VERSION: “After reading Alexander Field’s book, I distrust the simple correlation in timing between technological improvements and productivity growth. In any case, Gordon is over-focused on the immediate implications of information technology, and he’s missed the one under his nose: biotechnology.
“The Green Revolution doesn’t show up much in American productivity statistics, since our crops had already been improved through a hundred-plus years of scientific selective breeding, but it does show up in global productivity. (Though biotechnology is probably keeping our crop yields high in the biological arms race against disease and pests, but that’s hard to quantify with typical productivity measurements.)
“Cheap effective medicines is another. For many health issues, we can actually buy a few more years of life — or more than a few — at prices far below the break-even point. If we take his question about toilets versus computers and switch it to toilets versus postwar improvements in the applied biological sciences, I think we’d be carrying our chamberpots to the outhouse.
“And now we’re entering the second wave of improvements in biotechnology, ones based on the information technology revolution. It might be a dud. I strongly suspect not.”