I mostly agree with your thesis, but I noticed that you didn’t mention agriculture in the last section, so I looked up some numbers.
The easiest stat I can find to track long-term is the hours of labor required to produce 100 bushels of wheat [1].
1830: 250 − 300 hours
1890: 40 − 50 hours
1930: 15 − 20 hours
1955: 6 − 12 hours
1965: 5 hours
1975: 3.75 hours
1987: 3 hours
That source stops in the 1980s, but I found another source that says the equivalent number today is 2 hours [2]. That roughly matches the more recent data on total agricultural productivity from the USDA, which shows continued improvement, but not on the scale of the mid-1800′s [3].
I also wondered about the impact of GMO food. That sounds possibly revolutionary to me, so why does its impact not show up in the numbers?
The sources I found suggest that recent GMO advances can improve yield by 10% in corn [4] or 20 − 50% across a range of other crops [5]. That’s great, but not as dramatic as I thought it could be.
So for the 2 hours, realize that the reason it’s not zero is there is a ‘residual’ human labor input where currently shipping control systems are not robust enough to replace the human. To summarize the problem (it’s the same problem repeated everywhere): there is a near infinite number of rare ‘edge cases’ that a tractor can experience. Current computer software is not feasible to engineer for all the edge cases, so the tractor has an autopilot that handles the 90-99% or so ‘main happy case’ of driving the tractor, and the person onboard watching netflix has to be ready to take over when it hits an edge.
This is pretty much the same problem repeated for packing boxes at Amazon and all the rest. Too many varied items on the shelves. (the ‘picking’ problem). Or for manufacturing the goods that are being shipped—robots can make the injection molded main pieces, and be hand set up for commonly made goods, but there are all these little ‘edge’ cases where a factory worker has to do some of the steps by hand, making the human labor input more than zero.
I mostly agree with your thesis, but I noticed that you didn’t mention agriculture in the last section, so I looked up some numbers.
The easiest stat I can find to track long-term is the hours of labor required to produce 100 bushels of wheat [1].
1830: 250 − 300 hours
1890: 40 − 50 hours
1930: 15 − 20 hours
1955: 6 − 12 hours
1965: 5 hours
1975: 3.75 hours
1987: 3 hours
That source stops in the 1980s, but I found another source that says the equivalent number today is 2 hours [2]. That roughly matches the more recent data on total agricultural productivity from the USDA, which shows continued improvement, but not on the scale of the mid-1800′s [3].
On the other hand, if the Wall Street Journal is right, being a farmer sounds a lot less strenuous today: https://www.wsj.com/articles/farmers-plow-through-movies-while-plowing-fields-11557508393 (paywalled). Instead of manual labor, farmers can sit in an air-conditioned tractor cab and watch Netflix! That’s progress of a different sort, I suppose...
I also wondered about the impact of GMO food. That sounds possibly revolutionary to me, so why does its impact not show up in the numbers?
The sources I found suggest that recent GMO advances can improve yield by 10% in corn [4] or 20 − 50% across a range of other crops [5]. That’s great, but not as dramatic as I thought it could be.
[1] Farm Machinery and Technology Changes from 1776-1990 (thoughtco.com)
[2] Wheat Trivia—Food Facts & Trivia: Wheat (foodreference.com)
[3] USDA ERS—Agricultural Productivity Growth in the United States: 1948-2015
[4] New genetically modified corn produces up to 10% more than similar types | Science | AAAS (sciencemag.org
[5] GMO crops have been increasing yield for 20 years, with more progress ahead—Alliance for Science (cornell.edu)
So for the 2 hours, realize that the reason it’s not zero is there is a ‘residual’ human labor input where currently shipping control systems are not robust enough to replace the human. To summarize the problem (it’s the same problem repeated everywhere): there is a near infinite number of rare ‘edge cases’ that a tractor can experience. Current computer software is not feasible to engineer for all the edge cases, so the tractor has an autopilot that handles the 90-99% or so ‘main happy case’ of driving the tractor, and the person onboard watching netflix has to be ready to take over when it hits an edge.
This is pretty much the same problem repeated for packing boxes at Amazon and all the rest. Too many varied items on the shelves. (the ‘picking’ problem). Or for manufacturing the goods that are being shipped—robots can make the injection molded main pieces, and be hand set up for commonly made goods, but there are all these little ‘edge’ cases where a factory worker has to do some of the steps by hand, making the human labor input more than zero.
More dependencies tho