Comment by Animats
1 day ago
India has the problem with farming that the US is starting to have with AI. Farming in India is still far too labor intensive by world standards. 43% of workers still work in agriculture. [1] For the US, that number is under 2%. China is at 22% as of 2023, and dropping steadily.
This inefficient agricultural system is not by accident. It is supported by heavy subsidies. Attempts to cut the subsidies resulted in riots.[2] Trouble is ongoing. Comments from someone who knows more about this than I do would help here.
The US and most of the EU went through that transition over several generations, and farming is still heavily subsidized in both areas. The transition happened faster in China, and a hukou system was put into place to prevent people from migrating from farms to cities faster than the cities could absorb them.
Looking at how countries coped with a fast transition from labor intensive agriculture to an urban society gives hints on how an AI transition may look. All the Asian countries that went from poor to rich in a generation did this, with different approaches. How that took place may provide more useful info than philosophy.
[1] https://economictimes.indiatimes.com/news/economy/indicators...
[2] https://en.wikipedia.org/wiki/2024%E2%80%942025_Indian_farme...
I liked reading "The Box" about the transition to container shipping.
It was interesting to see this totally-unrelated-to-our-times process from the outside.
From our place in time, container shipping is obvious.
At the time, to people who wanted to ship something, it was ridiculously hard and expensive and risky.
If you were shipping something from cleveland to paris, you might just give up.
Say you were shipping alcohol - only part might arrive, the rest would disappear.
The shipping industry had all KINDS of forces at work to keep the status quo. trucking companies, trains, shipping companies, freight forwarders, longshoremen, stevedores, unions, people with older non-container boats, etc.
and they didn't want standards.
https://en.wikipedia.org/wiki/The_Box_(Levinson_book)
They seem to have made their peace with standards, but now a lot of them don’t want robots. “Automation hurts families” is a common slogan.
https://youtu.be/hr-isyMV1y8?si=grnJqn8AOEuIwkRw
In all fairness, this is exactly why insurance was invented: unreliable shipping. You just took out a policy, and the shipment didn’t make it you took the payout.
It creates a system that diminishes risk, but simultaneously diminishes incentives for improvement.
Ideally the incentives would be driven by the insurance companies: "Put a tamper-proof lock on your crate and get 50% off your theft insurance!"
The industrial revolution was enabled by more efficient agriculture feeing labour to do other work.
>The industrial revolution was enabled by more efficient agriculture feeing labour to do other work
you're not wrong, but that's not exactly what happened. Agriculture itself was mechanized by the industrial revolution, affordable tractors (tillers, farrowers, etc.) and harvesters. mechanized railroads put more perishable agricultural goods "closer" to urban areas, etc.
if you look at the growth industry before that, it was mercantilist overseas trade.
The big productivity gains of the Agricultural Revoution https://en.wikipedia.org/wiki/British_Agricultural_Revolutio... started considerably before the Industrial Revolution, though naturally the Industrial Revolution in turn fed back into agricultural productivity, in time.
3 replies →
The first London coffeehouse opened in the 1650s and the industrial revolution started in ~1760. It just took a while for the habit to catch on.
The mechanization of the ag industry took place 100 years after the industrial revolution. Try again.
3 replies →
Seems like this and parent are referring to two different waves of improvement. Pre-IR improvements in plowing and cultivation (still draft and human traction), followed by post-IR mechanization for harvesting and transport?
Farm labor demands wax and wane with the seasons with harvest typically being the most labor intensive; so, it was actually stationary motors that did the most to reduce labor demands. Tractors mostly helped farmers eliminate the need for horses on a farm.
One of the leading economic history theories of why the industrial revolution happened was that it was largely a result of the Black Plague.
The theory is roughly that before the Black Plague, the population was stuck in Malthusian dynamics at the top of the logistic curve - population had expanded to the level that land could support.
The massive deaths allowed the remaining population to only farm the most productive land, leading to a massive surplus. The elite were able to capture that surplus and fund things like art, science, etc. Some of those scientists were able to create technology that led to further efficiency gains, so that technology could make the economy grow faster than population growth could catch up.
There are a ton of things that allowed that surplus to translate into technology and economic growth. But AFAIK the leading theory is that without the massive shock from population decline due to the bubonic plague, that surplus would have never existed to begin with, so how it was allocated would have been moot.
23 replies →
The flipside is that there must be other work to be done or people starve to death.
Historically, more efficient agriculture meant a population boom. That's kinda the opposite of people starving to death. A lot of agriculture historically and in poor countries like India today is subsistence agriculture, yeoman farmers living off what they grow directly. More efficiency allows them to sell their surplus and to invest the proceedings, kicking off economic growth.
57 replies →
No, people must be given food or they starve to death. Whether or not work is done is completely orthogonal.
16 replies →
Or we take some small portion of that new surplus in productivity and share it among everyone by divorcing the need to work from the need to not starve.
27 replies →
Not necessarily. In the UK:
> Over one million young people in this country are now neither employed, in education nor in training...
yet no starvation. I'm not sure it's a good situation but it is what it is.
5 replies →
Only if we tie being fed to "working" which is a rather inhumane and untenable thing to do
In unrelated news, we are expelling our agricultural and service industry workforces at the moment.
>The flipside is that there must be other work to be done or people starve to death
false. people are not helpless and jobs are not fixed in number nor social welfare gifts. Human creativity and industriousness can be put to task to produce things that other people want. In an absurd example, you could live next door to a new efficient sweater mill, and you could still knit handmade sweaters, customized with people's initials, etc., and their sale would measure the value of your output in. dollars. People don't do this very often because such an economy produces more lucrative jobs than that.
doomers have foreseen the end of the world in every generation going back. The bad speculations have never come true, but there have been some very negative outcomes of fearful people believing the doom and gloom, look no further than the seeds of Marxist revolutions, Fascism, and Naziism, they all start with people feeling economic uncertainty.
9 replies →
Indeed: https://en.wikipedia.org/wiki/Dual-sector_model
What is the next large labour-absorbing sector supposed to be?
It won't be one large one, it will be thousands of little ones.
Every time this happens throughout history (and I mean going all the way back way past industrial revolution, to dawn of agriculture, to the earliest documented history, to the mitochondria, to the earliest stars exploding...) the result of a better way to get work done is more complexity and more diversity in work done (processes for increasing entropy).
The author said not to confuse laws of nature with observations of history, and I take issue with the implication. My perspective is grounded deep in physics, chemistry, biology and anthropology and after spending 10 years fretting over what AI would do to our civilization this decade I am not worried about labor displacement.
What I am worried about is power struggles and brainwashing.
4 replies →
If anything, it will be the trades. We're still a solid time away from being able to replicating what muscles and skin do - and fundamentally, there will always be a need for someone to run cable, terminate wiring and unclog a sewer pipe. At the same time, the trades are desperate for staff after the "academization" push of the last decades.
4 replies →
What's the point of this comment? It's just a basic fact that everyone agrees with that wasn't put into contention by parent comment.
In industrial revolution analogy we are more like the horses and oxen. How did being replaced by machines go for them?
What does that have to do with this article? One of the points was unlike past transitions this is hitting everyone.
People move from a farm to the city. If there's no more white collar jobs where do we go for work? It's unprecedented
That greatly depends on how much handholding is required and for how long.
The difference between mostly right and actually useable without supervision is why self driving cars still aren’t ready. When someone says AI can do job X, they rarely mean it’s good enough for anyone to blindly trust the results of it doing that job.
Every country subsidies their agriculture for national security purpose. You don't want an enemy to starve you in case of a big war.
Name one country which is fully self-sustaining.
Agricultural subsidies exist primarily due to lobbying, the "we don't want to starve" argument is nothing more than an excuse used to justify the fortunes handed out to corporate-scale farmers.
France, USA, Australia to name 3 (https://en.wikipedia.org/wiki/List_of_countries_by_food_self...).
9 replies →
> Agricultural subsidies exist primarily due to lobbying,
That's mostly true, but it's also true that we don't want to starve. There are 330 million hungry mouths in the US and we've got to keep production way above that level or it becomes a big political problem real quick.
If we just let the market set prices, in years where farms are all producing bumper crops, oversupply would push profits way down. This would force many producers to sell their farms (most likely to corporate-scale farmers) and leave the sector. Subsidies keep a nonzero number of producers producing independently. Granted, the corporate-scale farmers (who also accumulate funds via subsidies) can buy out producers who want to sell, but with subsidies, more producers can afford to say no and stay independent.
You're moving the goalpost from "prevent starving" to "fully self-sustaining".
You don't need 99% variety of cuisine in case of a big war, you need calories. A lot of calories.
UPDATE: and BTW, if world population is growing (no global starvation), then it's clearly self-sustaining, no? So some countries must be self-sustaining just by math. At least one country must produce more than it consumes, otherwise, if everyone produced less, then we would have global starvation.
Even if one grants this, it does not make the case that any particular set of subsidies is justified by that reasoning.
You want to have stable food prices so people don't have to worry about basic survival.
3 replies →
It is also the only alternative to a granary system to smooth out the variability of yields each year that might not average out for anything less than 10-15 year spans.
And the granary system regularly still resulted in shortages and famine. While crop subsidization has a bullet proof record of surplus.
From your Wikipedia link:
> Demands: […] Government to ensure at least 50% profit over their overall cost of production.
They demanded 50% guaranteed annual RoR on all farming activities? That’s a wild demand.
[dead]
You have to look at the population split between urban/rural. In China it is 67/33 and India it is completely reversed at 30/70. And agri continues to be the number one occupation.
Additionally, lack of opportunities is also a problem. India has been focused on services and trailed behind on industrialisation. The current government has been pushing for more industrialisation but they are behind in the curve.
Note that China is where it is because of efforts to do this, on purpose, over decades. 20 years ago, their urban percentage was somewhere in the 40s. We are even seeing more migration to cities in Europe and the US, even though it's unplanned, and it leads to big changes in cost of living thanks to this lack of planning.
So if China took 30 years, give or take, to get to where it's at, with its state capacity, I suspect India will take quite a bit longer.
> Farming in India is still far too labor intensive by world standards. 43% of workers still work in agriculture. [1] For the US, that number is under 2%. China is at 22% as of 2023, and dropping steadily.
Question about farming in India. How much of the process of mechanizing and scaling up agriculture in India is predicated on something like the more widespread use of diesel or other oil/fossil-fuel powered tractors? To replace manual hand labor. If energy costs continue to rise as they are, and all-electric/battery based systems remain costly and out of the reach of the purchasing power of many small to mid sized farmers, what will happen?
Most land holdings are very very small. 1 acre to 5 acres maybe the vast majority. These are all odd sized and shaped and most likely doing different crops based on water availability. To leverage the benefits of mechanization, we need larger land holdings. The farmers have no other ability or income sources, so they hang on it it. Electricity is free. There is no income tax on farming. Govts provide many incentives to get farmers votes. Each state does different things, but they end up copying each others schemes and it gets worse and worse.
Farmers in most regions are no longer poor. Land prices exploded 100 - 500x in a 100 - 150 km diameter around metro areas. Most farmers are now millionaires, yes millionaires in USD. They held on to their land because they didn't know better, the land was useless (no water) and nobody bought it. Now they are going to HODL.
There are mechanization instruments suitable for 5 acre plots, but manufacturers are limited and the potential gains are smaller. Its the same problem as small cheap consumer cars, big investors and capitalists don't want to put money into high volume low margin business with limited market cap. And without mass production costs are significantly higher. Thats why people still maintain 70 year old farmalls and planters and shit spreaders and plows, modern replacements are either crap to be cheap or expensive from low volume.
One difference though is that the agriculture transitions had somewhere for labor to go: factories, construction, urban services, export manufacturing, etc
This transition does too. It's just not clear to most where yet.
He talked about exactly this in the article
3 replies →
Is it a secret you can't share with us?
I see this transition as more like what would happen to livestock if they banned eating meat.
2 replies →
There is no guarantee that this transition will lead to any type of desirable or meaningful job.
Around the time "Bullshit Jobs" was published, more than a third of people said they believed their job was not meaningfully contributing to the world. Graeber goes as far as saying that more than half of white-collar jobs are actually harmful and kept around only because people associate work with self-worth. There is no way that this number will go down with increased automation.
It's not uncommon to hear Boomers say things like "kids these days don't want to work hard anymore. Everyone wants to be an youtuber, no one wants to be a teacher or a doctor or an engineer". Well, guess what? We are heading to a world where being an youtuber might be the only option.
2 replies →
It doesn't. If it did there'd be massive unmet demand for labor in $sector. There is no value for $sector that is currently reporting being short roughly 100 million headcount. So unless you're counting currently non-existence social safety programs or CCC-style government make-work programs that light at the end of the tunnel is an oncoming train.
This time, there is nowhere to move to. And there is no interest in supporting the displaced people in any way.
> Looking at how countries coped with a fast transition from labor intensive agriculture to an urban society gives hints on how an AI transition may look. All the Asian countries that went from poor to rich in a generation did this, with different approaches. How that took place may provide more useful info than philosophy.
There's only one problem with comparing urbanization with the AI transition: there were still jobs that the workers moving from farms to urban centers could do. Instead of planting and harvesting, they made things in factories or became professionals.
The idea of the GenAI bet that most companies are making is that you just don't have people doing work anymore. There aren't any jobs for the laborers to do anymore, at least not ones that are likely to fit their skillset and provide a standard of living that they're used to. If you're a software engineer - one of the higher-paying fields of the last half-century - and get laid off because the c-suite thinks AI can do your job for less, you're going to contract your spending.
The article mentions this, but it doesn't take into account that there will be some work (mainly manual labor) that will face at least some resistance to automation for the next decade. These people will try to get into those jobs, because they have bills to pay. It won't pay six figures. It very well might pay less than it is now due to the glut of candidates who are desperate to make any income at all.
The person who went from catered meals and foosball at the office to framing a house when it's 20 degrees F for a third of the money, none of the future, and a lot more body aches is going to be angry. In a society like the United States, it's the kind of angry that you can't solve with an internal passport system. It's going to mean violence.
Eventually, they'll figure out how to do more manual labor with automated systems. That means that there will be even fewer opportunities.
This is nothing like anything we've seen before, and no one wants to acknowledge that.
> There's only one problem with comparing urbanization with the AI transition: there were still jobs that the workers moving from farms to urban centers could do. Instead of planting and harvesting, they made things in factories or became professionals.
Mostly for export, in the case of the Asian tigers and China. Once wages reached developed world levels, an export-driven economy gradually became harder to sustain, because there was no longer a labor cost advantage. This is why Party leadership talks about "dual circulation", building up domestic demand within China, and about obtaining a technological edge that continues to make exports profitable. There's been considerable progress on both goals, especially in consumer electronics and the auto industry.
> The person who went from catered meals and foosball at the office to framing a house when it's 20 degrees F for a third of the money, none of the future, and a lot more body aches is going to be angry.
Yes. That happened to Egypt, where the government paid for college and then hired the graduates. Then the oil ran out. It's called "elite overproduction" when new college graduates can't get jobs that actually use their education. Already, in the US, about half of college graduates have jobs that don't really need a college education.
The hope is the robot armies will keep the hordes away from the lords' manors. We will see if it works.
System set in place
to slow migration from farms
cities can't absorb
Large population countries / economies reach point where there are more people than jobs, excess people gets dumped into subsistence farming or other inefficient sunk cost make job programs so angry horde doesn't burn it all down. You can replace 95% of subsistence farmers, i.e. billions of people with machinery. Probably replace 95% of knowledge workers, i.e. 100s of millions in OECD with AI, but maybe it's just better/stabler for political serenity for horde to keep generating useless make work email chains. Smaller pop countries can probably meander through for a while specializing in a few high value sectors, larger countries will have to deal with disproportionate idle hands, but also more are in favorable position to exploit / consolidate industrial / resource advances.. Hopefully end game dwindling demographics supported by fully automated luxury communism within sustainable carrying capacity. But there's a lot of probably violent steps between draw some circles and draw the rest of the owl. Ultimately we're likely entering period of placating surplus people and managing demographic relative automation / ai progress.
Data is always nice, but empirical results are literally useless without philosophy to understand and apply them.
That is unless 'we all move to south korea 20 years ago' is an option, I suppose!
> 43% of workers still work in agriculture. For the US, that number is under 2%. China is at 22% as of 2023, and dropping steadily.
Don't you have to contrast these figures with import and export of produce, and environmental/ecological factors? Technology is one thing, but increasing yields by wasting resources (e.g. water, phosphorous, soil erosion, ...) may increase nominal productivity, but not efficiency. Not saying the conclusion is wrong, but I think your numbers are not necessarily causally linked to productivity/efficiency. I mean, the US also has a declining domestic fabrication percentage, but that's not merely indicative of productivity, but mostly outsourcing/loss of capabilities, I think.
Anyway, apparently India also doesn't score very well for food self-sufficiency: https://www.nature.com/articles/s43016-025-01173-4
In the US case the large industrial farms are the ones more concerned with things like soil erosion and fertilizer runoff. Both are things we measure and put a number on what is washing away. Smaller farmers know it can be measured but either are stuck in their ways, or just see that they are making money so they don't care.
My point was the possible disconnect of productivity/yield, and efficiency. A small potato and a large potato likely need the same amount of harvesting work. Now, if you dig up finite reservoirs of phosphorous in Morocco to increase yield in the US, you are not increasing efficiency. That's just planetary debt for a wasteful sprint. I think, in the US this is most evident with water usage, where now many aquifers and reservoirs are spent. Hardly indicative of desirable outcome. Sustainable farming usually has much lower yields inherently. And of course, soil quality and erosion dynamics naturally differ greatly across the globe.
2 replies →
There's no manufacturing sector to employ them if you displace them from agriculture. They'd be displaced into gig economy. This would just increase the population of a handful of metropolitan cities which are already congested. India should fix its cities first.