Not to be too nitpicky here, but given the hours they work and the profit margins, most dairy farmers almost certainly do this work for an effective wage less than that amount. I think most farmers are quite aware how little money there is to be made in farming.
Ah no sorry for the scare quotes. Was just meaning to link it what you are saying.
it is physically demanding, but worse is LONG hours (and early hours, milking starts at 5). I think whats also bad about the way they are employed now is that US ag doesn’t have upward mobility (promotion to manage)
I was more just pointing out that, thankfully, the people employed in this story weren't being grossly mistreated.
Personally, I have worked on a dairy farm for that wage, plus done the "backbreaking labor" required. I would be happy for someone to be paid more than that!
Cover of the book "Sweet Tyranny" by Kathleen Mapes
The reason why we only let immigrant workers in for a temporary time is a bigger question. I assign my students to read a chapter from this book, "Sweet Tyranny," which discusses its origins.
It seems mostly to be a compromise between the eugenics-loving progressives and farms that needed labor.
The dairy industry is one of the most dependent on undocumented workers, according to David Bier, an immigration expert at the Cato Institute. Visas are available for farm work, but only if it is seasonal. Rank-and-file dairy workers do not qualify because of the year-round demands, despite the industry’s efforts to get them included.
In the article I think it tries to explain it here but doesn't do the best job.
There really is no legal way to hire year-round agricultural labor, as the only legal pathway is the H2A program (which doesn't allow year round work).
This is basically the "fix" most ag wants: a legal way to hire.
Workers can make in an hour what they made in an day in Mexico. The O’Harrows pay between $14 and $20 an hour. The minimum wage in Mexico is about $17 a day.
Not to imply that ag workers aren't mistreated, but this article does not indicate the workers were underpaid for their work.
$14-$20/hour is typical for ag work, and far above the minimum wage for Wisconsin.
IMO one of the most useful ways to think about LLM use-cases is "Would someone in Star Trek use the ship's computer to do this? If they did, would the episode be about how that's a lame thing to do?"
Photo from the train which is empty
also the beautiful train is almost completely empty.
Picture of my TexRail ticket (2 DOLLARS) out the window from my seat (which has a table)
Shout out to TexRail.
I’m in Fort Worth for a conference and I almost took a 60 dollar Uber to sit in DFW traffic, instead of this lovely train that picks up in the airport and goes downtown for 2 DOLLARS.
Ha, what a fun job!
I get the sense that most of their default beliefs are "farmers are doing important work and they should be protected" and don't examine the nuances beyond that. I expect this is the case for many people in the US.
What I find most surprising is that this does not really depend on ag background. I see the suburban kids from Chicagoland area be just as ardent in their support for protecting ag interests as the ones from farming backgrounds.
Very few of them brought up the cost of food, which I thought was very interesting.
At least with my undergrads, I find that a portion of them are not swayed even when they see the evidence. For those, it seems that work requirements is a moral position.
The students also didn't seem to appreciate that farmers would almost certainly still stay in business. The government support programs would make sure of that.
One of their other worries is the impact on the biofuel industry, which employs a lot people here in Illinois.
My guess is also that there's probably a difficulty imagining a Midwest that does not grow this much corn.
Note that I frame this by citing Lark et al (2022), which I mention in lecture as a study showing negative environmental effects of RFS. In their own readings of more evidence, they mostly don't contest this.
They also bring up that they want to protect biofuels in general, including cellulosic.
@mikegrunwald.bsky.social you have your work cut out for you.
Another note on something students mostly wont condemn: the Renewable Fuel Standard.
Just graded their midterms where they were asked to evaluate the effect of taking away RFS, and most of the answers were "yes this hurts the environment, but it helps farmers and gas prices so it should stay."
Agree.
I’ve had proponents of work requirements deliberately misinterpret my work to support their claims.
It’s not from lack of evidence or knowledge about evidence, it’s twisting evidence in favor of policies they want to support anyway.
donmoynihan.substack.com/p/what-aei-g...
I think this is particular disheartening to see among students who will eventually be the next generation of voters and policymakers.
... They end up saying something like "I know this doesn't increase employment, but people should still have to work to get SNAP."
In other words, it is a MORAL argument they are making, not an economic one. They don't care if it fails its purpose, because it enforces a moral objective.
I teach an undergraduate course on agricultural and food policy. In recent years, their assignment has been assessing the impacts of work requirements on employment.
What is a bsolutely fascinating: a certain portion STILL support work requirements, even when they see research like this. Why? ...
I’m pretty uninformed here, but isnt the massive training data what makes Claude and ChatGPT so useful? taking away that corpus and running it locally, how would it ever compete?
So im not sure I completely understand, as I thought the whole deal was that these models are good because of their massive training data. How are they going to be as good as the mainstream products being only a local install without this massive corpus?
* generally right. I suppose and LLM would have caught that if I had been using one :p
Again, not trying to be too abrasive to @paulgp.com, as I think he is generally write that the tech is useful. I'm just not convinced it's net positive just yet.
Note that this is not theoretical. We already have work showing us that this is the case:
www.sciencedirect.com/science/arti...
4) Offloading writing is NOT automatically a net positive. I see this with my students constantly. The "banging your head against the wall" process is essential to connecting concepts. Bypassing it with LLMs cheats everyone of this essential learning process.
3) on the point of using it to do reviews: is this really going to align with journal's privacy positions? Did I, as a reviewer, sign up to have my words used as training data for Anthropic or OpenAI?
Given, I think the tech is useful for this kind of task, but the frameworks have not caught up.
