AI: The World Will End Yesterday. Plan Accordingly.
Well, if you watch the artificial news, Artificial Intelligence is going to take your job and your girlfriend, at least when it’s not too busy taking over some bunker in North Dakota and launching nukes willy nilly. You could form the opinion that AI already has taken over the world. You can’t turn on anything on these here intertunnels without some demented form of Clippy the AI assistant offering to correct your grammer, and maybe write that email for you that you’ve been meaning to send, but you can’t for the life of you remember how to spell Deer Sur.
There have been many, many laundry lists published of all the jobs that are going to be wiped out by one chatbot or another. Most everyone outside of longshoremen and prostitutes are slated to be standing on streetcorners holding signs that read: Will photoshop the background out of pictures of female footwear for use on your Shopify store for food. The usual commentarazzi are furiously analyzing the inroads that Large Language Models (LLMs) are making into the economy, and publishing their search engine optimized articles, written by ChatGPT, natch, with headers like: AI: The World Will End Yesterday. Plan Accordingly.
What is missing is some form of sober analysis. Just adding a new definition of slop to the dictionary isn’t helpful, any more than adding a new definition of vaccine kept you from getting the flu. I’m interested in the topic, however, and I finally found one lonely source that at least attempted to answer the only cogent question:
HOW ADAPTABLE ARE AMERICAN WORKERS TO AI-INDUCED JOB DISPLACEMENT?
I remember the good old days on the intertunnel when I’d have to warn you that the link goes to a PDF. It’s 2026, I think. I’m never sure until about February. If it is indeed 2026, I think you should have gotten over your fear of Adobe Acrobat by now. I suppose I could skip the warning about the format of the document, and offer a more timely warning for today’s internauts: It’s not only a PDF, it’s a 54-page working paper from a think tank, and it’s got a lot of words, some of them polysyllabic. It’s likely your lips will get really tired while reading it. It’s got numbers in brackets all over it, too, which I think lead to footnotes at the end. I can’t be sure, I never get that far without my eyes glazing over.
The working paper is from NBER. That’s an acronym for the National Bureau of Economic Research. They’re a think tank in Cambridge, Massachusetts. The locale makes my spider sense tingle. That zip code is ground zero for educated lunatic worldviews. But NBER doesn’t appear to be a big building full of Sovereign Citizens or people eating avocado toast and plotting to dye their hair pink or anything. It’s a loose agglomeration of academics and public policy wonks that seems at least modestly open-minded. So I figured it might be worth the time it took to read the report. Honestly, the question itself, how adaptable are American workers to AI-induced job displacement, demonstrates some clear thinking from the get-go. It’s long past time to stop arguing whether LLMs are real, or here to stay, or bankrupting only themselves or the whole nation, or useless, or whatever. LLMs are real, and they’re spectacular, sorta. Let’s move on. Whose ox is gonna get gored?
The paper doesn’t have a monomania for simple exposure to AI, which is great, because AI is exposing itself in more places than Hunter Biden. That ship has sailed. They’ve come up with an Adaptive Capacity Index, to see how well many types of workers will be able to adapt themselves to the new workplace now that LLMs rule most every roost. The analysis is interesting.
First they predict (or observe, really, at this point) the potential for tasks in an occupation to be affected by AI. Then they measure the Adaptive Capacity of that guy that always takes the last donut in the break room, and everyone like him. Adaptive Capacity is an amalgam of workers’ ability to adjust after the modification of their jobs (or outright displacement) caused by AI. It includes factors like liquid financial resources, skill transferability, geographic labor market opportunities, and age distribution within occupations. So far, so good. The index they came up with covers 356 occupations. They claim that’s about 96% of U.S. workforce. That’s a lot. I’ll admit I jumped to conclusions earlier, and I’m not really sure if longshoremen or prostitutes might be included after all.
If you’re of the USA Today generation, they’re looking out for you. First, a bubble chart:
If you’re of the Facebook generation, don’t worry. They’ve got a map showing the distribution of the population that isn’t expected to survive the chatbot apocalypse. It has colors and a thermometer, rendering it still more fascinating:
Got that? If you live where the buildings are tall enough to cast shadows, you’re in danger. If you live in New Mexico, you already knew you were in danger, just by looking out the window. A failed state, that.
They’ve got lists, if you’re from the Tumblr generation. Who’s got high exposure, but high adaptability to boot? Here goes:
So much for all the news blurbs about software developers and various other computer nerds being put out of a job by chatbots. They’ve got the highest exposure, and the highest ability to adapt to that exposure.
So who’s on the other end of the spectrum, and the dookie stick? Who’s getting Skynetted first? Here’s who really needs to adapt, but won’t be able to:
I suppose it would be impolitic of me to mention that there are several job descriptions on that last list that I’d like to sign up for manned missions to the surface of the sun.
Once the report has identified the problem, they go on to mention the only solutions anyone ever mentions. The government has to step in with retraining and handouts for these benighted souls flummoxed by ChatGPT. One can’t help but notice that a lot of those job descriptions are more or less either government jobs, or private sector jobs made necessary only by government regulation. Retraining? Handouts? It’s a maladaptive snake eating itself, and turtles all the way down. Maybe they can all open daycare centers in Minneapolis. It pays well, I hear.
One is also tempted to observe that the people on the first list are preponderantly male, and the second list is loaded with the distaff set, and in many cases, just plain loaded. I was tempted, but I got over it. So I won’t mention it, or parallel parking, or any other divisive topic.
See? I’ve adapted to the internet. It has girls on it now.

Recent Comments