I’m willing to buy your whole argument. I graduated into the start of the recession in ‘07 and it was brutal, I didn’t really get on my feet until ‘12, so I lived the pain you’re describing.
Non competes are terrible, we should ban them. I’m good with stronger anti-trust and more dynamism.
But I’m concerned at the emphasis on interests rates, not just in your piece, but everywhere. Because ZIRP was *really bad!* and I think we’re all still dealing with a hangover from having interest rates historically broken for a very long time.
A 6% mortgage is not *historically* expensive, it’s just *recently* expensive. House prices mechanically ratcheted up, in part because, since the historically high rates of the 80s, every drop in rates meant the seller could demand more cash for the same monthly payment.
Also, zero interest fuels consolidation and crushes dynamism. At 5-8% interest only real businesses can work. At 0% you can get VC money to scale up an imaginary business for years, as long as it has good vibes.
I think we are still dealing with the longer-run pain of having to shift back to a real economy after ZIRP, but we *need* to make that transition.
To your broader point, though, I very much agree that the social consequences for young workers are brutal and we should do more to help them.
To my mind, banning non-competes is a no brainer. As a multi-time entrepreneur, I’d also love to see health insurance decoupled from employment, with a reasonably priced government option, because health benefits are the number one thing where I can never compete with big companies.
What else can we do to create more economic dynamism that can thrive within historically normal interest rates?
To push back against interest rates and housing, I think supply is a better explanation. If we go solely off your theory, prices should have fallen, but they didn’t. Housing has continued to rise even as interest rates rise. The same phenomenon occurred in the 1980s when supply really started to be constrained. If anything, there just need to be more houses built. At this point, I would argue housing is weakly interest rates sensitive. Prices fell in 2008-2012 because there was so much supply even though rates fell that whole time too.
Re: supply in 2008 you should read Kevin Erdman’s work, he will show you the data that there was no oversupply, and instead the housing crash was caused by the government greatly reducing who was eligible for mortgages compared to historic norms (ie they way overcorrected against subprime lending).
Re: supply, yes we need more supply. But housing has more than one variable. My argument is only that interest rates are *part* of the problem, which we see in thousands of news stories and anecdotes of people not being willing to move even though they want to, because they aren’t willing to let go of their low mortgage rate.
The WARN Act detail is the piece that should end the debate. Companies cite AI freely in earnings calls and press releases where the audience is investors who reward "efficiency narratives." But when the same companies file legally binding layoff notices in New York, where they now have an explicit checkbox for "technological innovation or automation," zero out of 160 checked it. That gap between the investor story and the legal record tells you everything about what AI displacement actually is right now: a communications strategy, not a labor market event.
The Iscenko-Millet thought experiment deserves more attention too. A hiring freeze with zero layoffs mechanically produces a 25% decline in entry-level headcount within one year. You don't need a single line of code to be automated. You just need the Fed to raise rates in March 2022 and watch the math play out. The fact that the same occupational pattern appears in 2020, when generative AI wasn't even theoretically available, should be dispositive.
From LATAM the view is even sharper. The same "AI is replacing jobs" narrative is being imported wholesale into markets where enterprise AI adoption is closer to 3-4%, and where rate sensitivity and capital flight do far more damage to youth employment than any chatbot. The wrong diagnosis travels faster than the evidence.
I agree AI seems like an excuse to fire people overhired during the pandemic.
I agree with you on non-competes, they should be universally illegal. I found my own annoying but mostly something I could get around. I did turn down an offer from amazon once over a very restrictive non-compete.
I'm indifferent on whether resources shocks should affect Fed policy.
However, I think their tightening cycle post pandemic was correct and they never should have lowered rates during the pandemic. That was a huge mistake. I don't think rates should be set artificially low to try and achieve 3% unemployment.
And I don't think firms are wrong not to make large potentially bad ROI investments in energy infrastructure every time there is a potentially temporary price spike.
As you know I'm deeply against the Iran war.
On firm concentration I'm generally against the idea that we should let the Elizabeth Warrens of the world run around calling our best companies monopolies and trying to tear them down.
"These are not organizations with 50 to 75 percent surplus headcount. They are organizations where management layers, internal politics, and misallocation have made it hard to ship anything well. The people are not unnecessary. The structure wastes them."
This is a terrible argument. If firms can't find ways to use people, then they are overstaffed.
I would even go further to say that having too many people actually decreases productivity in a lot of circumstances. The best teams I've ever worked on have been lean.
Your structural case is compelling. Four decades of eroding job mobility, rising employer concentration, and noncompete proliferation have done genuine damage that no amount of AI literacy training will fix. Your debunking of AI-as-culprit is thorough and well-evidenced. But there is a blind spot sitting right at the center of the piece, and it is visible in the title.
Featuring interest rate hikes as the headline villain while acknowledging they were necessary is an odd choice. Blaming the Fed while admitting it had no choice is not really an argument, it is just describing a difficult situation. It also buries the more uncomfortable point your own evidence is making.
That point is sitting in your own evidence. Oil executives are pledging buybacks at $200 oil, a deliberate decision to return cash to shareholders rather than invest in workforce. Hiring managers told Resume.org they blame AI specifically because it protects the stock price. Companies are redefining "entry-level" to mean mid-career because it is cheaper than developing junior talent. You document all of this, then decline to name it for what it is: corporate decision-making that compounds the structural damage, not a mechanical response to forces beyond anyone's control.
You are willing to call out concentration and noncompetes as policy failures requiring intervention. You are considerably less willing to say that the people running these organizations are making decisions - about shareholder returns, workforce development, and public communication - that compound the damage. That asymmetry is hard to justify. Structural forces require legislation to fix. Managerial priorities require only a different decision.
The scarring research you cite is genuinely alarming. The cohort entering the labor market now will carry the consequences for decades. They deserve a diagnosis that follows the evidence all the way to where it points — rather than stopping just short of the people making the calls.
If noncompetes are suppressing wages/salaries, that should mean more employment for new grads, not less. If employees cost less, more of the marginal ones should generate revenue greater than the cost of employing them.
Well not necessarily. Young people aren’t job switching as often because of noncompetes. This reduces opportunities for new grads because the old grads are still there
I’m willing to buy your whole argument. I graduated into the start of the recession in ‘07 and it was brutal, I didn’t really get on my feet until ‘12, so I lived the pain you’re describing.
Non competes are terrible, we should ban them. I’m good with stronger anti-trust and more dynamism.
But I’m concerned at the emphasis on interests rates, not just in your piece, but everywhere. Because ZIRP was *really bad!* and I think we’re all still dealing with a hangover from having interest rates historically broken for a very long time.
A 6% mortgage is not *historically* expensive, it’s just *recently* expensive. House prices mechanically ratcheted up, in part because, since the historically high rates of the 80s, every drop in rates meant the seller could demand more cash for the same monthly payment.
Also, zero interest fuels consolidation and crushes dynamism. At 5-8% interest only real businesses can work. At 0% you can get VC money to scale up an imaginary business for years, as long as it has good vibes.
I think we are still dealing with the longer-run pain of having to shift back to a real economy after ZIRP, but we *need* to make that transition.
To your broader point, though, I very much agree that the social consequences for young workers are brutal and we should do more to help them.
To my mind, banning non-competes is a no brainer. As a multi-time entrepreneur, I’d also love to see health insurance decoupled from employment, with a reasonably priced government option, because health benefits are the number one thing where I can never compete with big companies.
What else can we do to create more economic dynamism that can thrive within historically normal interest rates?
To push back against interest rates and housing, I think supply is a better explanation. If we go solely off your theory, prices should have fallen, but they didn’t. Housing has continued to rise even as interest rates rise. The same phenomenon occurred in the 1980s when supply really started to be constrained. If anything, there just need to be more houses built. At this point, I would argue housing is weakly interest rates sensitive. Prices fell in 2008-2012 because there was so much supply even though rates fell that whole time too.
Re: supply in 2008 you should read Kevin Erdman’s work, he will show you the data that there was no oversupply, and instead the housing crash was caused by the government greatly reducing who was eligible for mortgages compared to historic norms (ie they way overcorrected against subprime lending).
Re: supply, yes we need more supply. But housing has more than one variable. My argument is only that interest rates are *part* of the problem, which we see in thousands of news stories and anecdotes of people not being willing to move even though they want to, because they aren’t willing to let go of their low mortgage rate.
The WARN Act detail is the piece that should end the debate. Companies cite AI freely in earnings calls and press releases where the audience is investors who reward "efficiency narratives." But when the same companies file legally binding layoff notices in New York, where they now have an explicit checkbox for "technological innovation or automation," zero out of 160 checked it. That gap between the investor story and the legal record tells you everything about what AI displacement actually is right now: a communications strategy, not a labor market event.
The Iscenko-Millet thought experiment deserves more attention too. A hiring freeze with zero layoffs mechanically produces a 25% decline in entry-level headcount within one year. You don't need a single line of code to be automated. You just need the Fed to raise rates in March 2022 and watch the math play out. The fact that the same occupational pattern appears in 2020, when generative AI wasn't even theoretically available, should be dispositive.
From LATAM the view is even sharper. The same "AI is replacing jobs" narrative is being imported wholesale into markets where enterprise AI adoption is closer to 3-4%, and where rate sensitivity and capital flight do far more damage to youth employment than any chatbot. The wrong diagnosis travels faster than the evidence.
I agree AI seems like an excuse to fire people overhired during the pandemic.
I agree with you on non-competes, they should be universally illegal. I found my own annoying but mostly something I could get around. I did turn down an offer from amazon once over a very restrictive non-compete.
I'm indifferent on whether resources shocks should affect Fed policy.
However, I think their tightening cycle post pandemic was correct and they never should have lowered rates during the pandemic. That was a huge mistake. I don't think rates should be set artificially low to try and achieve 3% unemployment.
And I don't think firms are wrong not to make large potentially bad ROI investments in energy infrastructure every time there is a potentially temporary price spike.
As you know I'm deeply against the Iran war.
On firm concentration I'm generally against the idea that we should let the Elizabeth Warrens of the world run around calling our best companies monopolies and trying to tear them down.
"These are not organizations with 50 to 75 percent surplus headcount. They are organizations where management layers, internal politics, and misallocation have made it hard to ship anything well. The people are not unnecessary. The structure wastes them."
This is a terrible argument. If firms can't find ways to use people, then they are overstaffed.
I would even go further to say that having too many people actually decreases productivity in a lot of circumstances. The best teams I've ever worked on have been lean.
Yes to the lean part. Too few people have read the Mythical Man Month ! 😎
I would argue that the root cause is even more basic: an economy based on fiat currency, aka: monopoly money.
I wonder how much remote work and outsourcing played a role?
My company fills entry level roles with Brazilians and Mexicans.
Thank you AI for publishing this
I keep coming back to this shift:
“Scaling” becomes “bloating”
the moment people are no longer needed.
Same companies. Same systems.
Just a different role for labor.
To me, that says something quite uncomfortable:
work isn’t valued in itself,
only in relation to what the system needs.
And this line keeps coming back to me:
“The wrong diagnosis leads to the wrong treatment, and young workers cannot afford the delay.”
I think that’s exactly it.
If we keep measuring the wrong things,
we’ll keep solving the wrong problem.
I’ve been exploring that here:
https://saraeson.substack.com/p/what-we-measure-what-we-miss?r=1t83gn&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true
Your structural case is compelling. Four decades of eroding job mobility, rising employer concentration, and noncompete proliferation have done genuine damage that no amount of AI literacy training will fix. Your debunking of AI-as-culprit is thorough and well-evidenced. But there is a blind spot sitting right at the center of the piece, and it is visible in the title.
Featuring interest rate hikes as the headline villain while acknowledging they were necessary is an odd choice. Blaming the Fed while admitting it had no choice is not really an argument, it is just describing a difficult situation. It also buries the more uncomfortable point your own evidence is making.
That point is sitting in your own evidence. Oil executives are pledging buybacks at $200 oil, a deliberate decision to return cash to shareholders rather than invest in workforce. Hiring managers told Resume.org they blame AI specifically because it protects the stock price. Companies are redefining "entry-level" to mean mid-career because it is cheaper than developing junior talent. You document all of this, then decline to name it for what it is: corporate decision-making that compounds the structural damage, not a mechanical response to forces beyond anyone's control.
You are willing to call out concentration and noncompetes as policy failures requiring intervention. You are considerably less willing to say that the people running these organizations are making decisions - about shareholder returns, workforce development, and public communication - that compound the damage. That asymmetry is hard to justify. Structural forces require legislation to fix. Managerial priorities require only a different decision.
The scarring research you cite is genuinely alarming. The cohort entering the labor market now will carry the consequences for decades. They deserve a diagnosis that follows the evidence all the way to where it points — rather than stopping just short of the people making the calls.
If noncompetes are suppressing wages/salaries, that should mean more employment for new grads, not less. If employees cost less, more of the marginal ones should generate revenue greater than the cost of employing them.
Well not necessarily. Young people aren’t job switching as often because of noncompetes. This reduces opportunities for new grads because the old grads are still there
Doesn’t follow. Old grads not switching jobs doesn't decrease the overall pool of net-profitable jobs.
Hm I’m not sure I follow. It still reduces early career opportunities and explains experience inflation
All great ideas that both parties should be able to implement!