The AI job losses are here.
How bad will it get.
And where will the new jobs come from.
Last week Salesforce.com disclosed that AI is now doing between a third and half of its internal work. They've already reassigned 500 workers to other roles, and laid off 1,000.
Microsoft plans 15,000 layoffs -- including 6,000 software developers -- citing AI.
Clippy's revenge.
And so much for learn to code.
Visa cut 1,400, Facebook 3,600. Intel 5,000. Even Chevron and British Petroleum cut a combined 17,000, citing AI.
In all, the Wall Street Journal estimates US public companies have laid off 3.5% of their workers over the past 3 years.
Interestingly, they've laid off 4.5% of executives and over 6% of middle managers.
AI hitting white collars shouldn't be a surprise; Mike Rowe, host of Dirty Jobs, has long predicted that AI would wipe out white collars, with electricians and pipefitters inheriting the earth.
Still, there's a lot of white collars. McKinsey says AI could affect 12 million jobs by 2030 -- that's 2 million per year. Brookings predicts 36 million. The OECD estimates one in four jobs are at risk of automation.
Automation since the Dawn of Civilization
Now, this all sounds dire.
But remember this isn’t our first rodeo: Humanity has faced comparable waves of automation many times, from agricultural mechanization to the internet in the 90's -- which was also supposed to wipe out the jobs.
In fact, tech making humans useless has been predicted for literally thousands of years: Aristotle worried about ox-plows replacing agricultural workers. Medieval Europe had a jobs scare over... watermills.
And it never happens.
The reason it never happens human wants are unlimited. So as tech makes us richer, whether its ox-plows or supply chains, activities that used to pay little or nothing —because you were selling to relatively poor people — now pay enough to be a “job.”
Consider actors used to be in line with prostitutes, now they're millionaires.
Or that personal trainers in high-automation America make 10 times what they make in low-automation Honduras.
In short, automation is an escalator: you take one step down, but earnings for everything else goes up even faster. For thousands of years and counting.
What are the New Jobs?
So what will be the new jobs after AI?
Easy: Anything where human presence is the value.
Consider that hand-made items cost more than factory-made even though hand-made almost always has more errors.
Or that vending machines have existed for a century yet bartenders per capita has tripled since 1900.
Or that in the 1960's stadium concerts barely existed -- there were 3 between 1965 and 1970. Today there's 500 a year despite the fact you can watch concerts for free on Youtube.
In 1970 there were a dozen comedy clubs in the US. Now there's over a thousand.
So we're talking bartenders and youtubers, nurses and caregivers, people who make things by hand.
Therapists, tutors, private chefs, travel guides.
Stylists, decorators, personal trainers, dance teachers, diet coaches.
Here’s a handy list from the US Career Institute:
What’s Next
In short, AI will replace today’s cubicle farms with an economy made of humans.
And when the dust settles we're probably looking at a temporary half-percent rise in unemployment while people shift into new jobs.
Jobs that are more soul-nourishing, whether hand-building chairs or tending bar. And, going by 2,000 years of history, jobs that ultimately pay much better specifically because of the background automation.
In the near term this transfers wealth to blue collars -- and away from journalists.
In the long run it makes our economy far more humanized than the soul-crushing cubicle farms that future generations will look upon with pity.
Every week I write a newsletter on the top stories in economics with regular deep dives into history like the Fall of Rome, the Weimar hyperinflation, or FDR’s Great Depression.
Subscribe for free, or choose $5 monthly to help keep us on the air. And Thank You to current subscribers!
I also make short daily videos on economics and freedom:
Roundup Podcast of the week’s videos on Spotify, Apple, and Google









With your degree in "oiling and lubrication", all those fancy musk-made robots are going to need you so they don't become paralyzed and rusty. It is a fantasy that A/i will replace most jobs. It is far too retarded mentally and is emotionally distant. Heck, the silly A/i customer service goons are some of the dumbest creations ever.
In the vertical infographic: "65 Jobs With The Lowest Risk of Automation", I am surprised to see that "Choreographers" is number-two, ranked for projected growth by 2031. HAHA! So, come the Great Reset we'll all want to hire a dance teacher to help us do viral shimmies to up our social credit score so we can spend our $CBDC? I mean, I'm all for dancing (for its own sake, with humans, and not for social approval) but what an odd profession to take the silver medal. Joking aside, I find this hard to believe – nobody dances anymore, what the hell is this?
That aside, great article, and I love the historic and economic analytic lens you regard the future with here. Your predictions line up mostly with my layman's intuition, but regarding the situation from the ground (i.e. inductively looking a special cases, rather than deductively from models or theories) seems to foretell a different horizon. Take the handmade instrument professionals "Luthiers" for example – there was a spurt of Great American Luthiers from about 1980-2010, but they are retiring and dying off, and nobody is filling their ranks. Very few people play (acoustic) music anymore; only niche weirdos (like myself). We've become almost entirely consumers of music and not "players", hence the "makers" (i.e. luthiers) going out of business. Similarly we have increasingly become consumers of sex (bring back the players!) Same goes for sport (we watch it, but seldom play casually within our community). You get the idea.
All in all, I fear we are being shunted into becoming exclusively consumers. Very few people know how to make anything that doesn't require social approval to know whether it succeeded or not – ya know? I cannot see the current trendline radically changing course come AI, rather—only getting worse. And that current trajectory is as Toby Rogers cynically sums up: "We used to make things, now we just make people sick and profit from that."