Dear Reader,
The August jobs report came out Friday morning.
162,000 new jobs.
Wall Street was looking for 53,000.
So roughly triple the estimate, and the strongest month since March.
The unemployment rate sat right where it was, at 4.1%.
Every headline I saw called it strong.
They were right. It was.
Then I did what I always do, which is read past the headline.
Let me tell you what's actually in there.
Where The Jobs Came From
Bars and restaurants led the country in job creation.
Manufacturing added 16,000. Health care added 13,000.
And information lost 23,000.
Information means computing infrastructure, data processing, publishing, broadcasting. The white-collar work. The jobs that ask for a degree and pay for one.
So in the month of August, the American economy hired people to pour drinks and let go of people who write code.
Now, ordinarily this is the point where I'd tell you not to build a worldview out of one BLS release.
One month of data is one month of data. I've watched people forget that and then spend two years defending a position they took on a Friday morning.
But I can't give you that caveat here, and I want to explain why.
Because information hasn't lost jobs for a month.
It has averaged a loss of 8,000 jobs a month for the last twelve.
That's a year.
And look at what the rest of the report did in that same stretch.
June got revised up to a gain of 31,000. July, which was first reported as a loss of 23,000, was revised to a gain of 21,000.
The average month over the past year added 31,000 jobs.
So the labor market didn't limp into August and get lucky.
It has been quietly fine this whole time.
Which is exactly what makes the information number impossible to wave away.
You can explain away a shrinking sector in a shrinking economy.
This isn't that.
The Part Nobody Wants To Say Out Loud
The explanation being offered for those losses is artificial intelligence.
I've spent the better part of two years reading about what AI was going to do to the labor market.
Most of it was speculation. Consultants with slide decks. Projections for 2030.
This is different.
This is the Bureau of Labor Statistics, in a routine monthly report, showing one category of employment going down while nearly everything else went up.
That isn't a forecast.
That's a receipt.
And I want to point out something about that receipt, because I think most people will read it backwards.
It does not tell you that AI is profitable.
It tells you that AI is already good enough that somebody was willing to trade a salary for it.
Those are two very different statements.
The second one is the one that matters to an investor.
What It Costs To Replace A Person
Here's what I mean by that.
When a company lets a programmer go and buys software instead, that money doesn't disappear.
It moves.
It moves to whoever sold the software. And then most of it keeps moving — to whoever runs the data center, and to whoever keeps that data center supplied with electricity.
A model doesn't work for free.
It works for power.
Every one of those substitutions quietly converts a wage into a hardware order and an electricity bill.
That's the part of this story I want to own.
Not the company promising to change the world.
The company that gets paid every time somebody else tries to.
Which brings me to what I've been working on.
On January 2, 2027, the federal government pulls the plug on Elon Musk's supercomputer.
He has one way out, and it runs through a single small company.
He does not have a second option, because there isn't one to have.
I put the whole story together here>>>
The Bottom Line
I'm not going to sit here and tell you a jobs report predicts anything.
It doesn't.
What it does is tell you what already happened.
And what already happened is that the broad American labor market held up just fine for a year — while the most educated slice of it got smaller every single month.
You can argue about whether that's AI.
You should argue about it.
What you shouldn't do is call it a rounding error and then find out in three years that it was the whole story.
Watch the composition, not the headline.
The headline was written for people who'll have forgotten it by Monday.
Have a wonderful Saturday.
I'll see you tomorrow.
"The Buck Stops Here,"
Found this helpful? Share it with others.
Written by Dylan Jovine