Unemployment rate: 6.1% (up from 5.4% three months prior)
Notable sector changes:
- Information services: -156,000
- Professional and business services: -134,000
- Financial activities: -78,000
- Retail trade: -45,000
- Healthcare and social assistance: +28,000
- Construction: +12,000
"The acceleration of job losses in knowledge-work sectors continues to exceed projections. While overall unemployment remains below historical crisis levels, the concentration of losses among college-educated workers is unprecedented. The divergence between labor market outcomes for different skill categories warrants close monitoring."
Not asking for sympathy. I knew this was coming eventually. I just didn't think it would be this fast.
Three months ago they rolled out the AI tools as "productivity assistants." We were all told our jobs were safe, that AI would help us do more, not replace us.
Now they've done the math. One person with AI tools can do what five people did before. So they're keeping the one person and cutting the rest.
The severance is decent. The references will be fine. But I've been in marketing for 12 years and I genuinely don't know what I'm supposed to do next. Every job posting in my field either doesn't exist anymore or wants someone who's already expert in the AI tools I just started learning.
I'm 38. I have a mortgage. My kids are 7 and 9. What's the play here?
Options that are actually working for people:
1. Become the AI person. Not just a user-the expert who helps other companies figure this out. There's demand for people who understand both the tools and the business context.
2. Pivot to something AI can't easily do. Trades are obvious. But also: high-stakes human interaction, regulatory compliance, anything requiring physical presence and trust.
3. Go independent. Use the tools to run a one-person operation that does what your whole department used to do.
None of these are easy. All of them are better than waiting for the market to stabilize, because it won't.
Still unemployed. The truth is there are too many of us and not enough landing spots. Some of us aren't going to make it to the other side.
The "winners" in this aren't winning. We're just losing slower.
By Harrison Blake, Senior Fellow, American Enterprise Institute
The numbers are stark: corporate productivity is up 23% year-over-year. Profits are at record levels. And yet median household income has declined for the first time in a decade.
This is the productivity paradox of the AI era. The machines are doing more work than ever. The people are not benefiting.
Classical economics would predict that productivity gains flow to workers through competition for labor and consumer purchasing power. When companies can produce more with less, prices should fall and wages should rise as firms compete for scarce talent.
But talent isn't scarce anymore. The AI systems that have eliminated hundreds of thousands of knowledge-work jobs have created tens of thousands of new ones. The math doesn't balance. And the surplus labor is suppressing wages across the economy, even in sectors not directly affected by automation.
We are witnessing something economists have debated for centuries: genuine technological unemployment on a scale that market forces cannot quickly absorb.
What we do about this is a political question, not an economic one. The wealth is being created. The question is who gets to keep it.
The White House announced yesterday that it's convening an emergency task force on AI-related unemployment. We're joined by Labor Secretary Rachel Martinez. Secretary Martinez, what exactly is this task force meant to accomplish?
Good morning. The task force has three objectives: first, to get accurate data on the scope and pace of AI-driven job displacement. Second, to evaluate existing safety net programs and determine where they're falling short. Third, to develop policy options for the President and Congress to consider.
Critics say the administration has been slow to respond to what many see as a crisis. Why the delay?
I'd push back on the characterization of delay. The administration has been monitoring this closely. But there's a difference between monitoring and intervention. We wanted to be sure we were seeing a genuine structural shift rather than normal labor market fluctuations.
And now you're sure?
The data is increasingly clear that this is structural. The question is what we do about it.
What options are on the table? Universal Basic Income?
Right now, the focus is on retraining programs-helping displaced workers transition to sectors where human skills remain essential. We're also looking at extended unemployment benefits, portable healthcare, and partnerships with community colleges. Some advocates are pushing for more ambitious proposals, but we need to be realistic about what's politically viable in the near term.
What about regulating the AI companies directly? Slowing down deployment?
That's a more complicated question. The technology is global. If we slow down domestic development, other countries continue forward. We could end up with the job losses anyway, but without the technological leadership.
So we're trapped?
I wouldn't say trapped. I'd say we're navigating. And navigation requires careful choices, not panic.
NARRATOR: In a small office in Austin, Texas, a startup called NovaSynth is doing something that would have seemed like science fiction three years ago: operating a profitable company with exactly one human employee.
MARCUS CHEN, FOUNDER: I started NovaSynth eighteen months ago. I had an idea for a data analytics service but no capital to hire a team. So I thought-what if the AI is the team?
NARRATOR: Chen's company uses a network of AI agents to handle everything from customer acquisition to product development to financial management. His role, he says, is more like a conductor than a traditional CEO.
CHEN: I set the direction. I make judgment calls when the agents flag uncertainty. I handle the rare situations that require a legal signature or a human face on a video call. But the operational work-eighty, ninety percent of it-that's the agents.
NARRATOR: NovaSynth reported $2.3 million in revenue last year with operating expenses under $400,000. Chen's personal income exceeded $1.5 million-wealth he's careful to acknowledge came at a cost.
CHEN: I'm not naive about what this means. If one person can do what a team of twenty used to do, that's nineteen people who need to find something else. I don't have a good answer to that. I just know I'd rather be building the future than being displaced by it.
NARRATOR: Critics call this model "wealth concentration disguised as innovation." Supporters see it as inevitable efficiency gains. What's clear is that more companies are following Chen's template. The era of the autonomous organization has begun.
I got my UBI application approved last month.
Not the federal program-that's still being debated. This is California's pilot, available to residents who meet certain displacement criteria. $1,400 a month, no strings attached.
It's not enough to live on in the Bay Area. It's not even close. But combined with the freelance work I can still find, it keeps me housed. It keeps me fed. It keeps me from the slow spiral of debt and desperation I've seen consume others.
I used to be ashamed of it. Government assistance, the safety net, whatever you want to call it. I grew up with the American mythology that hard work equals success, that needing help means you didn't try hard enough.
I don't believe that anymore.
I worked hard. I have two degrees. I spent fifteen years building skills that were valuable right up until they weren't. The thing that changed wasn't my effort or my talent. The thing that changed was the technology.
The UBI doesn't make me feel like a failure. It makes me feel like someone who got caught in a wave that was bigger than any individual swimmer. The shame isn't in accepting help. The shame would be in a society that creates this much wealth and lets people drown.
I don't know what comes next. Nobody does. But I know that the old stories-work hard, get ahead, build a career-those stories were written for a different economy. We need new stories now.
Maybe this is one of them.
This journal does not typically engage with economic policy. But the events of the past year have made clear that scientific progress and social stability are not separate concerns.
The AI systems transforming research-enabling breakthroughs in protein folding, materials science, drug discovery, and a dozen other fields-are the same systems eliminating millions of jobs. The researchers celebrating new capabilities are the same citizens watching their neighbors lose livelihoods.
We cannot, as a scientific community, pretend these are separate phenomena.
The rate of AI capability advancement continues to exceed projections. Benchmarks that were expected to take years are being saturated in months. Research directions that seemed speculative are becoming practical. The systems are better than we expected, faster than we expected, and more general than we expected.
This is, from a purely scientific standpoint, remarkable. But science does not exist in a vacuum. The same acceleration that enables our breakthroughs is outpacing society's ability to adapt.
We call on our colleagues to engage seriously with this tension. The technology we are building will shape the world our children inherit. We have a responsibility to ensure that world is one worth inheriting.
They're being cagey about the details, but reading between the lines: this is probably the learning/generalization work they've been hinting at. If they've actually cracked continuous learning without capability loss...
> If they've actually cracked continuous learning without capability loss...
Then what? Genuinely asking. What changes?
Everything. Right now models are smart but frozen. They can't really learn from experience, can't develop expertise in your specific problem over time. If that changes, the models stop being tools and start being... something else.
Collaborators maybe. Colleagues. The thing sci-fi has been promising for decades.
Or the thing dystopian sci-fi has been warning about. Depends on your perspective I guess.
Either way, we're about to find out. The announcement is scheduled for next month. Shiftman's doing the keynote himself.
Curious timing. Right before the Senate hearings on AI unemployment. Almost like they're trying to control the narrative.
That's exactly what they're trying to do. "Look at this amazing breakthrough" is a much better headline than "look at all the jobs we're eliminating."
Welcome to the future. It's complicated.