Maybe you, like me, read Matt Shumer’s viral piece about AI eliminating 50% of white-collar jobs in the next 1-5 years and maybe you, like me, thought… Oh. Isht.
It’s well-crafted, detailed, and structured to make you feel like you’re getting the inside scoop from someone who knows. (And that everyone else is sleepwalking into disaster.)
It’s also designed to make you panic.
I’m here to say: Do not panic. Let’s talk about it.
First things first: I’m not dismissing AI. I use AI. I think the technology is remarkable. It will continue to improve—of course it will. Matt is right about a few things, including that most people underestimate how much has changed in the past year.
But at the same time… he’s not right about everything. And he’s asking the wrong questions. The questions you ask determine the life you build.
The COVID Trap
Matt opens by comparing February 2026 to February 2020—remember when you thought people were overreacting about a virus? And then the world shut down?
In some ways, it’s an incredible analogy (I love a good analogy), effective because it’s a gut-punch that triggers a visceral feeling inside us all: I don’t want to be the person who dismissed the warning signs again.
But if you think more about it, it’s a terrible analogy.
COVID works on the surface as a comparison because, yeah, it was a once-in-a-century event. Global pandemics that shut down the entire world almost never happen. It surprised and shocked everyone.
“This AI moment is as rare and catastrophic as a global pandemic,” is what Matt says.
Okay. But let’s unpack it a little more….
COVID had a clear, measurable, physical mechanism. Virus spreads. People got sick. Hospitals overflowed and set up sick bays in their parking lots. Governments shut things down. Restaurants set up tables outside in freezing weather; we ate pasta carbonara with gloves on. We hoarded toilet paper and hand sanitizer. You could see it happening in real time, country by country.
AI job displacement is much murkier. When does it “happen”? When the technology is capable? When companies adopt it? When people actually lose jobs? Those are very different timelines, aren’t they? Matt mashes them all together into one formless blob and plops them out onto a Panic Conveyor Belt.
He’s saying: “The technology can do the work now… ergo, the jobs are gone!”
But that skips about five major steps of economic and social change.
But Capitalism
But capitalism will force adoption, you might think. In other words: companies will adopt AI to save money, creating competitive pressure that forces everyone else to follow or whither away.
But that assumes “cheaper and faster” always wins. That’s not actually true—especially not in knowledge work where trust, relationships, and stakes matter.
COVID spread because you can’t choose not to be exposed to a virus. AI adoption spreads only if companies choose it. And that choice is complicated. (More on that in a sec.)
The Race Nobody Asked to Run
Matt’s entire argument rests on one assumption: AI can do it faster and better, therefore it wins. It’s a refrain we hear from a lot of AI companies.
DEEP SIGH.
It’s the same logic as the AI book-mill romance “writer” profiled in the NY Times recently who asked:
“If I can generate a book in a day and you need six months to write one, who’s going to win the race?”
Matt is also asking: Who’s going to win the race? The answer for both the AI book-mill operator and Matt is, of course, AI. The robot wins. ROBOTS ALWAYS WIN!
But I keep coming back to: What race? Who said we’re racing? What are we racing toward?
Matt’s piece is saturated with racing language: “get ahead,” “head start,” “falling behind,” “the people who will struggle most are the ones who refuse to engage.”
The entire framing assumes speed is the only metric that matters. That faster is always better. That if you’re not optimizing for velocity, you’re losing.
I could not disagree more with this worldview.
What Gets Lost
Here’s what’s absent from Matt’s piece:
- Any question about what work is actually worth doing, and why.
- Any acknowledgment that some friction creates value rather than destroying it.
- Any consideration that the things that take time might take time for a reason.
- Any curiosity about what we lose when we optimize purely for speed and efficiency.
He writes: “Have no ego about it. The managing partner at that law firm isn’t too proud to spend hours a day with AI.”
But what if it’s not about ego? What if some kinds of work require the time because that’s where the judgment gets built?
Matt thinks a managing partner at a law firm is valuable because she can research case law faster or draft briefs more efficiently. But that’s not actually why clients hire her, is it?
They hire her because she knows which judge will respond to which argument. Because she can read a room. Because when things go sideways, she knows who to call. Because she’s been practicing for 30 years and has judgment that comes from making mistakes and living with consequences. Because clients trust her.
AI can research the law. It can’t sign the filing. It can’t go to court. It can’t absorb the malpractice liability. It can’t build a relationship over a decade.
What About Marketers?
That’s law. But what about marketers? Matt’s argument would say that a senior marketer is valuable because he can write copy faster or analyze campaign data more efficiently. But that’s not actually why companies hire him.
Companies hire him because he knows which message will land with which audience. Because he can feel when a campaign is off before the data proves it. Because when a launch goes sideways, he knows how to salvage it. Because he’s run enough campaigns to know the difference between a metric that matters and a dumb vanity number. Because he understands the customer in ways that don’t show up in a brief. Because he knows how to lead a team.
AI can’t sit in the room when the CEO wants to pivot the entire strategy three weeks before launch. It can’t rebuild trust with a customer after a PR crisis. It can’t read the culture and know that this is the wrong time to be clever. It can’t take accountability when the campaign fails. It can’t lead that team.
Those aren’t small details, friends. They are structural foundations.
One more example: this time a personal one.
I just spent almost 2 years writing a book. Was it inefficient? “Yes,” Matt would say. “AI can do it faster.”
Listen: You are wrong, my friend. It was efficient. Because I was doing the work of figuring out what I actually think and how to help a reader and say it in a way that could come only from me. The friction between idea and expression, between intention and execution is not frivolous. That’s where the thinking happens.
Matt would say: “But AI can help you think faster.”
Mmm. Maybe.
Or maybe it helps you skip the thinking and arrive at answers that sound right but haven’t been tested against the resistance of actually working through them.
See the issue?
The Timeline Shell Game
One more thing: Matt is absolutely certain about timelines. “1-5 years.” “By the end of this year.” “Doubling every 7 months, accelerating to 4 months.”
It all sounds very scientific. Very data-driven. You nod along as you read. You feel the panic rise.
Here’s what he mentions but doesn’t fully reckon with: he runs an AI company and invests in AI startups. He discloses this upfront, to his credit. But he doesn’t seem to acknowledge how profoundly his financial and psychological incentives shape his certainty that the revolution is happening right now.
The people who were closest to crypto in 2021 were absolutely certain it would replace the financial system.
The people closest to VR in 2016 were certain we’d all be working in the metaverse by now.
Proximity creates conviction. I kept whispering that to myself all the way through the reading of it.
Matt writes like the exponential curve is guaranteed to continue at exactly this pace. But these curves always hit walls— not because of technological limits, but because real-world adoption is messy and human-shaped. Things like regulation, litigation, integration costs, resistance to change, and the fact that humans are messy and uneven and irrational and are wired a certain way.
The gap between “this works in a demo” and “this has replaced 50% of jobs” is as enormous as an ocean. And Matt hops over it like it’s a puddle.
What I Think Is Actually True
Some jobs will disappear entirely. Many more will be transformed—you’ll do the same role but with AI handling parts of it.
Each of us will face a “learn this or become less valuable” moment.
The timeline will be longer than 2-5 years. How long? No one knows (harder still to know when you’re inside the bubble.) And the change will be partial and uneven, not the clean apocalypse Matt describes.
No one actually knows anything. We are all guessing.
Except there is one thing I am certain about:
When speed becomes cheap, judgment carries a premium.
Matt is right that AI is making speed cheap. He’s wrong that the only response is to move faster.
Sometimes the right response to acceleration is to slow down enough to ask better questions:
- What am I actually trying to make?
- What’s worth protecting because it compounds over time?
- Where does friction create value instead of destroying it?
- What work do I love doing, regardless of whether AI could do it faster?
The Choice Matt Doesn’t Offer
Matt presents a binary: adapt or die. Use AI aggressively or get left behind.
But that’s a ridiculous choice.
You can use AI as a tool and still protect the friction that creates value.
You can learn what it’s good at and maintain clarity about what you’re trying to make and why.
You can take the capabilities seriously and reject the panic.
The real danger isn’t AI. It’s letting someone else’s timeline and someone else’s definition of winning determine how you spend your life.
What I’m Not Saying
I’m not saying ignore AI. I’m not saying it won’t change things. I’m not saying Matt is wrong about capabilities.
I’m saying: DO NOT LET PANIC DRIVE YOUR DECISIONS.
(I put that in ALL CAPS. Because it’s important.)
Don’t restructure your entire life around predictions from someone who has strong incentives to believe the revolution is coming right now.
Don’t abandon the work that takes time just because something else can be done faster.
And for Pete’s sake, don’t accept “who’s going to win the race?” as the only question worth asking.
A Different Question
Matt ends his piece: “The future is already here. It just hasn’t knocked on your door yet. It’s about to.”
I’ll end with this:
The future is always arriving. The question is: what are you building that’s worth the time it takes?