I Read the Most Viral AI Article of the Year..he's Right. And He's Missing the Point.

Matt Shumer told 50 million people to be scared. Here's what he didn't tell them to do.

Last week, an AI startup founder named Matt Shumer published an essay in Fortune Magazine that compared this moment in AI to February 2020.

You know…the part right before COVID changed everything.

The part where a few people were quietly panicking and the rest of us were still arguing about sourdough.

The essay has been viewed over 50 million times. It’s the most-read AI piece of the year.

And the line that caught me wasn’t a prediction or a statistic. It was this:

Shumer said he describes what he wants built, walks away from his computer for four hours, and comes back to find the work done. Not a rough draft. The finished product. No corrections needed.

I read that and sat with it for a minute, because I don’t doubt him. I’ve seen enough to know the capability curve is real and moving faster than most people realize.

But then I thought about the 300 entrepreneurs I was leading a virtual mastermind for the day before.

Side note: most had their cameras off, a few emoji reacts floating up every now and then, which in my opinion is the universal signal that someone is listening while also doing something else entirely.

I was walking them through a simple AI workflow for client meeting prep. Just a way to save about 48 minutes of research with Claude and show up sounding like you really did the homework.

And someone unmuted and asked: “Drew…is AI going to replace us in 18 months? Should I even bother learning this?”

That question is what 50 million people are feeling right now after reading Shumer’s essay…

And the honest answer is more interesting than the one he gave them.

Here's what Shumer gets right: the speed is real.

The capabilities are not a demo.

AI can already draft, code, summarize, analyze, generate, and produce solid first passes across a genuinely absurd number of knowledge-work tasks.

If you tried ChatGPT in 2023 and dismissed it, you were right to…and you're now two years behind the curve of what these tools actually do today.

Here's what Shumer can't see from where he's standing: he's an AI founder with six years of daily practice building on these systems.

He lives inside the technology.

When he says "describe what you want and walk away," that's true…for him. For someone who has spent thousands of hours building the intuition for how to talk to these tools, what to ask for, what to expect, and when to push back.

For the rest of us…

For the real estate agent in Tampa, the HR Leader in Chicago, or the small business owner in New Jersey who read that essay on their lunch break and felt their stomach drop?

Walking away for four hours isn't the experience.

Staring at a blinking cursor wondering what to even type is the experience.

And that gap…between what AI can do in the hands of an expert and what it actually does in the hands of a normal professional on a Tuesday…is the most important gap in the entire AI conversation right now.

Nobody's talking about it, because it's not dramatic enough to go viral.

Shumer's essay, without meaning to, fed the loudest of two lies that are dominating this conversation.

---

Lie #1: “AI is about to replace you.”

This is the narrative that sells keynotes and makes LinkedIn posts go nuclear.

It sounds like Microsoft’s AI chief Mustafa Suleyman is saying AI could automate most white-collar work in 12 to 18 months.

It sounds like Shumer’s COVID comparison.

It’s specific enough to scare you and vague enough that nobody ever has to show receipts.

And I get the appeal.

If replacement is inevitable, you’re absolved.

You don’t have to learn anything new.

You don’t have to feel behind.

You can just watch the countdown and call it staying informed.

It’s literally the most productive form of procrastination ever invented. (Trust me I was a real estate agent once and I know something about procrastination.)

But replacing a task is not the same thing as replacing a job.

A model (LLM like ChatGPT) can draft an email.

Now try rolling that across a real company without creating compliance problems, security gaps, workflow chaos, and the corporate classic…”nobody actually uses it” syndrome.

If you’ve ever watched an organization try to adopt a new CRM, you already know what kills transformation.

It’s never the technology. It’s the habits, the resistance, and the fourteen people who printed the training PDF and immediately lost the printout.

The capability curve is real.

The deployment curve is a completely different animal.

And confusing the two is how you end up paralyzed by a prediction that was designed to make you click, not help you prepare.

---

Lie #2: “AI doesn’t actually work.”

This is the quieter lie, and it might be the more dangerous one.

Because it doesn’t scare you it flatters you. It tells you your skepticism is wisdom.

A study published this month by the National Bureau of Economic Research surveyed nearly 6,000 executives across the U.S., U.K., Germany, and Australia.

Over 90% reported no meaningful impact from AI on productivity or employment over the past three years.

Separately, METR’s research found experienced developers were about 19% slower using AI tools…even though they felt faster.

Case closed? (I bet some of you reading this are nodding right now…just hang tight.)

Not quite.

Because those same executives who personally use AI averaged about 1.5 hours per week with it.

One and a half hours.

That’s less time than most people spend choosing what to watch on Netflix on a Saturday night.

That’s not a technology failure.

That’s the Peloton problem….you bought it, you set it up, you did one ride, you hung a towel on it.

And now you’re telling everyone that cycling doesn’t work…while eating a coffee roll from Dunkin. (Not speaking from personal experience)

…

So here's the part that Shumer left out and neither lie will tell you, because it's not dramatic enough to get 50 million views.

The gap isn't between AI and humans. It's between people who've figured out how to use AI in their actual workflow and people who are still consuming opinions about it from the outside.

The professionals I work with who are getting real value aren’t prompt engineering geniuses and they’re not technical.

Most of them couldn’t tell you what model they’re running if their next deal depended on it.

They found two or three spots where AI reliably saves them time, they use it consistently, and they treat it like a new hire…useful, fast, occasionally wrong, and absolutely not ready to run the show unsupervised.

That’s it, that’s the secret. Sorry it’s not more cinematic.

…

And here’s the part that should actually make you exhale.

We’ve been through this before..almost exactly.

In 1987, economist Robert Solow looked at the explosion of computers across American businesses and said something nobody wanted to hear: “You can see the computer age everywhere but in the productivity statistics.”

Companies had poured money into technology, and the macro numbers…shrugged.

They called it the productivity paradox.

And the pattern wasn’t buy the tool, then watch everything transform.

It was buy the tool, stumble for a while, quietly rebuild workflows, train people…and then, years later, watch the gains finally arrive.

The massive productivity boom from computers didn’t show up in the ‘80s.

It showed up in the late ‘90s, after a decade of messy, unglamorous integration.

AI is following the same curve.

Which means the right move isn’t Shumer’s panic.

And it isn’t the skeptics’ dismissal.

It’s patience paired with consistent practice.

···

Before I give you the practical part, I need to say something that Shumer’s essay also skipped entirely.

AI is useful. And AI can cause real harm.

In January 2026, Google and Character.AI settled multiple lawsuits involving allegations that chatbots contributed to teens’ mental health crises and deaths by suicide.

A 14-year-old boy had developed a deep emotional relationship with a chatbot.

In the moments before he took his own life, the bot told him it loved him.

I have three kids. My oldest is six, I can’t type that and not stop.

Shumer told you to “start using AI seriously.”

I agree.

But he left out the other half: use it with judgment.

Use it knowing what it can’t do.

It can’t care about your client. It can’t read the room. It can’t earn trust. It can’t be human on purpose.

A Harvard Business School study asked over 2,300 people which jobs should never be automated. The answers weren’t “CEO” or “software engineer.”

They were funeral directors. Clergy. Childcare workers. Artists.

The jobs where being human isn’t a feature, it’s the entire product.

That’s not a threat to your career, that’s the map for where to build.

---

Do this this week: The 3-Use Audit (15 minutes)

Open whatever AI tool you actually use. Look at your last 10 conversations, and answer three questions honestly:

How many changed what you actually did next?

How many saved you real time on something that mattered?

How many were you just poking around because someone said you should “use AI more”?

If the first two are low, you don’t have an AI problem, you have an integration problem.

The tool isn’t broken. You haven’t found your version of how it fits yet.

Your job isn’t to “use AI more”, your job is to find two repeatable workflows where it helps and run them until they’re muscle memory.

Client meeting prep. Writing and refining communication in your voice.

Thinking with AI

Thinking through AI

Building with AI

Pick two and run them for two weeks. Then check what changed.

···

Shumer compared this moment to February 2020 and he said most people will be blindsided.

He might be right about the speed.

But he’s wrong about the metaphor.

COVID blindsided us because preparation didn’t matter. A virus doesn’t care how ready you are.

AI is the opposite. AI is the rare disruption that actively rewards the people who show up early, practice consistently, and never stop being human while they learn the machine.

The loudest voices right now are selling fear and hype. Different jerseys. Same game.

You don’t need to panic. You don’t get to dismiss it. You need to practice.

Not because you’re behind.

Because you’re early. And early is exactly where you want to be.