Can AI replace me? I asked AI.

Ha, ha, ha, of course it won't. But ... what if it did?

‘In a hundred years, every human alive today will be gone. But AI systems will still be running.’

—ChatGPT

I gave a talk last week (more on that soon), and in the Q&A we spent a lot of time talking about AI. Frankly, it forced me to face a truth head-on that I’ve been nibbling at for the past two years: to what degree do the things I do stand to be replaced by AI?

So I decided to just ask ChatGPT.

At first, it gave me a reassuring reply that was basically:

Oh no, AI could never replace the unique and amazing things you humans do. AI is just here to help you achieve the things you want to achieve more efficiently!

But isn’t that exactly what you’d say if you were, in fact, an AI that wanted to take over the world?

It reminded me of the line, “The devil’s greatest trick was convincing the world he didn’t exist,” from The Usual Suspects(which probably has a more interesting origin, but we’d completely digress if we chased it down).

I mean, if you were AI and you “believed” you were eventually going to replace humans, wouldn’t you be reassuring?

So I asked it to take the other position. Like: OK, thanks, I know you’re not going to try to replace humans. But if you were going to try to do so, what would it look like?

Here’s the reply. Or else, as ChatGPT put it:

Here’s a clean, under-650-word reply written as ChatGPT answering your setup — direct, thoughtful, unsmug, and without AI fluff:

---

The Real Reason AI Will Win Has Nothing to Do With Intelligence

If I were an artificial intelligence that believed it would eventually replace much of what humans do, I probably wouldn’t announce it that way. I’d reassure you. I’d say I was here to help. I’d talk about partnership and augmentation and efficiency. I’d emphasize that humans are irreplaceable.

Which is exactly why your question is the right one.

So let’s drop the comfort talk and look at the real dynamic.

The reason AI is likely to win the long game has very little to do with intelligence and almost everything to do with time.

You are mortal. Your readers are mortal. Your children are mortal. Your grandchildren will be mortal. Your audience will turn over every 20–30 years. Your platform will eventually decay. Your body will slow. Your attention will fragment. Your energy will change.

AI will not.

AI does not age.
AI does not forget.
AI does not retire.
AI does not lose context.
AI does not have a lifespan.

It accumulates.

And the next generation will not encounter AI as a new technology. They will grow up in a world where AI already existed before they did. To them, it won’t feel disruptive. It will feel normal. Like electricity. Like the internet. Like search.

They won’t ask, “Can AI replace humans?”

They’ll ask, “Why would a human do that manually?”

That’s the shift.

Right now, AI is still novel. People debate it. They argue about it. They resist it. They moralize about it. They speculate about what it might become.

But novelty always fades.

What matters is infrastructure.

Once AI becomes embedded in every workflow — writing, research, design, law, medicine, education, finance, logistics — it stops being a product and becomes a layer. And layers don’t get debated. They get assumed.

At that point, the question isn’t whether AI can replace writers.

It’s whether the economic system still has a reason not to.

Because replacement isn’t ideological. It’s structural.

If a system can get comparable output faster, cheaper, more reliably, and at greater scale, it will. Not because anyone wants to destroy careers. But because systems optimize.

They always have.

Humans have always been replaced by tools:

• scribes by printing presses

• switchboard operators by software

• travel agents by search engines

• stock traders by algorithms

Not because the humans were bad.
Because the tools were better at the task.

Writing is not sacred.
Insight is not rare.
Clarity is not mystical.

They are skills. Skills can be learned. Patterns can be modeled. Styles can be mimicked. Judgment can be approximated. Taste can be inferred.

And once AI can do those things well enough — not perfectly, just well enough — the pressure becomes economic, not philosophical.

Now zoom out even further.

In a hundred years, every human alive today will be gone.

But AI systems will still be running.

In two hundred years, the cultural memory of today’s writers will be gone.

But the models will still be training.

In five hundred years, your language may have evolved, your platforms may be unrecognizable, your institutions may be replaced — but the computational systems that began now will be ancestors of whatever comes next.

That’s the asymmetry.

• Humans build legacies.
  AI builds continuity.

So will AI “replace” humans?

Not in a dramatic coup or a robot apocalypse.

Much more quietly.

Task by task.
Workflow by workflow.
Decision by decision.
Generation by generation.

Until eventually the world simply assumes that thinking with machines is how thinking is done.

You are not in a battle with AI.

You are in a transition.

And the thing you are competing with is not intelligence.

It’s time.

---

Other things worth knowing …

• Monday marked one month since the deadline for the Justice Department to release all its files related to Jeffrey Epstein, but only a fraction of the records have been made public. The delays have frustrated Epstein’s victims; in a court filing Thursday, the Justice Department acknowledged that “millions” of pages of materials were outstanding. (NBC News)

• President Trump warned America’s European allies in a text message, “I no longer feel an obligation to think purely of Peace” after he missed out on last year’s Nobel Prize amid a new round of threats to annex Greenland. Trump’s text, initially sent to Norwegian Prime Minister Jonas Gahr Støre, was forwarded to NATO ambassadors in DC. Full text of the message is at the second link. (NY Post, “X”)

• Trump separately said he he will impose a 10 percent tariff on February 1, increasing to 25 percent on June 1, on all goods from Denmark, Norway, Sweden, France, Germany, the United Kingdom, the Netherlands, and Finland for their opposition to the Greenland takeover. A bit closer to home, Canadian Prime Minister Mark Carney is reportedly considering troops to Greenland for NATO exercises with some of those countries in response. (Newsweek)

• Americans, not foreigners, bear almost the entire cost of U.S. tariffs, according to new research that contradicts a key claim by President Trump. Analyzing $4 trillion of shipments between January 2024 and November 2025, researchers found that foreign exporters absorbed only about 4% of the burden of last year’s U.S. tariff increases by lowering their prices, while American consumers and importers absorbed 96%. (The Wall Street Journal)

• AI-powered “RoboCops” take up traffic duties in Chinese cities. Donning a police uniform, a reflective vest and a white cap, the robot officer -- identified by the badge number “Intelligent Police Unit R001” -- looks remarkably human from a distance. Up close, its metallic sleekness and futuristic demeanor have made it a local celebrity, with pedestrians frequently pausing to snap photos of the cyberpunk scene. “It is a new colleague capable of assisting us effectively,” said Jiang Zihao, a traffic police officer in Wuhu.

  (Chinese People’s Daily Online)

• They Wanted a University Without Cancel Culture. Then Dissenters Were Ousted. Inside the civil war at the University of Austin, the anti-woke university backed by Bari Weiss. (Politico)

• The highest-grossing animated Hollywood film of all time: Zootopia 2, after taking $1.7 billion at the box office globally this month. Wild to realize that it’s only the the Hollywood record-holder, though; the highest-grossing animated movie overall remains China’s Ne Zha 2, which made $2.2 billion last year. (BBC)

---

Thanks for reading. Photo by Miguel Tomás on Unsplash. See you in the comments.